Scatter, Adapt, and Remember: How Humans Will Survive a Mass Extinction

IF YOU THINK that humans are destroying the planet in a way that’s historically unprecedented, you’re suffering from a species-level delusion of grandeur. We’re not even the first creatures to pollute the Earth so much that other creatures go extinct. Weirdly, it turns out that’s a good thing. If it hadn’t been for a bunch of upstart microbes causing an environmental apocalypse over 2 billion years ago, human beings and our ancestors never would have evolved. Indeed, Earth’s history is full of apocalyptic scenarios where mass death leads to new kinds of life. To appreciate how these strange catastrophes work, we’ll have to travel back in time to our planet’s beginnings.

The Proterozoic Eon (2.5 billion–540 million years ago): Oxygen Apocalypse

Earth is roughly 4.5 billion years old, and for most of its life the atmosphere would have been noxious for humans and all the creatures who live here now. Vast acidic oceans roiled in what today’s environmental scientists would call an extreme greenhouse climate: the air was superheated and filled with methane and carbon. Our planet’s surface, now covered in cool water and crusty soil, was bubbling with magma. The solar system had formed relatively recently, and chunks of rock hurtled between the young planets—often landing on them with fiery explosions. (One such impact on Earth was so enormous, and threw off so much debris, that it formed the Moon.) It was on this poisonous, inhospitable world that life began.

About 2.5 billion years ago, early in an eon that geologists call the Proterozoic, a few hardy microbes who could breathe in this environment drifted to the surface of the oceans. These microbes, called cyanobacteria (or blue-green algae), knit themselves into wrinkled mats of vegetation. They looked like black, frothy coats of slime on the water, trailing long, feathery tendrils beneath the waves. All that remains of this primordial ooze are enigmatic fossils that hide inside a distinctive type of ancient, spherical rock called a stromatolite. If you slice a stromatolite down the middle, you’ll see thin, dark lines curving across its inner surface like the whorls in a fingerprint—these are all that remain of those algal mats. Only a few people in the world would recognize them as the traces of impossibly old life that they are, and Roger Summons is one of them. He’s a geobiologist at the Massachusetts Institute of Technology who has spent decades studying the origins of life on Earth, as well as the extinction events that wipe it out.

An Australian with a dry sense of humor, Summons has an office you can only reach by walking through his lab, a big, airy room full of tanks of hydrogen and bulky mass spectrometers that look like old-school Xerox machines covered in tubes. When I visited him to talk about ancient Earth, he plucked some slices of stromatolite from the top of a filing cabinet to show me the traces of algae that spidered across their surfaces. “This one is eight hundred million years old, and this one is two-point-four billion,” he said, pointing at each ragged half sphere of rock. “Oh, and this one is probably three billion years old, but it’s a crap sample.”

Even with a “crap sample,” Summons can pin a date on the fossils of creatures who lived more than 2 billion years ago by examining the sediments that have preserved them. In his lab, researchers grind up ancient rocks, subjecting them to vacuum, freezing, lasers, and a strong magnetic field before running them through the mass spectrometers. At that point, often nothing remains of a stromatolite but ionized gas. And that’s exactly what mass spectrometers need to decode the atoms in each sample. Atoms in minerals decay at a fixed rate, and reading the state of a rock’s atoms can tell scientists how long it has been since it formed. Geologists don’t put fossils themselves beneath the laser. They use machines like the ones in Summons’s lab to figure out the ages of the rocks next to the fossils. Call it dating by association.

Knowing when the oldest stromatolites were created helps us date an event which changed Earth forever. The mats of algae that became stromatolites weren’t just methane-loving scum. They were also filling the atmosphere with a gas that was deadly to them: oxygen. This is how the first environmental disaster on Earth began.

Just like plants today, ancient blue-green algae nourished themselves using photosynthesis, a molecular process that converts light and water into chemical energy. Cyanobacteria were the first organisms to evolve photosynthesis, and they did it by absorbing photons from sunlight and water molecules from the ocean. Water molecules are made up of three atoms—two hydrogen atoms and one oxygen atom (hence the chemical formula H₂O). To nourish themselves, the algae used photons to smash water molecules apart, taking the hydrogen to use as an energy source and releasing the oxygen molecules. This proved to be such a winning adaptation to Earth’s primordial environment that cyanobacteria spread across the face of the planet, eventually exhaling enough oxygen to set off a cascade of chemical processes that leached methane and other greenhouse gases from the atmosphere. The dominant form of life on Earth ultimately released so much oxygen that it changed the climate dramatically, soon extinguishing most of the life-forms that thrived in a carbon-rich atmosphere. Today we worry that cow farts are destroying the environment with methane; back in the Proterozoic, it’s certain that algae farts ruined it with oxygen.

Greenhouse Becomes Icehouse (and Vice Versa)

What happened after the rise of oxygen was an event shrouded in mystery until the late 1980s, when a Caltech geologist named Joe Kirschvink asked his student Dawn Sumner to research a rock whose existence seemed to be impossible—at least, given the prevailing theories about early Earth. Found near the equator, the rock’s surface was scored with marks that suggested it had once been scraped by the weight of a slow-moving glacier. In a short paper that eventually revolutionized geologists’ understanding of climate change, Kirschvink suggested that this rock offered a window on a late-Proterozoic phenomenon he called Snowball Earth.

Snowball Earth is what happens when our planet’s climate enters a very extreme “icehouse” state, the opposite of a greenhouse. A carbon-rich atmosphere can heat our climate up into a sweltering greenhouse, but an oxygen-rich atmosphere cools it down and causes what’s called an icehouse. Throughout its life, the planet has vacillated between greenhouses and icehouses as part of a geological process called the carbon cycle. Put in the simplest possible terms, a greenhouse happens when carbon is free in the air, and an icehouse occurs when carbon has been locked down or sequestered in the oceans and rocks. During an icehouse, ice collects at the poles, sometimes creeping down into lower latitudes during an ice age. But our recent ice ages were nothing compared with Snowball Earth.

Two billion years ago the sun was dimmer than it is today. As more and more cyanobacteria pumped out oxygen, the whole place began to cool down. Because the sun was a relatively weak heat source, this effect was magnified into a “runaway icehouse.” Ice from the poles began to spread outward, solidifying the top layer of the oceans and burying the land beneath vast, frozen sheets. The more ice that formed, the more it reflected sunlight—lowering the planet’s temperature further. Finally, ice stretched from the poles nearly all the way to the equator, pulverizing rocks beneath its weight. If you looked at Earth from space at that time, you’d have seen a slushy white ball, its circumference banded by a narrow equatorial ocean of algae-infested sludge. At that moment in geological history, our planet resembled Saturn’s icy moon Europa. It was an alien world called Snowball Earth.

I visited Kirschvink at the California Institute of Technology to find out what happened next. In the basement of the geology building, his generously sized desk was piled with fossils, family photographs, papers, and his prized possession, a cheap plastic vuvuzela from South Africa. “This is real!” he enthused, gesturing at the instrument whose droning sound annoyed and delighted audiences during the 2010 World Cup. Kirschvink lit up when he talked about the provenance of objects, whether pop culture ephemera or 3-billion-year-old fossils. Maybe it was his off-kilter imagination that allowed him to look for environmental patterns in Earth’s history that nobody had thought possible.

Kirschvink believes that there may have been as many as three snowball phases on Earth. “It was the longest, weirdest perturbation in the carbon cycle,” Kirschvink said. “And my explanation for it is simple. It’s the time between when the biosphere learned to make atmospheric oxygen and the time when everybody else learned to breathe it and use it.” Without any creatures around to breathe oxygen, the cyanobacteria likely created an atmosphere far more oxygenated than any we’ve ever known.

