Inside the conference center’s auditorium in Nantes, the scientists were struggling. Not with the new findings on the inner workings of the Earth. Or with the elegant math that described them. But with the tiny font size of the print on the screen at the front of the cavernous room. Some of the conference participants were covertly bringing out binoculars. Others were taking photographs with their iPhones and then zooming in on the information using their touchscreens. This is the mind of the scientist: If there are barriers to getting data, you figure out your own way of leaping over them. It was that instinct that drove the scientific community after Brunhes’s paper on reversals in 1906. Skepticism reigned. Had the planet’s magnetic field really reversed direction? And if so, how can we be sure?
Theoretical questions hinged on whether such a dramatic perturbation of the poles was even possible. If it was, what was the mechanism? What was the purpose of a reversal? Could it have happened more than once? Could it even be a recurring feature of the planet’s magnetic landscape?
The practical questions were no less pressing. What if Brunhes’s terracotta didn’t mean what he thought it meant? Modern inquisitors might question whether Brunhes’s finding meant that the poles were stable but the European continent had rotated 180 degrees on the Earth’s surface. But at the beginning of the twentieth century, most geologists thought the continents were fixed, so they were looking for a different explanation for Brunhes’s discovery. What if scientists had a faulty understanding of the way the magnetic memory of rocks worked? What if Brunhes’s chunk of rock had merely been hit by lightning and that had shifted its dip? What if rocks could change their magnetic memory on their own, without any influence from the Earth’s poles?
It was this latter issue that dominated magnetic studies for decades after Brunhes’s paper. If a rock could spontaneously change its record of magnetic coordinates, then the whole idea that the field had reversed would be in question, as would many other aspects of rock magnetism. Before that question could be settled, other findings supporting Brunhes began trickling in from other parts of the world. Scientists had begun their meticulous job of collecting new data points. The most compelling findings were in a modest three-page paper in Proceedings of the Imperial Academy by the Japanese geologist Motonori Matuyama in 1929. A professor at Kyoto University, Matuyama also studied at the University of Chicago.
Japan is a global volcano hot spot. It sits at the juncture of four tectonic plates along what’s known as the Pacific Ring of Fire. Recently analyzed undersea marine sediments show that volcanoes have been active in the area for 10 million years and that the past 2 million years have been a period of extreme volcanic activity. In other words, the Japanese are keenly interested in what happens under the Earth’s crust, and that island nation has produced some of the world’s most eminent experts in all things inner-Earth. Including lava.
In theory, as Melloni and then Brunhes had reasoned, lava would take on a record of the intensity and direction of the magnetic field from the time and place where it cooled, in effect becoming a sort of sophisticated fossil compass showing inclination, declination, and field strength. So in 1926, Matuyama went looking for ancient basalt in a cave celebrated for that type of rock in Japan. He carefully measured its magnetic coordinates while it was in the cave and then took a sample for later examination. Its field pointed in the exact opposite direction from where the Earth’s field pointed in 1926. Matuyama then embarked on a systematic examination of basalts that had spewed forth from volcanoes over many millions of years in Japan, Korea, and Northeast China, then called Manchuria. His findings were that some of the rocks’ fields were aligned with today’s north and some of them were aligned with the south. Few were aligned anywhere in between.
The south-aligned rocks were from different geological periods: Some were Miocene, meaning they were as much as 23 million years old. Some were from the Quaternary, making them as much as 2.6 million years old. His conclusions were staggering. Not only was there more proof that the poles had reversed, but now there was evidence of more than one reversal. Each appeared to have lasted for a long period. Even more astonishing, Matuyama could put rough dates on when some of those reversals had taken place. All of a sudden, it looked as though geologists might be able to make a clock going back over the Earth’s distant past, describing where the poles had been during each era. It was a new way of seeing the planet, akin to the first maps Edmond Halley had produced showing the wavy contours of declination across the Atlantic Ocean.
One hitch in this analysis was the possibility that rocks could change their own magnetic memory. Through the 1930s and 1940s, this was an intractable problem. Rocks were tricky. Even iron, the standard material for compass needles, could lose its magnetic sensitivity. That’s why sailors in centuries past had carried a lodestone as a “keeper” to keep the iron magnetized. They would stroke it across the compass’s needle every now and again to remagnetize the iron. Geophysicists handled the confusion over spontaneous rock reversals by ignoring pole reversals until they could get more data, a phenomenon the American geophysicist Allan Cox and his colleagues later put down to “the embarrassing lack, even at so late a date, of a theory adequate to account for the present geomagnetic field, let alone reversed magnetic fields which may or may not have existed earlier in the earth’s history.”
