The Nuvvuagittuq greenstone belt doesn't look like a battlefield. It lies in peaceful, roadless isolation along the northeastern edge of Hudson Bay in Canada, more than 20 miles from Inukjuak, the nearest human settlement. From the shoreline, the open ground swells into low hills, some covered by lichens, some scraped bare by Ice Age glaciers. The exposed rocks are beautiful in their stretched and folded complexity. Some are gray and black, shot through with light veins. Others are pinkish, sprinkled with garnets. For most of the year the only visitors here are caribou and mosquitoes.

But this tranquil site is indeed a battleground—a scientific one. For almost a decade rival teams of geologists have traveled to Inukjuak, where they have loaded canoes with camping gear and laboratory equipment and trekked along the coast of the bay to the belt itself. Their goal: to prove just how old the rocks are. One team, headed by University of Colorado geologist Stephen J. Mojzsis, is certain that the age is 3.8 billion years. That is pretty ancient, though not record setting.

Jonathan O'Neil, who leads the competing team at the University of Ottawa, argues that the Nuvvuagittuq rocks formed as long as 4.4 billion years ago. That would make them by far the oldest rocks ever found on Earth. And that is not the least of it. Rocks that old would tell us how the planet's surface formed out of its violent infancy and just how soon after that life emerged—a pivotal chapter in Earth's biography that has so far remained beyond reach.

The first half a billion years of Earth's history—from its formation 4.568 billion years ago to four billion years ago—was a time when water rained down to create the oceans, when the first dry land heaved above the surface of the sea to form continents. It was a time when comets and asteroids crashed into Earth and when a failed planet the size of Mars may have collided with ours, creating the moon from the wreckage. But geologists have very few clues about the timing of these events, such as a few specks of minerals that suggest oceans might have formed before the moon. They find themselves in much the same situation as biographers of ancient Greek philosophers, trying to squeeze as much meaning as they can from scraps of parchment and secondhand stories.

If O'Neil is right and Nuvvuagittuq's rocks are indeed 4.4 billion years old, they will read not like scraps but like entire books. Thousands of acres of the minerals are waiting to be studied, perhaps holding answers to long-running mysteries. Did plate tectonics start early on, or did Earth mature for hundreds of millions of years before the continents and ocean crust began moving around? What was the chemistry of the youngest oceans and the atmosphere? And how soon did life emerge after Earth formed?

If Mojzsis is right, the earliest chapter in Earth's history will remain shut for now. If O'Neil is right, the rocks of Nuvvuagittuq are among the most precious treasures of geology.

Imprisoned in Stone

Like the rocks that make up much of Earth's crust, the rocks at Nuvvuagittuq generally arose in one of two ways. In some cases, fine particles settled to the bottom of oceans, where they were gradually pressed into layers of sedimentary rock. In other cases, molten magma rose from Earth's mantle, cooling and crystallizing into igneous rock as it ascended.

Only tiny portions of ancient crust in places such as Nuvvuagittuq have remained intact, whereas the rest has vanished. Some rocks were slowly eroded by rain and wind and delivered back to the ocean for new sedimentation. Many others were carried back down under Earth's crust by tectonic plates sinking into the hot mantle, where the rocks melted, their original identity wiped out like an ice cube tossed into a warm pond. Their atoms mixed into the magma and rose again as fresh, young stone.

Rocks on the early Earth were also wiped out by giant asteroids that smashed into the planet and melted large fractions of the crust. About 4.4 billion years ago one collision—called the Giant Impact—hurled a huge amount of material into orbit, which became the moon. “The Giant Impact probably made a real mess of Earth,” says Richard W. Carlson of the Carnegie Institution for Science. “You would not have wanted to be here. You might want to have watched it from Venus.”

Given that so much ancient rock was destroyed in one way or another, it is not surprising that samples are rare. That is why the Nuvvuagittuq findings are so prized—and so hotly contested. Just a few other sites around the world have provided samples that are 3.8 billion years old. The oldest is from the tundra of the Northwest Territories, dating back 3.92 billion years.