For 1.5 billion years after cyanobacteria evolved, Earth’s biosphere was in chaos. At least two more snowballs crept across the face of the planet, followed by intensely hot greenhouse conditions caused when volcanoes pumped carbon back into the air. Meanwhile, microbes were slowly learning to use oxygen to their advantage. A new kind of cell called a eukaryote began to populate the seas. Unlike cyanobacteria, which are basically just genetic material contained inside a membrane, a eukaryotic cell contains a nucleus packed with DNA as well as tiny organs called, appropriately enough, organelles. One of those organelles, called a mitochondrion, could turn free oxygen and other nutrients into energy. At last, Earth was inhabited by oxygen-breathers. The planet we know today was taking shape.

While the eukaryotes got busy swapping genetic material and sucking oxygen from the air, the old methane-breathers were dying out. A few migrated to the sea floor, finding niches near superheated volcanic vents where they could live in the remaining fragments of a once-global methane ecosystem. But the rest went extinct. It was the most extreme form of atmospheric pollution in Earth’s history, soon killing off almost every form of life that couldn’t breathe oxygen.

By “soon,” I mean within a billion years, or possibly 2 billion—a period of time that’s almost impossible to wrap our minds around. Still, that is the timescale required to understand Earth’s environmental transformations. Many of the catastrophic changes we’ll discuss over the next few chapters took millions of years to unfold. To geologists, we are all living in fast motion, our lives so short that it’s usually impossible for us to personally experience environmental change. Often, these scientists will contrast “human-scale” time with what they clearly view as real time, or time that unfolds on a planetary scale.

One of our most incredible accomplishments as a species, however, is an ability to think beyond our own life spans. We may not live in geologic time, but we can know it. And the more we learn about our planet’s past, the more it seems that Earth has been many different planets with dramatically different climates and ecosystems. This idea offers a much broader perspective than what you find in the work of environmentalists like Bill McKibben, who argues in his book Eaarth that humans have burned so much fossil fuel that we’re turning our planet into something fundamentally different (requiring the new name Eaarth). In that book and elsewhere, he laments the loss of “nature,” by which he means the ecosystems that existed on Earth before human meddling. But before humans took center stage on Earth, there were many permutations of nature. Climate disasters were the norm. Indeed, the only way Earth could ever transform enough to merit a new name like Eaarth would be if the planet’s environment suddenly stopped changing.

Undeniably, our planet is undergoing potentially deadly environmental changes today. But it’s incorrect to say that this is the first or even the worst time it’s happened. For the creatures who perished during the Proterozoic, and other periods we’ll learn about in the coming chapters, McKibben’s ideal of nature would be deadly. Over the course of its history, Earth has always vacillated between a carbon-rich greenhouse and its opposite, the oxygen-rich icehouse where humanity is more comfortable. We’re simply the first species on Earth to figure out how this climate cycle works, and to realize that our survival depends on preventing the next environmental shift.

Defining Mass Extinction

As bad as the oxygen apocalypse was, neither Kirschvink nor the geobiologist Roger Summons would call it a mass extinction. So how can an entire world full of life go extinct without it being a mass extinction? This brings us to the question of what mass extinction really is. In a remarkable paper published in Nature in the spring of 2011, a group of biologists from across North and South America exhaustively summed up all the data available from the fossil record and present-day extinctions and came up with a clear definition. They agreed that mass extinctions on Earth can be defined as events in which 75 percent or more species go extinct in less than 2 million years. The oxygen apocalypse didn’t happen fast enough to qualify.

The statistician and paleontologist Charles Marshall, a coauthor on that Nature paper, warns that the definition of “mass extinction” is highly contextual and slippery. Sitting with his back to an enormous window overlooking the UC Berkeley campus, Marshall told me that the key to understanding mass extinction always begins with a calculation of what researchers call the “background extinction rate.” Species naturally pass into extinction all the time, at a rate of about 1.8 extinctions per million species every year. On top of that, there are also natural cycles of elevated extinction rates that fall roughly every 62 million years in the fossil record. So just because a bunch of creatures are going extinct, even in numbers above the background extinction rate, doesn’t mean you’re looking at a mass extinction. The only time you’re really seeing a mass extinction, Marshall said, is when “you see a big spike sticking out of the background distribution.” While on Earth those big spikes tend to be times when 75 percent or more species go extinct, it’s all relative. “You could imagine a planet where the biggest spikes sat at thirty percent,” Marshall speculated. “On that planet, thirty percent of species dying out would constitute a mass extinction.”

There are some ways that the fossil record can trick us into seeing a mass extinction where there isn’t one. Take, for example, the bombs at Hiroshima and Nagasaki. The rates of death were high, but they were low in terms of the world’s population. If we looked at these atomic bomb strikes in the fossil record, it might appear that there had been a mass extinction, but that’s because we’d be mistaking the rates in one local area for a global phenomenon. When geologists study mass extinction in the fossil record, they constantly have to ask themselves whether the extinctions they’re seeing are a statistical anomaly like Hiroshima, or something more widespread. Mass extinction is not an absolute idea, and to measure it we have to prove that the extinctions aren’t just localized. Plus, we have to compare the rate of death to the normal background extinction rate.

Still, the oxygen apocalypse does resemble a mass extinction in one way. It ushered in a completely different world, populated by an entirely new set of life-forms. It gave rise to the atmosphere that allowed life as we know it to develop. The change was so dramatic, said Marshall, that “you’re measuring less by magnitude and more by the idea of a world changed forever.” In every mass extinction, the world is changed forever—but over a short, terrifying two million years, rather than a slow billion. In the next few chapters, we’re going to see exactly what that looks like.

NEARLY TWO CENTURIES ago, scientists trying to learn about Earth’s history visited England’s famous “white cliffs of Dover,” where the wind and water had eaten away at the land’s edge, revealing rocks unseen for millennia. There, these early geologists discovered that the cliffs weren’t just big, crumbling slabs of chalk. They were actually formed from distinct layers of rock, each containing very different sets of fossils, providing a chronological record of how the land was built up from mineral deposits and ecosystems through the ages.

Each geological period is named for one of these rocky layers. They’re chunks of frozen time, identified by their unique combinations of life-forms and mineral deposits. Generally, fossils change dramatically from layer to layer because there has been an extinction event. Though only five of these demarcations qualify as mass extinctions, there have been dozens of smaller extinction events where, say, 20 or 30 percent of all species die out. You might say that geological time is measured in extinctions. But if you were to visit Dover, and allowed your eyes to wander up the cliff face, you’d also see layer after layer of evidence that life always emerges from mass death.

The more we learn about these layers, the more it seems that there are two basic causes that can set a mass extinction in motion. The first is an unexpected calamity from the inanimate physical world, often taking the form of fatal climate patterns, megavolcanoes, or even debris from explosions in space. As for the second, as we learned in the last chapter from the cyanobacteria that poisoned the planet, biological life can transform the physical world so much that extinctions are inevitable. Of course, many extinctions are a combination of the two causes—one often leads to the other.

To see examples of both, we’ll journey back hundreds of millions of years to the first two mass extinctions that gripped the Earth. The Ordovician (beginning roughly 490 million years ago) and Devonian (beginning roughly 415 million years ago) were both periods when life was exploding with unprecedented diversity. And both ended in holocausts. The Ordovician was scourged by natural disasters from Earth and space; the Devonian was choked to death by invasive species that turned the planet into an environmental monoculture.

The Ordovician Period (490 Million–445 Million Years Ago): How the Appalachians Destroyed the World

Before the fecund Ordovician period, the seas had stopped looking like goo-covered murk and flowered into underwater forests full of aquatic plants, shellfish, coral, and lobsterlike arthropods called trilobites. New species were evolving at a rapid clip. It was a greenhouse world, with carbon dioxide levels in the air at fifteen times higher than they are today. But a warm, high-carbon climate was exactly what those Ordovician plants and animals needed.

