“Welcome to the Black Triangle,” said paleobiologist Cindy Looy as our van slowed to a stop in the gentle hills of the northern Czech Republic, a few miles from the German and Polish borders. The Black Triangle gets its name from the coal burned by nearby power plants. Decades of acid rain generated by power-plant emissions have devastated the region’s ecosystems. Yet the treeless hills looked healthy and green.

I tried to hide my surprise. For months I’d been on the trail of the greatest natural disaster in Earth’s history. About 250 million years ago, at the end of the Permian period, something killed some 90 percent of the planet’s species. Less than 5 percent of the animal species in the seas survived. On land less than a third of the large animal species made it. Nearly all the trees died. Looy had told me that the Black Triangle was the best place today to see what the world would have looked like after the Permian extinction. This didn’t look like apocalypse.

We saw the first signs of death as we walked into the hills—hundreds of fallen timbers lay hidden in the undergrowth. A forest once grew here. Half a mile (0.8 kilometers) uphill we found the trunks of a stand of spruce, killed by acid rain. No birds called, no insects hummed. The only sound was the wind through the acid-tolerant weeds.

“The forest that grew here a few decades ago contained dozens of species of plants,” said Looy. “Now there are only a few grassy species.”

Looy picked up a spruce cone. Pollen from the trees around us might be preserved inside. She believes that the Permian extinction was caused by acid rain following a massive release of volcanic gases. She wants to compare tree pollen from a modern forest killed by acid rain with fossil pollen found in Permian rocks.

Like a homicide detective at a crime scene, Looy sealed the cone in a plastic bag for later lab work. “You could say we’re working on the greatest murder mystery of all time,” she said.

Looy is one of many scientists trying to identify the killer responsible for the largest of the many mass extinctions that have struck the planet. The most famous die-off ended the reign of the dinosaurs 65 million years ago between the Cretaceous and Tertiary periods. Most researchers consider that case closed. Rocks of that age contain traces of an asteroid that struck Earth, generating catastrophic events from global wildfires to climate change. But the Permian detectives are faced with a host of suspects and not enough evidence to convict any of them.

To understand this extinction, I wanted first to get a sense of its scale. That’s difficult—sediments containing fossils from the end of the Permian are rare and often inaccessible. One site that preserves the extinction’s victims lies about a half day’s drive inland from Cape Town, South Africa, in a scrubland known as the Karoo.

“The Karoo is the kind of place where people fall asleep at the wheel,” said Roger Smith, a paleontologist at the South African Museum, as we drove across the treeless land. “But it may be the best place to see the terrestrial realm’s transition from the Permian to the Triassic period.”

We ascended through sheep-ranching country toward the Lootsberg Pass. The rocks that surrounded us date from the late Permian. For every yard of altitude we gained, we traveled tens of thousands of years forward in time, heading for the Permian’s conclusion.

If we had driven here before the extinction, we would have seen animals as abundant and diverse as those of today’s Serengeti, except most would have belonged to a group known as synapsids. Often called mammal-like reptiles—they looked like a cross between a dog and a lizard—the synapsids were Earth’s first great dynasty of land vertebrates.

“We’ve found fossils of many kinds of synapsids in these rocks, especially tortoise-beaked dicynodonts, which likely lived in herds and browsed on vegetation along the riverbanks,” said Smith. “There were also a lot of smaller grazers and root grubbers, like Diictodon, a dachshund-shaped dicynodont that probably dug up roots and shoots. They were preyed upon by gorgonopsians—fleet-footed synapsid carnivores with needle-sharp teeth.”

The late Permian rocks we passed as we neared Lootsberg Pass capture the synapsids at the height of their reign. For more than 60 million years they were Earth’s dominant land vertebrates, occupying the same ecological niches as their successors, the dinosaurs.

Smith slowed at a switchback, rolled down the window, and pointed to a horizontally banded cliff. “See that road cut?” he asked. “That’s your Permo-Triassic transition zone. Brace yourself, you’re about to go through the extinction.” The fossils embedded in this road cut suggest that synapsids took a savage hit at the end of the Permian.