The rarity of early rocks has driven geologists to look for other clues to what the planet was like in its first few hundred million years. Some of those clues have come from tiny crystals called zircons. These rugged, zirconium-based minerals will sometimes form in cooling magma. When the resulting rocks later erode away, some of the zircons may remain intact, even as they settle back on the ocean floor and are incorporated into younger sedimentary rocks.

The chemical bonds that make up zircons can trap radioactive atoms such as uranium. The decay of those atoms acts like a clock that geologists can use to measure the age of the zircons. The crystals also trap other chemicals, which can provide a few clues to what Earth was like when they formed. “Zircons are great because they're time capsules,” Mojzsis says.

In the outback of Australia, geologists have found sedimentary rocks that are sprinkled with immensely old zircons. Some of the zircons (but not the surrounding rock) date back as far as 4.4 billion years, making them the oldest traces of geologic history ever found. Scientists have squeezed remarkable information from these tiny gems since their discovery in 2001. Their structure suggests that the rock in which they originally formed solidified about four miles below the surface. Mojzsis and his colleagues have found chemical fingerprints of water in some of the Australian zircons, too.

The information that scientists can extract from sedimentary zircons is vastly better than nothing, but it is vastly less than they could get from the original rock in which the zircons grew. Rock contains many other minerals, which together can reveal far more about what Earth was like when it formed. “Unless you have the rocks, you don't have the complete story,” says Larry Heaman of the University of Alberta. Which brings us back to Nuvvuagittuq.

Sheer Luck

In the late 1990s the Quebec government launched a massive geologic expedition to make the first detailed maps of the northern reaches of the province. The region has an onionlike geology, with ancient cores of continental crust surrounded by layers of younger rock. Much of the rock proved to be around 2.8 billion years old. But Pierre Nadeau, then a Ph.D. candidate at Simon Fraser University in British Columbia, brought back a sample that dated back 3.8 billion years. By sheer luck, he had been sent to the Nuvvuagittuq greenstone belt. “Finding these rocks is like having a jewel dropped in my lap,” Nadeau's co-worker Ross Stevenson told the BBC in 2002 after they released their results.

Other geologists began to make the long journey to Nuvvuagittuq. Among those pilgrims was O'Neil, who was earning his Ph.D. at McGill University. He was struck by the chemical similarity between the Nuvvuagittuq rocks and 3.8-billion-year-old rocks in Greenland. Perhaps they belonged to the same ancient landmass.

To probe the chemistry, O'Neil teamed up with Carlson from the Carnegie Institution, who is an expert at precision measurements of ancient rocks. The only clear way to determine if a specific stone from Nuvvuagittuq is ancient or not is to date it. To do so, scientists count the levels of radioactive isotopes trapped inside a stone. Radioactive isotopes are variations of atoms that were part of the dusty cloud from which our solar system was born. They became incorporated into solidifying planets and meteorites, and when rocks on Earth crystallized, they became imprisoned inside. As time passed, the isotopes gradually broke down at a regular, clocklike pace. Measuring the levels remaining today reveals a rock's age.

In a Carnegie Institution lab, O'Neil and Carlson tallied up the concentrations of different isotopes. That is when they realized something was very strange about the Nuvvuagittuq samples. Among the isotopes was one known as neodymium 142. It forms from the breakdown of samarium 146. There is no natural samarium 146 left on Earth, because its half-life is short, by some estimates only 68 million years. “It's long gone,” Carlson says. “Samarium 146 was present when Earth formed because it was injected by the supernova that started the solar system. But then it decayed away within 500 million years.”

Carlson and his colleagues found that different Nuvvuagittuq rocks had different proportions of neodymium 142 and other neodymium isotopes. That variation could have come about only if the rocks formed at a time when there was still samarium 146 on Earth. O'Neil, Carlson and their colleagues compared the proportions to estimate just how long ago the rocks had formed. The number was one none of them had anticipated: 4.28 billion years. To their surprise, they had discovered the oldest rocks on Earth.