Peter M. Sheehan, a geologist with the Milwaukee Public Museum, describes the Ordovician as having “the largest tropical shelf area in Earth’s history.” Put another way, it was a world of sultry beaches. Earth blossomed into this tropical paradise partly due to the climate, and partly due to continental drift, the process by which massive plates of the Earth’s crust slowly move around on top of the planet’s superheated molten layers. Lava gushing from underwater volcanoes applied so much pressure to the Earth’s crust that it pushed all the continents together, into the low latitudes of the warm southern hemisphere. Slowly drifting over the South Pole was a supercontinent called Gondwana, made up of land that became, among other places, Africa, South America, and Australia. Its balmy, world-wrapping coastline teemed with life.

Ordovician life was confined almost entirely to the oceans, though a few plants spread to the land. Trilobites scuttled into many different territories, evolving into a range of species: some became swimmers, while others wandered the floors of the shallow seas, developing sharp, defensive spines or shovel-shaped heads for rooting food out of the sediment. Shelled creatures and sea stars attached themselves to enormous coral reefs, and strange colony animals called graptolites built complicated, beehive-like structures out of proteins secreted from their bodies. Their hives, which looked like thorny, interconnected tubes, floated beneath the ocean surface while the graptolites poked their feathery heads out and snarfed up plankton.

The ancestors of sharks prowled the waters and fed on everything that moved (and some things that didn’t). Joining the sharks were jawless fish called agnathans, whose soft mouth slits and heads were covered in bony plates that probably looked like turtle shells. These armored fish were the first vertebrates, or animals with internal skeletons like we have. Plus, there were thousands of new kinds of plankton evolving all the time, creating an abundant food source for all the multicelled newcomers looking for easy-to-reach food floating through the waters.

But over a few hundred thousand years, over 80 percent of the species in the Ordovician coastal waters would go extinct.

We can place part of the blame for the slaughter on the Appalachian Mountains, a gently curving spine of peaks that stretched from Canada’s Newfoundland down to Alabama in the southern United States. These mountains were formed during the Ordovician when a smashup between two continental plates pushed ancient volcanic rock into jagged peaks above the continent. Almost immediately, rain and wind began eroding the soft, dark rock. The newly formed mountains ran with thick slurries of water and mud, which turned into rivers that picked up even more soil on their way to the seas. This natural process, called weathering, is actually one of the most powerful ways to change our planet’s atmosphere. As exposed earth crumbles beneath the weather’s onslaught, tiny rocks pull carbon dioxide from the air and take it with them into sediments deep beneath the sea. Sliding into the sea along with all that carbon was the Ordovician’s warm, life-nourishing climate.

Seth Young, a geology research associate at Indiana University, observed, “We are seeing a mechanism that changed a greenhouse state to an icehouse state, and it’s linked to the weathering of these unique volcanic rocks.” The Ordovician Appalachians weathered so rapidly, in fact, that they were worn down to a flat plain within a few hundred million years. The Appalachians we know today are the result of a second tectonic-plate smashup, which raised a new set of mountains about 65 million years ago. Washing carbon out of the atmosphere sounds like a good dream in our fossil-fueled times, but it was the worst thing that could happen in the Ordovician. Without greenhouse gases to keep the planet warm, disaster struck in the form of the fastest glaciation in the planet’s history. About 450 million years ago, ice caps began to spread outward from the poles. Gondwana and its hot, humid shorelines were at ground zero of the ice apocalypse.

As the glaciers grew, they locked up liquid water and lowered sea levels dramatically, drying out the lush coastal areas beloved by corals, graptolites, and shelled creatures. Most affected were stationary animals like the shellfish in a coral reef, which remain anchored in place for most of their lives. Because they couldn’t move, they died with their habitats. In all, Peter Sheehan estimates that about 85 percent of marine species died over a million years as massive ice sheets sucked the liquid out of their environments. Not all the Ordovician species died at once. There were two “extinction pulses,” as geologists put it. The first came when ice abruptly destroyed sea life. The second came when the ice melted just as suddenly as it had come, causing sea currents to slow and stagnate. Fewer currents meant that less oxygen was churned into the water and vast “dead zones” of anoxic (low-oxygen) water suffocated life throughout the oceans. First came ice, then came stagnation. Together, they created a mass extinction.

Despite what we know about the Appalachian Mountains, the wholesale slaughter at the end of the Ordovician remains a mystery. We understand why rapid freezing and thawing would kill so many life-forms. But typical ice ages are millions of years in the making, and this one lasted for less than a million, making it ridiculously rapid in geological time. Could weathering alone have caused the rapid glaciation in the first place? Probably not. It’s possible that the ice formation was hastened by invisible rays from space.

Cosmic Rays of Death

Adrian Melott, a professor of physics and astronomy at the University of Kansas, has long been fascinated by a weird fact about mass extinctions. They seem to fall roughly every 63 million years. Trying to explain why this might be, he stumbled upon one possible explanation for the swiftness of the Ordovician ice age. It has to do with the motion of our star through the swirling galactic disk of the Milky Way.

Every star in the galaxy has an orbit around the edges of the galactic disk. As our sun makes its vast circuit around the Milky Way, it bobs up and down, floating above or below the galaxy’s flat plane about every 60 million years. When it does this, our solar system brushes the edge of the protective magnetic field that envelops the galaxy, deflecting dangerous cosmic rays zooming through deep space (on a smaller scale, the Earth’s magnetic field protects us from these same particles). Cosmic radiation could help explain why extinction events are more likely to happen every 63 million years or so.

Cosmic rays are highly energetic subatomic particles that have been bouncing around in deep space since the early days of the universe. They can shoot right through a living creature’s body, damaging its DNA and causing cancer. And these particles aren’t much kinder to the molecules that make up Earth’s atmosphere. Cosmic rays can damage the ozone layer, which leaves the planet more vulnerable to deadly radiation. Melott hypothesizes that cosmic-ray bombardment could also whip up a thick cloud layer in the atmosphere, lowering temperatures and helping the ice caps to form more quickly.

As the planet cooled, extinctions would have been worsened by radioactivity hitting the planet’s surface. “At this point we’re thinking that … the radiation dose for organisms on the surface of the earth, or in the top kilometer of ocean water, could be very large. This causes cancer and mutations.” Melott paused, as if imagining a planet with gray skies racked by cancers and catastrophic erosion. Then he chuckled. “Or, you know, it could lead to giant ants that rampage across the Earth.”

His joke about the 1950s atomic monster movie Them!, featuring giant ants that take up residence in the sewers of Los Angeles, underscores the degree to which he thinks of his work as speculative. Cosmic rays, he conceded, were only one part of the problem that animals faced at the end of the Ordovician. “The analogy I like to give is that it’s like you have the flu and then you get shot. Cosmic-ray stress is like the flu.” But other factors—the bullet in Melott’s analogy—need to be in play. And these would likely be the volcanic activity that led to the uplift of the Appalachians, the weathering that flattened them, and the resurgence of volcanoes that shut down the Ordovician ice age as quickly as it began.

The Ordovician ended with an extremely rapid version of what happened during the snowball phases of Earth’s history. A swiftly changing climate, vacillating between icehouse and greenhouse, made it impossible for most species to survive. Because those deadly climate shifts happened so fast, geologists have dubbed this horrific period the Ordovician mass extinction, marking the first time our planet witnessed the deaths of so many species at once.