A synapsid known as Lystrosaurus appears in these sediments. Smith had a skull of the animal in his truck. Its flat face gave it the look of a bulldog with tusks. In the first few yards of the transition zone, only one or two Lystrosaurus fossils have been found scattered among all the diverse late Permian animals. Higher up, the diversity suddenly dwindles. Dozens of species of Permian synapsids disappear, leaving Lystrosaurus and a few others in early Triassic rocks. Animals were still abundant, but the community they formed was about as species rich as a cornfield.

Plants were also hit by the extinction. Evidence for the scale of damage to the world’s forests comes from the Italian Alps. I joined a research team led by Henk Visscher of the University of Utrecht at the Butterloch gorge, where exposed fossil beds cover the transition from the Permian to the Triassic. The beds lie high on a cliff, accessible only by climbing piles of debris. I anxiously followed veteran climber Mark Sephton up a slope of loose rocks to a ledge. Sephton used his hammer to chip bits of rock from the layers that chronicle the extinction. Each fragment contains microscopic fossils—pieces of plants and fungi. The lower layers, dating from prior to the extinction, contain lots of pollen, typical of a healthy conifer forest. But in rocks from the Permo-Triassic boundary the pollen is replaced by strands of fossilized fungi—as many as a million segments in some golf-ball-size rocks.

All that fungi in boundary rocks may represent an exploding population of scavengers feasting on an epic meal of dead trees. “We think it’s a wood-decaying fungus,” says Looy, who works with Visscher. “When a tree dies, it falls. As it decays, fungi grow into it from spores on the ground, decomposing it.”

Visscher and his colleagues have found elevated levels of fungal remains in Permo-Triassic rocks from all over the world. They call it a “fungal spike.” The same rocks yield few tree pollen grains. Visscher’s conclusion: Nearly all the world’s trees died en masse.

On the drive from Butterloch a team member handed me a soft, brown banana—a leftover from lunch. “This is how you can imagine the Permian extinction,” he said. “Rotting biomass.”

“It’s not easy to kill so many species,” says Doug Erwin, a Smithsonian Institution paleontologist. “It had to be something catastrophic.” Erwin and geologist Samuel Bowring of the Massachusetts Institute of Technology have dated volcanic ash in Chinese sediments laid down during the extinction. Bowring thinks the extinction took place in as little as 100,000 years—quicker than the click of a camera shutter on a geologic scale of time. Suspects must be capable of killing with staggering swiftness both on land and in the seas. As I spoke with some of the researchers on the killer’s trail, I learned how many suspects there are—and how difficult it is to develop a tight case.

An enormous asteroid impact is the prime suspect of Gregory Retallack, a geologist at the University of Oregon. The collision would have sent billions of particles into the atmosphere, he explains. They would have spread around the planet, then rained down on land and sea.

Retallack has discovered tiny quartz crystals marked with microscopic fractures in rocks from the time of the extinction in Australia and Antarctica. “You need staggering force, many times greater than a nuclear explosion, to create this shocked quartz,” said Retallack. “Only an impact could deform it this way.” A team of researchers recently found what may be that impact’s footprint buried below Australia—a 75-mile-wide (120-kilometer-wide) crater left by an asteroid more than three miles (4.8 kilometers) across.

I asked Retallack what an impact would be like if we had been standing a few hundred miles from ground zero. “You’d feel a shudder,” he replied. “Clouds of noxious gases would billow in and block out the sun for months. Temperatures would drop, and corrosive acid snow and rain would fall. After the clouds cleared, the atmosphere would be thick with carbon dioxide from fires and decaying matter. CO2 is a greenhouse gas; it would have contributed to global warming that lasted millions of years.”

The short-term effects alone—cold, darkness, and acid rain—would kill plants and photosynthetic plankton, the base of most food chains. Herbivores would starve, as would the carnivores that fed on the plant-eaters.

Other Permian detectives suspect the killer oozed up from the sea. For years scientists have known that the deep ocean lacked oxygen in the late Permian. But most life is concentrated in shallow water, in places like reefs. In 1996 English geologists Paul Wignall and Richard Twitchett of the University of Leeds reported the first evidence of oxygen depletion, or anoxia, in rocks that formed under shallow water at the time of the extinction.