“This was totally not what we were expecting to find,” O'Neil says. He and his colleagues reported their discovery in 2008. Since then, they have analyzed other samples and now estimate that the Nuvvuagittuq rocks are as old as 4.4 billion years.

Old—or Oldest Ever?

O'Neil and his colleagues first announced their results at a geology conference in Vancouver. Mojzsis can still remember the shock he felt at the news: “My jaw drops. I look around, and people are stunned. I think, ‘This is peculiar.’”

Mojzsis had particular reason to be surprised. He was among the few geologists who had traveled to Nuvvuagittuq to follow up on Nadeau's research. Mojzsis and his colleagues had identified a vein of igneous rock that had thrust through the crust after the crust had formed. It turned out to contain zircons. Back home in Colorado, Mojzsis and his colleagues determined that the zircons were 3.75 billion years old—a result that squared nicely with Nadeau's original estimates of 3.8 billion years.

Now O'Neil was standing before Mojzsis and the rest of the scientific community, declaring that the Nuvvuagittuq rocks were half a billion years older.

Mojzsis's collaborator Bernard Bourdon of the École Normale Supérieure in Lyon, France, asked for some of O'Neil's samples and tested them again. The measurements of neodymium were correct. Still, “it didn't make sense to me,” Mojzsis says.

So, in 2011, Mojzsis and his students returned to Nuvvuagittuq to study the site further. They mapped the terrain and layers of rock around the samples O'Neil had dated. In the rocks that were reportedly 4.4 billion years old, they saw bright green bands of quartzite. That, Mojzsis decided, offered a way to test whether the Nuvvuagittuq rocks were the oldest on the planet.

Geologists have seen similar arrangements in much younger formations. They occur when underwater volcanoes spread molten rock across the ocean floor. Sometimes the volcanoes die down, and sediments from the land settle on top of the igneous rocks. Then the volcanoes rev up again, burying the sedimentary rock in a fresh layer of igneous rock.

If that was the case in Nuvvuagittuq, then the quartzite had come from sediments from an ancient landmass during one of those volcanic pauses. And if that quartzite had zircons, those zircons would have to be older than the surrounding volcanic rock because they had a much longer history.

“We were on our hands and knees crawling over many outcrops,” Mojzsis says. After days of hunting, they found two patches of quartzite with zircons—one of which yielded thousands of the tiny minerals. When they brought those zircons back to Colorado, Mojzsis found that they were 3.8 billion years old. That is precisely what they would not expect to find in rocks that were 4.4 billion years old.

Mojzsis's group also approached the question of Nuvvuagittuq's age from other scientific directions. They used another clock to date the rocks, for example, based on the decay of lutetium into hafnium. Once again they came up with 3.8 billion years.

All this evidence has led Mojzsis to a new narrative of Nuvvuagittuq. Around 4.4 billion years ago some molten rock rose toward Earth's surface and turned solid. As it crystallized, it captured some short-lived radioactive samarium 146 that still existed in the early Earth. But then the ancient crust was pulled back down into the mantle. The material heated up to the point where it was no longer rock, but all of it did not get mixed into the surrounding mantle. A bit of it remained a distinct blob with its own peculiar levels of neodymium. Finally, 600 million years later, volcanic activity pushed the material back to the surface, creating rock that incorporated some of the ancient blob, along with the blob's 4.4-billion-year-old signature.

“That melt itself can have a memory of a previous existence,” Mojzsis says. As a result, a rock that is only 3.8 billion years ago can appear to be 4.4 billion years old.

Zircon Mystery Explained

Mojzsis and his colleagues have been presenting these results at geologic conferences, sometimes in the same sessions where O'Neil is presenting the opposing view—that the rocks formed 4.4 billion years ago and simply have remained there in Earth's crust ever since. O'Neil's team has returned to Nuvvuagittuq, building up its collection of ancient rocks from 10 to about 50. None of the new data have clashed with the original estimation for the age of the site. O'Neil also rejects the evidence that Mojzsis and his colleagues have used to argue that Nuvvuagittuq is only 3.8 billion years old. “We have strong disagreement about the geology of the region,” O'Neil says.