The Devonian Period (415 Million–358 Million Years Ago): Invasive Species

By the time the planet’s temperatures had stabilized, the Ordovician biosphere was gone forever. A few survivors remained, like the hardy trilobites. But for the most part, new animals and plants evolved to rule the seas, and a few creatures even crept up on land. Life diversified and flourished for 100 million years, a fairly long time even for a geologist. Consider that modern humans evolved only about 200,000 years ago, and you have an idea how many species evolved and died out during the 100 million years before the planet suffered its next mass extinction. Our next rendezvous with mega death came during the Devonian period. This time there were no dramatic claws of ice, cosmic rays, or greenhouse extremes—but that’s because this was the first mass extinction caused by life itself. By the end of the Devonian, 50 percent of marine genera (groups of species) and an estimated 75 percent of species were dead. Oddly, these species died out at an ordinary rate, probably no higher than the typical background extinction levels. So why is this even considered a mass extinction at all? Because almost no new species evolved to take the extinct ones’ places for as many as 25 million years. It was mass extinction by attrition.

Scientists call this phenomenon a “depression in speciation,” meaning a low point in the evolution of new species. If you had been floating around for thousands of years in one of the Earth’s oceans during the late Devonian, about 374 million years ago, you wouldn’t see corpses piling up. Nor would you see vast stretches of lifeless water as you would have during the late Ordovician. Instead, you’d see the same species slowly spreading everywhere, darting around in enormous coral reefs that were ten times more expansive than the ones we have today. There weren’t fewer life-forms during this mass extinction. There were just fewer kinds of them.

How did invasive species destroy the planet? It all had to do with the period’s peculiar ocean ecosystems. The massive sea creatures of the Devonian earned the period its nickname, the Age of Fishes. The eminent geologist Donald Canfield conducted a study of the atmosphere during this period, after which he and his colleagues concluded that the Devonian oceans contained a high amount of oxygen, which allowed the period’s enormous animals to evolve. A group of hardy armored fish called placoderms vied with sharks to become the ocean’s most forbidding predators. Placoderms grew up to 36 feet long and had faces entirely covered in armor; they were also among the first creatures to develop jaws. (Sharks won the scary toothed predator contest in the end, though—placoderms went extinct.) Reefs made from algae and early sponges—all species lost to this extinction—were dramatically unlike the coral-dominated reefs we know today. The ocean floors crawled with ammonites, which looked something like octopuses with spiraled shells.

Watery habitats were everywhere—even on the continents. Enormous tropical inland seas dominated the landmass that later became North America. Most of the Midwest and the central United States were fully submerged, which is why paleontologists today find some of the best fossilized fish in the middle of the Midwest’s rolling prairies, which are about as far from the coast as you can get. Yet by the end of the Devonian, almost none of the gigantic armored fish and swarms of tentacled ammonites were left. What happened?

One paleontologist, Ohio University’s Alycia Stigall, has a theory that could explain why life during this period went from diverse to homogenous. She believes that invasive species took over the world’s oceans and inland seas, the same way cockroaches, kudzu, rats, and humans have spread across the globe today.

Stigall lives in Ohio, at the bottom of what was once a shallow Devonian sea. In fact, the vista from her windows is the former seafloor of an inland ocean hit particularly hard by the mass extinction. “We don’t have a good modern analogue for these types of oceans,” she said. They are a completely vanished ecosystem, though she imagines they might have been something like Hudson Bay. At the end of the Devonian, it’s likely that sea levels were high, pooling several inland oceans together. Earthquakes thrust new mountains from the land, which also brought previously separated ecosystems into contact with each other.

Many highly adaptable or generalist species began invading new watery territories. That meant they were competing with the local specialist species, like trilobites, for food. A specialist species requires very specific temperatures or food sources. They couldn’t cruise all over the Devonian oceans eating anything that came into view the way sharks could. So when invasive species came into their territories and stole their food, the specialists had nowhere to go. Their populations dwindled and they went extinct. By the end of the Devonian mass extinction, the planet was covered with giant, homogenous inland oceans where you’d see the same generalist species no matter where you looked. And out of those circumstances, Stigall believes, you had the makings of a mass extinction.

Though Stigall’s account is only one of many theories about mass extinction in the Devonian, her claims are backed up by evidence that many of the creatures who survived the period were generalists like sharks—creatures who could live anywhere and eat almost anything. Another survivor was the humble crinoid, a starfish-like creature with several feeding arms surrounding its mouth that make it look a little like the “face-hugger” stage of the creature in Alien. The crinoids went through a floating larval stage that allowed them to drift into many new environments before attaching to the ground and feeding on the abundant plankton in the water.

Still, homogenous ecosystems like the ones in the Devonian would have left all life-forms primed for disaster. Generalist species may be hardy, but they also share the same vulnerabilities. Say, for example, a drought hits an ecosystem in the Midwest. If there is only one type of wheat species, and it can’t deal with higher temperatures or less moisture, a short-term climatic change could kill off every blade of wheat in a given region. Without a diverse range of grain species, which might have different moisture tolerances, drought slays all the wheat. This in turn kills the animals who feed on that wheat, whose deaths leave predators hungry, too. Soon, you have multiple extinctions because the whole food chain has been ravaged. “The more we cull diversity, the more we are vulnerable to extinction,” Stigall concluded. It’s very possible that this is how the Devonian came to a close, with only a few invasive species scraping by until speciation, or new species evolution, made the ecosystem diverse again.

For Stigall, there’s a lesson in the fossils that surround her home, remnants of an inland sea unlike anything that exists now. “We’ve got the same problem with invasive species today,” she said. “It’s not caused by sea levels, but by humans, because we like to move things around.” She described how invasive species like pigeons and rats, as well as some trees and grasses, have gone from a few local regions to expand across the Earth. If this trend continues, she predicted that “long term, what you expect is a huge decimation of total biodiversity.” We could be on our way back to the late Devonian, in the early stages of a mass extinction that begins with a depression in speciation and ends with deadly homogenization.

Sometimes, the urge to live by expanding into as many territories as possible can backfire. As the invasive species of the Devonian reveal, life does not always beget more life. Some ways of living can actually kill just as handily as climate change and radiation.

As devastating as the Ordovician and Devonian mass extinctions were, they were nothing compared to what ravaged the planet about 75 million years later. The “Great Dying,” as it is known among geologists, was the worst period of mass death the Earth has ever known. It’s very likely that this event had no single cause; it was set off by a combination of disasters from the physical world and the biological.

THE BERKELEY GEOCHRONOLOGY CENTER, a lab devoted to studying the ancient ages of Earth, is located on a pleasant tree-lined ridge overlooking UC Berkeley. Somewhat puzzlingly, it shares a building with the Church Divinity School of the Pacific. While seminary students strolled by in the hall outside, I met Paul Renne, the center’s head geologist, a big, jovial man in a T-shirt. As he walked me through a warren of labs—full of the lasers and mass spectrometers that I’d come to expect in such places—I realized there was a peculiar kind of symmetry to the Geochronology Center’s tenancy arrangement with the Divinity School. After all, I had come to ask Renne about the closest thing to a total apocalypse the planet has witnessed.

The Permian Period (299 Million–251 Million Years Ago): Life in the Time of Megavolcanoes

Two hundred fifty million years ago, at the end of the Permian period, Earth spent thousands of years dying. At the end of those millennia of carnage, almost 95 percent of the species on the planet were dead. It was the worst mass extinction in our planet’s history, earning it the moniker “the Great Dying.”

The first phase of the mass extinction was caused by a disaster that has left an indelible and easily deciphered mark upon the Earth. If you visit the vast area known as the Siberian Traps today, you’ll find a beautiful, hilly terrain covered in short grasses. But 250 million years ago, the region was drowning under liquid rock spewing from the ragged mouth of a megavolcano. As its name implies, a megavolcano is far more powerful than your typical lava-filled mountain. Renne and other geologists estimate that as much as 2.7 million square kilometers of basaltic lava swept across the land in a fiery deluge. Almost a million square kilometers of the hardened basalt rock still remains here, smoothed by erosion into plateaus and valleys. It’s unclear whether this almost unimaginable ocean of lava was unleashed by one or two enormous eruptions, or a single, ongoing eruption that lasted for centuries.

But the Great Dying wasn’t caused by flaming tides of death. Volcanic eruptions on a large scale release a lot of gases, including greenhouse gases like carbon dioxide and methane. Jonathan Payne, a geologist at Stanford, estimates that the eruptions unleashed 13,000 to 43,000 gigatons (a gigaton is 1 billion tons) of carbon into the atmosphere. As if that wasn’t enough, they also released highly reflective sulfur particles that remained suspended in the atmosphere, scattering light away and cooling the climate very rapidly. The culprit responsible for the Great Dying was climate change.

Ironically, the roiling fires from this Siberian megavolcano may have caused a brief ice age. As glaciation locked coastal waters into ice sheets, the sea level dropped, and another source of greenhouse gas was unleashed. It’s possible that the water dipped low enough to expose methane clathrates, huge deposits of frozen methane that cling to the edges of continental shelves deep beneath the ocean. The clathrates melted and released ancient methane, a powerful greenhouse gas. As quickly as it began, the Permian ice age would have ended with a more intense greenhouse than before. These radical transformations in the atmosphere and climate made it impossible for most creatures to survive. Food sources dwindled. Species upon species died out.

It was an ugly ending for the Permian, which had been a time of rapid animal evolution on land. When the megavolcano began erupting, the earliest ancestors of today’s mammals were walking the Earth. Gingkos and conifers covered the coasts in forests, while seeded ferns evolved, uncurling their leafy fronds beneath tall pines. Mammal-reptile hybrids called synapsids roamed the land, some looking like giant lizards, and some like small rhinos. One of them, the enormous, dragon-like predator dimetrodon, had a tall sail attached to its back like a bony fin, and was such a badass hunter that paleontologists believe it may have fed upon sharks. These creatures all thudded around on the same continent because plate tectonics had finally pushed the planet’s landmasses together into one enormous continent called Pangaea, which stretched from pole to pole. A globe-wrapping ocean called Panthalassa teemed with sea creatures, from tiny single-celled organisms to corals and large fish.

These new forms of life, the forerunners of so many animals and plants we take for granted today, almost didn’t make it. What was especially unusual about the Permian mass extinction was that it took out nearly every form of life. Unlike in other mass extinctions, which sometimes hit sea creatures but not land creatures, or animals but not plants, this extinction was absolute. As many species were lost at sea as on land. When the megavolcano pumped carbon into the atmosphere, a lot of that got dissolved into the oceans. The water grew warmer, which destroyed the habitats of shellfish, who are sensitive to temperature changes. It also grew more acidic. The shells of shellfish are made of calcium carbonate, which dissolves in acid. Many sea creatures didn’t survive simply because their offspring couldn’t form shells in a highly acidic ocean environment.

Meanwhile, on land, so many trees and plants died that the continent’s surface was “denuded,” as Payne put it. The result was shockingly rapid weathering. As acid-tinged rain poured from the sky, followed by hot winds, more soil poured into the oceans, further raising the levels of carbon and acid. Vast areas of the coastal seas became anoxic dead zones—regions completely purged of oxygen. With oxygen supplies low in the water, large fish could not survive, especially ones that lived close to the deeply damaged ocean’s surface.

Even insects, which generally survive everything, suffered extinctions. An estimated 9 out of 10 marine species and 7 out of 10 land species went extinct. Across the planet, carbon levels suddenly skyrocket in rocks from this period examined by Renne and his colleagues, which suggests that the dead bodies of plants and animals were quite literally piling up on land and at sea. As the plants rotted, they released even more carbon into the environment. The devastation was so complete that we see a “coal gap” in the layers of rock left behind from this era. Plant life, which decays into coal, was so sparse in the 10 million years following the end of the Permian that none of the fossil fuel could form.

The planet had already endured ice ages, greenhouses, cosmic rays, and speciation depression. But only in the Permian mass extinction were almost 95 percent of all species cut down. And it happened in just 100 thousand years—the blink of an eye in geological time.

Slime World Survivors

Still, there were survivors. The Stanford geologist Payne showed me a rock that’s a slice of geological time from this period, where a layer of ocean-floor sediment filled with tiny shells is topped by a black layer of what looks like pure sludge. It’s easy to see that a diverse community of creatures was abruptly replaced by nothing but, well, slime. Payne and his colleagues have nicknamed this era Slime World, because the oceans were dominated by dark, oozing bacterial colonies, feasting on the dead bodies of their multicellular cousins.

On land, one of the great survivors was Lystrosaurus, an animal that managed to thrive. A heavy, clubfooted creature with a beaked snout and two tusk-like teeth, Lystrosaurus was a four-legged synapsid, or mammal-reptile hybrid. About the size of pigs, lystrosaurs were burrowing animals whose muscular hindquarters ended in short, wiggly tails. And they somehow managed to endure when even the precursors of the hardy cockroach were dying. They were herbivores, and their beaks probably allowed them to chomp on rough vegetation and dig for roots to eat.

For several million years after the end of the Permian, lystrosaurs were alone on a dead world. But they didn’t cower or retreat. Instead, they spread out as far as they could across the landmass that would one day fracture into the continents we know today. Their fossils have been found in Africa, Asia, and even Antarctica, which was a tropical region at the time. With no predators and no competition for their favorite foods, lystrosaurs could waddle anywhere they liked. They are, as far as we know, the only creatures ever to dominate our world so thoroughly: For millions of years, most four-legged land creatures were one type of lystrosaur or another.

Why did these creatures—our distant ancestors—survive when so many of their fellow creatures didn’t? Theories abound. The Permian expert Mike Benton said it’s possible that they were “just lucky.” More likely, he added, they were well adapted for a world with depleted oxygen. They lived in underground tunnels, so they had a natural way to escape the heat and fire of the initial volcanic eruptions. Plus, the air they were used to breathing in their burrows was likely to be low in oxygen and full of dust—sort of like the air after carbon has been saturating it for a few centuries. Their barrel chests held lungs of a tremendous capacity, which meant more oxygen uptake. Lystrosaurus had the right respiratory system at the right time.

Over time, Lystrosaurus’s progeny repopulated the southern part of Pangaea, diverging into many subspecies. Their favored half of the supercontinent eventually broke off from the northern half and became its own continent, Gondwana (named after the southern Ordovician continent), packed with dinosaurs and proto-mammals. It took 30 million years for our planet to grow a robust ecosystem again, packed with predators and herbivores and a wide range of flora and fauna.

The Early Triassic Period (250 Million–220 Million Years Ago): Unraveling Food Webs

Those 30 million years of ecosystem struggles are their own story. Though every mass extinction unfolds differently, they all end when a new community of creatures has established itself—generally, a community that statistician Charles Marshall described as “completely different life-forms.” After the Permian, during the early millennia of the Triassic period, new communities of completely different life-forms rose and fell with alarming regularity. A new ecosystem would come together only to collapse in a few million years. Then another ecosystem would arise. This mass extinction just wouldn’t end.

Why did it take the planet so long to recover from the Great Dying? For answers, I visited Peter Roopnarine, a zoologist at the California Academy of Sciences who has a rather singular occupation among scientists. He’s developed a computer program that simulates food webs, the complex interplay between predators and prey within an ecosystem. Using this program, Roopnarine studies why the worst part of mass extinctions isn’t necessarily the fire, or the eruptions. It’s what comes afterwards, in the centuries of what scientists call “indirect extinctions” caused by food webs that are too unstable to support life.