Take the quartzite layer where Mojzsis found his zircons. In formations as old as those of Nuvvuagittuq, it is not simple to identify what kind of rock makes up a formation, because it has been deformed so much over billions of years. O'Neil does not think the quartzite band is quartzite at all. Instead, he argues, it is a vein of magma that pushed itself into the ancient rock 3.8 billion years ago. The age of its zircons thus has no bearing on the age of the surrounding rock. “There's nothing bizarre or unusual” about his own rocks, O'Neil says. “They're just really old.”

Heaman, himself an expert on old rocks, thinks that O'Neil and his colleagues have made a good case. “I think their evidence is compelling,” he says. “They've done their due diligence.” But Heaman also thinks that some uncertainty will endure until scientists can find another way to date the rocks. It is possible that a few minerals are lurking in the contested Nuvvuagittuq rocks that contain uranium and lead. That combination is the most reliable way to tell ancient time because scientists have vast experience with it. “If somebody were able to go out there and find the right material and get an old date, then the scientific community [would] be more accepting of the idea that there's some ancient crust exposed there,” Heaman says.

When Life Formed

If the Nuvvuagittuq rocks are indeed 4.4 billion years old, O'Neil believes they have the potential to open a wide window on the early Earth because they would have formed shortly after the Giant Impact. The Australian zircons were also forming at that time, several miles down into the mantle. But O'Neil argues that the Nuvvuagittuq rocks formed on the surface. “The geochemistry of these rocks really looks like an ocean floor,” he says.

If that is true, it confirms that Earth acquired an ocean not long after the Giant Impact. O'Neil also finds that the chemistry of the rocks is remarkably similar to seafloor rocks that formed much more recently. That would suggest that when the world's oceans first arose, they were not drastically different than they are today. O'Neil even believes that the rocks show signs of plate tectonics, suggesting that this process started very early in the planet's history.

There is an even more exciting prospect if the Nuvvuagittuq rocks formed on the ocean floor 4.4 billion years ago: they could shed light on the origin of life. Right now the fossil trail runs cold at 3.5 billion years ago. In rocks younger than that, scientists find preserved bacteria. In rocks older than that, they have found none.

But fossils are not the only traces that life can leave behind. As bacteria feed on carbon, they can alter the balance of carbon isotopes in their environment, and that imbalance can be preserved in rocks that form at the time. Some researchers have claimed that the 3.8-billion-year-old rocks from Greenland carry that imbalance, a signature of life.

That still leaves no evidence of life for the first 700 million years of Earth's existence. Scientists thus cannot say whether life got a quick start on Earth shortly after the planet formed or if it was delayed for hundreds of millions of years. They also have yet to figure out where life began on the planet. Some researchers have suggested that biological molecules emerged in deserts or tidal pools. Others have contended that deep-sea hydrothermal vents were the original nurseries.

If the Nuvvuagittuq rocks formed on the ocean floor 4.4 billion years ago, they are the perfect material to study to tackle these big questions. O'Neil hopes to collaborate with researchers to see whether the rocks could have formed at hydrothermal vents. “We cannot ignore these rocks. It's the ideal place where life could have formed,” he says.

Finding the earliest traces of life is an obsession of Mojzsis, too, but he will not be looking in Nuvvuagittuq for them. “I'll spend the rest of my career pursuing that jabberwocky,” he says.

Mojzsis does see an important benefit arising from his disagreement with O'Neil and others, however. As they spar, they are developing better methods for dating old rocks in general. “It's really an amazing debate,” Mojzsis says. Future generations of geologists who venture into the remote corners of the world and bring back enigmatic samples will be able to use those methods to finally lift the curtain back on the early Earth. And on that point, at least, O'Neil and Mojzsis agree. “All these small enclaves of old rocks are probably all over the place,” O'Neil says. “They're just really easy to miss.”