The old computer game Wator offers a perfect example of a simple food web simulation. In it, red pixels stand in for sharks (predators) and green pixels for fish (prey) as they battle it out for supremacy of the sea. You can set a few simple parameters, such as how many sharks and fish there are to start, how often they breed, and how long it takes before they starve. Then you press “start” and watch generations unfold in seconds. When there are too many sharks, or the fish breed too slowly, the population of sharks eventually dwindles to zero and the waters of “Wator” become a sheet of uniform green. And that means you fail. What Wator reveals is that predators are as much at the mercy of prey as the reverse. Food webs can be knocked out of balance by life-forms at any point in the food chain.

Roopnarine’s simulations are infinitely more complex than Wator, incorporating the smallest planktons to the largest predators, and everything in between. In them, he describes relationships between predators and prey that lived millions of years ago. And from these models, he’s generated a theory about why the Triassic burned through so many food webs.

He began by coming up with a way to generate a realistic food web for species that no longer exist. He included every known form of life from the fossil record, and then he extrapolated predator–prey relationships based on what we know about how animals behave today. “You can’t ever know exactly what a fossil animal was eating—you can’t even know that with animals today,” Roopnarine explained. “But we can use the body size, tooth shapes, and other things to decide who their prey might have been.” Predators’ body sizes are helpful because, obviously, a small predator will prefer small prey, while a larger predator might be a generalist who can eat creatures of many sizes.

There are always complications, Roopnarine admitted. Many species share the same potential prey or predators, and it’s hard to know which species might have been generalists with many food sources, or specialists with just a few. But the beauty of using computers to simulate food webs is that you can go through as many iterations as you like, creating different worlds each time paleontologists discover more about a fossil predator’s range or appetites. Plus, there are a few rules of thumb, including the fact that there are usually far more specialists than generalists. Once the ancient food web has been set up in his program, Roopnarine said, he can simulate food-web disturbances like the one in the early Triassic.

Based on what he’s figured out so far, Roopnarine’s theory is that a basic imbalance in early Triassic food webs led to millions of years of maimed ecosystems rising and collapsing in rapid succession. Initially, the problem was that so few creatures had survived the Permian mass extinction. Of the survivors, he said, “you have small carnivores and some seriously big, bad amphibians who are the precursors of crocodiles.” Among herbivores, he said, lystrosaurs were the only game in town. The problem was that nobody around seemed to be eating Lystrosaurus, perhaps because they were the wrong size or in the wrong environments for most predators. In fact, the food web began to unravel because there were so many carnivores and very few prey.

Those “big, bad amphibians,” known as crurotarsans, were in fierce competition with each other. With their huge, toothy mouths and muscular tails, they would have been deadly predators—and Roopnarine says that the competition between carnivores during the early Triassic was more intense than in any other food web he’s looked at. The carnivores competed with each other so intensely for the tiny amount of available prey that they wound up driving each other to extinction over and over. New creatures would evolve, then get crushed out of existence. Only the herbivore Lystrosaurus, and eventually other herbivores, really recouped their losses. It took tens of millions of years before there were a small enough number of carnivores for food webs to stabilize.

Community Selection

This raises the question of what makes for a stable food web over the long term. And there’s an easy answer. “Diversity,” Roopnarine said firmly. A food web needs to be “robust,” full of many kinds of carnivores, herbivores, and plants, in order to withstand an environment that can often hammer creatures with everything from volcanoes to drought and sea-level shifts. As long as there are many nodes in a food web, a healthy balance of predator and prey, you have a community of life-forms that can remain steady even when the environment wobbles.

“So does that suggest some communities are better than others when it comes to survival?” I asked.

Roopnarine offered a conspiratorial nod. “This can be controversial, but yes, you could say this is natural selection at the community level.” Food webs don’t compete the same way two species might because they don’t exist next to each other, trying to eat the same things and live in the same caves. Instead, they compete with each other temporally, replacing each other in the same geographical places over time. To “win” the natural-selection game, a food web must outlast other food webs, remaining stable for as long as possible in the same place. Looked at from this perspective, you might consider all of Earth’s geologic history a competition between food webs struggling to last through as many environmental disasters as possible, simply by retaining their robustness in the face of calamity.

Survival is never just a matter of one species being exceptionally adept. We only survive in the context of our food webs. And when a food web starts to unravel, the extinction of one creature will mean the “secondary extinctions” of others.

Roopnarine and his colleagues have run enough simulations of food-web collapse that they’ve discovered a pattern. You can take away up to 40 percent of the life-forms in a system, and the number of secondary extinctions doesn’t increase significantly. “But there’s a critical interval after that where things happen rapidly—a threshold effect,” Roopnarine said. “The secondary extinction numbers rise dramatically.”

Imagine a world like the one we live in today, with a variety of creatures in many different environments. Let’s say we begin to chip away at one of those environments, like the American prairies. People clear grasslands, kill both predator and prey animals, and destroy insect pests. Still, the food web seems stable. Creatures and plants go extinct in the region, but there seems to be no ripple effect. And then, after centuries, we hit a tipping point. Forty percent of the nodes in the prairie food web have been knocked out. Suddenly, there are predators with very few prey. Catastrophic deaths among predators result: They are competing for scarce or no resources. And then a drought hits, killing the few remaining prairie grasses. Now our tiny herbivore population goes mostly extinct. We are left with few predators and virtually no prey. The already unstable food web falls apart, one death leading to another—and making the community more vulnerable to climate fluctuations.

“Don’t expect the unraveling to be linear,” Roopnarine warned. The deaths will be exponential. Once we hit the threshold, our food web has lost in the war of community selection. A new food web will rise up to take its place, turning the American prairie into a whole new world full of strange predators and grasses unlike any we’ve ever seen.

So the Permian extinction event yields a double lesson in survival. First, it offers compelling evidence that climate change caused by greenhouse gases can kill nearly every creature on the planet. Regardless of how that greenhouse scenario starts—whether it’s a massive volcano or an industrial revolution—climate change can kill more effectively than a meteorite impact. Of course, atmospheric changes were only the first phase in a problem that lasted 30 million years. One could argue that food-web collapse is really what makes the Permian mass extinction a “Great Dying.” The scourge started by Permian megavolcanoes echoed for millions of years, rending food web after food web until at last equilibrium was achieved in the Triassic period.

Still, there were survivors. Humans and many other mammals on Earth owe our existence to a bunch of piglike creatures with beaks who loved the sunny southern climate. That Lystrosaurus survived for millions of years (much longer than Homo sapiens has been around) proves that complex life can make it through even the most terrible disasters. These lumbering proto-mammals also left behind a few tips for what to do when we hit that wall of toxic air. By following in the lystrosaurs’ footsteps, mammals dodged the next major mass extinction—even though many dinosaurs didn’t.

“IT IS VERY hard to imagine what happened,” the paleontologist Jan Smit said. He was describing the minutes and days following the impact of a massive meteorite, possibly 10 kilometers wide, that slammed into the Earth roughly 65 million years ago. Smit is one of the scientists who first discovered evidence for this violent event back in the 1970s. Today many of his colleagues agree that it’s what caused the Cretaceous-Tertiary (K-T) mass extinction—or, as it’s better known, the extinction that ended the dinosaurs.

The Cretaceous Period (145.5 Million–65.5 Million Years Ago): Meteorite Impact

Though nearly everyone is familiar with the story, Smit finds himself constantly correcting people’s misconceptions about it. “It wasn’t like [the movie] Armageddon at all,” he chuckled. The Earth wasn’t wrapped in fire. There were no enormous dust storms choking the life out of the soon-to-be-extinct dinosaurs. Instead, Smit said, most of the molten splash-back from the hit would have been hurled right back into space. And that’s why it was so deadly.

The energy released by the meteorite slamming itself thirty meters deep into Mexico’s Yucatán Peninsula was enough to punch a hole in the atmosphere. As Smit put it, “For this kind of impact, blowing away the atmosphere is a piece of cake.” Tiny droplets of liquified rock and metals shot into space, quickly wreathing our stratosphere in a thick layer of extremely high clouds. The biggest problem was that the meteorite hit Earth in a particularly tender spot, geologically speaking. Beneath the Yucatán, Smit explained, “are three kilometers of limestone, dolomite, magnesium, and gypsum, and salt. Gypsum is about half sulfur. There aren’t that many areas in the world that contain that much sulfur.” Essentially, the meteorite vaporized a hidden cache of explosives and poisons, scattering them everywhere. Still, the mass death that swept the world afterward was not caused by acid rain or other poisons, according to Smit. Instead, it had to do with a peculiar property of vaporized sulfur: When reduced to tiny droplets in the upper atmosphere, the resulting cloud becomes highly reflective. “From space, the planet would have looked brilliantly white,” Smit speculated. For at least a month, Earth became a giant reflector, and little to no sunlight could have penetrated the sulfur-laced clouds. It would have been a very extreme version of what happened after the Permian megavolcano shot sulfur into the atmosphere and cooled the planet.

Below the cloud, it would have been dark for weeks or months. Death would have come quickly to anything that drew sustenance from sunlight, including most plants. Next to die would be plant-eaters whose food sources were gone, followed by starvation among the dinosaur predators at the top of the food chain. Imagine a near-instantaneous food-web collapse, a fast-motion version of the collapses that Roopnarine described as choking off early Triassic life over millions of years. What this means is that one of the planet’s most notorious disasters, complete with cinematic explosions, caused global mass extinction simply by shutting down photosynthesis.

Perhaps more than any of the other mass extinctions we’ve talked about so far, the K-T extinction dramatizes how mass death on Earth is tied to environmental changes. The area around the Yucatán would have been devastated after the meteorite hit. Toxic gas, fire, and extreme tidal waves would have sterilized the region around what is now called the Chicxulub crater. But even the death by darkness that followed would have been just the opening act. It would have taken centuries, and perhaps millennia, before the K-T event achieved full mass-extinction status. The dinosaurs did not die out during one long, sulfur-enhanced night. In fact, Smit underscored that the truly devastating effect of the sulfur cloud was most likely a temperature drop of 10 degrees Celsius that lasted for at least half a century, and probably a lot longer. The lush, green tropics of the Cretaceous cooled, ocean temperatures dropped, and animals who couldn’t migrate found themselves trapped in hostile ecosystems. The mass extinction took out as many as 76 percent of species, including all the non-avian dinosaurs. Meanwhile, a group of mouse-sized furry creatures we know today as mammals began to thrive and grow.

The flaming-ball-of-death controversy

The K-T mass extinction is the most recent one in Earth history, and the evidence it left behind is richer than what we’ve got for any comparable event. As a result, scientists who study the K-T have had to confront the full complexity of life when it collapses. Not surprisingly, this has led to some of the bitterest debates in paleontology.

Smit and a UC Berkeley colleague, the geologist Walter Alvarez, endured years of doubt and ridicule when they first began speculating that the dinosaurs’ demise began with a meteorite. Previously, paleontologists accounted for the mass extinction by suggesting everything from cosmic-ray bombardment to starvation. It took a while for the scientific community to accept the idea of a flaming ball from space. But Smit and Alvarez had pretty compelling evidence. Working on opposite sides of the globe—Smit in Spain, and Alvarez in the Americas—the two researchers uncovered physical remains of the impact and overturned the previous theories about why dinosaurs suddenly went extinct after ruling the planet for over 100 millennia. Alvarez worked with his Nobel Prize–winning physicist father, Luis Alvarez, and the two published a history-making paper in 1980 showing that rock layers at the K-T boundary contained a high concentration of iridium, a metal found almost exclusively in space. Meanwhile, Smit had discovered “spherules,” tiny balls of rock that had been heated up and then cooled down quickly, in the same layer all over the world. Smit published a paper about the spherules the same year Alvarez published his about what’s come to be known as the “iridium anomaly.” The one-two punch of these papers—chronicling metals from space and the remains of superheated rock scattered across the planet—suggested an event whose magnitude could easily have accounted for global mass death.

But the flaming-ball controversy is still far from over. In the late 1980s, Princeton geologist Gerta Keller began publishing papers questioning whether the meteorite impact actually had a global effect after all. She claimed she had a better explanation: megavolcanoes in India. And she spent the next two decades gathering evidence to prove her hypothesis, despite widespread scorn from the scientific community. UC Berkeley paleontologist Charles Marshall said that “nobody in the scientific community takes [Keller] seriously,” and Smit told the BBC that her ideas “are barely scientific.” Like Smit and Alvarez before her, Keller cheerfully met doubt with documentation.

When I spoke to Keller, she had recently returned from India, where she’d made a series of incredible discoveries. She and a group of local scientists managed to get samples three kilometers deep underground in a region called the Deccan Plateau, long known to be the site of an ancient megavolcano. The area has been off limits to scientists ever since India’s Oil and Natural Gas Corporation started drilling there. An outspoken person who clearly loves a good scientific fight, Keller put her considerable powers of persuasion into a campaign to gain access to what she suspected might reveal the truth about how the dinosaurs died. She finally got her wish and, joined by a group of Indian scientists, she found more than she’d ever hoped.

Using special tools that produce “cores,” cylindrical rock samples pulled up from deep underground using cannulated drills, Keller and her colleagues discovered that the Deccan Plateau was the result of at least four major eruptions following closely on each other. One eruption was so enormous that the team found a single uninterrupted lava flow stretching 1,500 kilometers from the volcanic vent all the way to the sea. But the most valuable discovery was the layers of sediment in between each lava flow—in those sediments, Keller and her colleagues found fossils that helped date the volcanic eruptions. Based on the evidence so far, it appears that the Deccan supervolcano began spewing lava and toxic gas about 67.4 million years ago—about 1.6 million years before the K-T boundary. The timing was right. And so were the deadly patterns of extinction she observed in those layers of sediment. After each lava flow, fewer and fewer animals were recovering from the devastation. “By the time the fourth flow came, nothing was left,” she said.

Keller believes that the flows may have come so rapidly that life nearby had no chance to recover—and that the toxins and carbon released by the explosions wound up killing off creatures across the globe with environmental changes similar to those at the end of the Permian. A runaway greenhouse effect, combined with acid rain and ocean dead zones, would have made the planet unlivable for the majority of its inhabitants. “That’s the likely killing mechanism,” Keller concluded matter-of-factly.

Who is right? It’s entirely possible that both Smit and Alvarez on one side and Keller on the other have identified causes of the K-T mass extinction. There are other theories, too. A fungal spike in the fossil record during the mass extinction has led at least one scientist to suggest that the dinosaurs died of fungal infections like the ones that are causing extinctions among amphibians and bats today. When the evidence in the geological record is relatively fresh, it becomes obvious that most mass extinctions on Earth have multiple causes. And as Keller’s work suggests, evidence gathered outside Europe and the Americas can offer a new perspective on old theories. We know that the bodies start piling up when environments change, but many events all over the world may have set those changes in motion.

The Late Triassic Period (220 Million–200 Million Years Ago): The Beginning of Our World

One of the difficulties in sorting out what happened to the dinosaurs has nothing to do with geological evidence and everything to do with human culture. Dinosaurs have been so widely misrepresented in pop culture about the prehistoric world that it’s hard for us to step back and appreciate this diverse array of creatures for what they actually were, and how they really died out.

To get the real story, we’ll return to the chaotic Triassic period that followed the Great Dying, when many species evolved and died out rapidly. Late in the Triassic, about 220 million years ago, dinosaurs began to evolve. At this time, they were, as Brown University geologist Jessica Whiteside put it, “about the size of German shepherds and not very diverse.” Their main competitors were the crurotarsans, those fierce carnivores that eventually evolved into crocodiles and alligators. How did a relatively small group of mini-dinos prevail against these toothy, occasionally armored beasts? “If you were in the Triassic, you would bet on crurotarsans,” Whiteside said. But surprisingly, most crurotarsans did not survive the mass extinction that ended the Triassic, leaving the dinosaurs to take over lands once dominated by their mega-gator counterparts.

Whiteside attributes this bizarre turn of events to one of the most stupendous underwater volcanoes in Earth history. Known as the Central Atlantic magmatic province (CAMP), the eruption started about 200 million years ago in a narrow body of water separating the eastern Americas from West Africa. (At that time, the two continents were still joined into the Pangaea supercontinent.) The lava flow from CAMP was tremendous. It forced the continental plates so far apart that an entire ocean grew between once-connected regions known today as Canada and Morocco.

If this eruption could create one of Earth’s biggest oceans, just imagine the high volumes of carbon, methane, and sulfur it was pumping into the water and the atmosphere. A superhot greenhouse gripped the planet as the Triassic wound to a murky close. Whiteside ticked off the deaths that followed: As the temperatures climbed higher, the world-spanning tropical forests of the Triassic dried out and succumbed to enormous wildfires. The burned remains of forests slipped into the oceans along with carbon-rich soil. The oceans became acidic, which led to anoxia and die-offs there. Coral reefs were the first to go, and their deaths set off a cascading effect where anything that fed higher in the food chain died too. It was the perfect storm for destroying food webs, starting in the oceans and creeping up onto a warming land whose trees were being eaten by fire. Once again, climate change was killing the world.

The rise (and fall and rise) of the dinosaurs’ world

There are many well-preserved plant fossils from this era, so it’s possible to visualize how the extreme greenhouse conditions changed the environment between the Triassic and the Jurassic. Jennifer McElwain, a paleobotanist at University College Dublin, has spent years studying this transition in Greenland, excavating everything from leaves and flowers to microscopic bits of pollen, to reconstruct the world where dinosaurs ultimately triumphed. Today, coastal Greenland is hard tundra that’s too cold for trees, but in the late Triassic and early Jurassic, it was full of lush vegetation. McElwain called it “a cross between New Zealand conifers and the Florida Everglades.” It was a world of “broad, meandering rivers” and “big, wide floodplains” bordering forests full of towering trees and stubby, thick-trunked plants called cycadeoids with palmlike fronds bursting from their tops. And then came CAMP, with its carbon emissions and rising global temperatures.

Tens of thousands of years of greenhouse conditions led to fire after fire. Ultimately, McElwain believes, the environment of diverse trees, shady forests, and thick vegetation was reduced to swamps full of ferns. “There would have been ferns as far as the eye can see, with hardly any trees, and lots of fire,” McElwain said. There was no complex, multitiered canopy in the forests, so the landscape would have been much brighter. But within another 100,000 years, the region went back to being conifer-dominated. What emerges from this fast-motion vision of ancient forests rising, burning, and rising again is something approaching the truth of where the dinosaurs began. They were among the only survivors of radical environmental changes that drove their competitors to extinction. Most crurotarsans were extinguished in the burned threads of food webs, but those small, early dinosaurs were able to spread out and adapt to the new environments and continents.

When forests at last returned to the land, dinosaurs evolved to be much larger. They diversified into armored herbivores like triceratops and plate-backed stegosaurus, sneaky scavengers, and predators like the 40-foot-long T. rex that we’ve seen in movies from the 1933 version of King Kong to Jurassic Park. Dinosaurs were as diverse as mammals are today, and their behavior probably varied a great deal from species to species. Many of them walked on two legs, with body postures similar to birds—their heads would have been thrust far forward, their spines nearly horizontal, and their tails held out stiffly behind them rather than dragging on the ground. Indeed, most paleontologists today accept that birds evolved from therapods, a group of bipedal, feathered dinosaurs that included the infamous T. rex. If you ever want to imagine what it would be like to face down a dinosaur, imagine a hulking, 40-foot-long crow whose beak has become a toothy mouth.

Recent evidence suggests that many dinosaurs weren’t feathered in quite the way birds are today. Most had dark gray or reddish proto-feathers (often called dinofuzz) that looked something like spiny down. Indeed, dinosaurs may have had proto-feathers for millions of years before birds evolved the ability to fly. Also like their bird relatives, many dinosaurs made nests and laid eggs. Though it’s hard to piece together how these different Cretaceous animals might have behaved, some paleontologists theorize that they may have been social, like birds, forming flocks and possibly mating for life.

What we do know is that when the catastrophes of the Cretaceous period hit, dinosaurs were in a position similar to their old rivals, the mega-gator crurotarsans. A lot of dinosaurs had evolved into specialists, and were therefore deeply connected to food webs that were all too easy to unravel with a few shifts in global temperature. This time, a group of mouse-like, furry animals called mammals—the descendants of the Permian survivor Lystrosaurus—were the survivors.

The Earth these mammals began to colonize with their strange paws and non-feathered faces had come into being through extremely complex events, whose true impact can only be measured in tens of millions of years. Environments had died and been reborn from the effects of liquid rock deep in the Earth and flaming balls from space; the mixture of gases in the atmosphere had been altered by microbes, mountains, and plants; temperatures had fluctuated between extreme icehouses and greenhouses; and the very shape of the planet’s landmasses and oceans had transformed quite radically dozens of times. If there had been a paleogeologist among the last of the dinosaurs, she could hardly have pinned the blame for her peers’ demise on any single factor. The entire ambiguous history of the planet would have had to stand trial for murdering brachiosaurus and letting a bunch of little monkeys take over.

The dinosaurs who survived

Adding to our paleogeologist dinosaur’s conundrum would be another issue, which is that the dinosaurs didn’t entirely die out. An evolutionary offshoot of therapods—birds—survived into the present to become one of the most successful animal classes on the planet. They are highly diverse, exhibit a wide variety of social behaviors, and undertake some of the most incredible migratory journeys of any creatures we know. Of course their dinosaur forebears are extinct, much the way humans’ forebears are. But the dinosaur evolutionary line appears to have continued unbroken.

Perhaps the single most common misconception about dinosaurs among humans is that these creatures and their world have disappeared. Like mammals, dinosaurs are survivors. But their children, who flash through the skies and leave us in awe, are so different from their ancestors that we find it hard to draw a connection between them. What we should ponder, as we move from geological history into the world where human evolution takes place, is whether our understanding of survival is as clouded as our understanding of dinosaurs.

The dinosaurs survived two mass extinctions, but the crows who like to hang out in the tree next to my house are nothing like those dog-sized dinosaurs who beat out the crurotarsans. In fact, it’s not entirely accurate to say the crurotarsans have been pushed off the environmental stage either. Are we not witnessing a strange tableau of survival whenever a bird alights on the head of a crocodile, bringing together the evolutionary offspring of Triassic and Jurassic? Instead of saying the dinosaurs died out, it might be more accurate to say that dinosaurs changed.

Can humans possibly expect to remain unchanged as we face the next mass extinction? History suggests that it’s unlikely. But if survival means that our species will evolve into creatures like ourselves, but with new abilities—like, say, flight—that’s not so bad. Some would even call it an improvement. Survival may be far weirder, and better, than we ever imagined.