One clue to the solution came from the work of Louis Néel, once offered a job at the observatory in Clermont-Ferrand, where Brunhes had worked. Néel eventually went to Grenoble, where he set up that university’s world-famous geophysics program. That’s where Philippe Cardin worked, who gave the public lecture at the conference in Nantes. But in 1931 as Néel was considering a position in Clermont-Ferrand, Brunhes’s legacy in magnetism was on his mind. So was the mystery of precisely how and why a rock retained its magnetic memory. Taking a page out of quantum mechanics, Néel began to question whether every molecule in a substance was magnetized in precisely the same way. What if there were differences? In a series of discoveries that won him the Nobel Prize in 1970, Néel found that there were. In the years following the Second World War, he advanced the concept of ferromagnetism, and, in 1949, discovered ferrimagnetism, which is a related but slightly different phenomenon. In doing so, it’s said that Néel took the magic out of magnetism, because he could finally explain why a material could hold its magnetic charge.
The reason goes back to the unpaired spinning electron.
The electron’s motion makes a tiny circulating current. That, in turn, creates a magnetic field with two poles. In most materials that make up our universe, the magnetic fields of unpaired spinning electrons cancel each other out, so the material doesn’t hold a magnetic charge. It’s a nano zero-sum game. It’s why so few materials retain magnetization over time. Sometimes, though, when electrons are unpaired, they don’t cancel out but reinforce each other by lining up. It’s the opposite of what you’d expect, and that makes these substances odd. When the electrons line up rather than neutralizing each other, the material ends up being magnetized, either for a while or, sometimes, permanently—as long as it doesn’t get heated up past its Curie point. The permanent type is called remanent magnetism, after the Latin word for “remaining.” This can get a lot more complicated. Even the Encyclopedia of Geomagnetism and Paleomagnetism says there are too many types of remanent magnetism for it to review. The type we’re interested in here is the magnetism a rock acquires under natural conditions as it cools. It’s commonly called natural remanent magnetism. In the days since Brunhes, scientists have learned how to strip away from rocks little bits of magnetism that came from other outside sources in order to reveal natural remanent magnetism. Carlo Laj, a French geophysicist, went back to Pont Farin and redid Brunhes’s experiments after stripping away extraneous magnetic influences. His paper, published in 2002, showed that Brunhes’s findings were absolutely correct.
Néel found that there are crucial differences in how the electrons decide to arrange themselves in order to enhance their magnetic fields. The differences determine the tenacity of a material’s magnetic field. In some substances the orbitals where electrons live overlap across atoms. And in some of those cases, when orbitals overlap, the electrons in adjacent atoms are then forced to line up in the same direction. That magnifies the magnetic pull of a material. When that happens, the material is called “ferromagnetic.” The common ferromagnets are iron, nickel, and cobalt and some compounds they are in. The name comes from the Latin word for iron: ferrum. The iron in a compass needle is a ferromagnet.
There’s a catch, though. The enhanced magnetic field within the groups of atoms or molecules is confined to domains, or neighborhoods, within a material. And while the field is strong within that neighborhood, it can be offset by an opposite field in the next neighborhood. So the material as a whole is not necessarily magnetized. That’s why your car keys aren’t magnetized, as a rule. But ferromagnetic materials can be magnetized if you put them in the presence of a strong magnet. The magnet’s power can make the unpaired electrons spin in the same direction, no matter which domain they’re in. That’s how the keepers kept the compass needles working. Stroking the needle with the lodestone made the domains line up. Ferromagnets can keep this strong magnetic charge for a time, but not permanently.
And then there’s more permanent magnetism, like the lodestone. Sometimes the way the atoms line up means that the opposite spins of the electrons don’t fully cancel each other out. Instead, they line up in, you could say, teams of uneven sizes, in alternating rows. One team is spinning in one direction. The other team spins in the opposite. The direction of spin of the bigger team wins out for the material as a whole and the material locks in on its magnetic direction. This is called ferrimagnetism. This arrangement of spins is far more stable than that of the ferromagnetic materials. It’s less apt to be changed or lost. The best example on Earth is the lodestone, the same magnetite that Homer wrote about and that Gilbert experimented on and that first sparked human investigation of magnetism. Magnetite is a type of iron oxide made up of a molecule of three iron atoms—each with four unpaired spinning electrons—connected by four oxygen atoms. It can hold its magnetic charge for millions of years, unless it is heated up past its Curie point. Some rare earth elements are also ferrimagnetic.
Once Néel worked out the difference between the ferros and ferris—in my giddier moments I call them the pharaohs and the fairies—he looked at fine-grained volcanic rocks and found that they often contained enough of the right size of ferrimagnetic grains of iron oxides to bind their magnetic memory for millions of years, unless heated. The same phenomenon holds true in some types of sedimentary rocks, like iron-rich terracottas.
During the same period after the Second World War, John Graham, a keen young geology graduate student at the Carnegie Institute of Washington, DC, launched a series of expeditions to test the magnetism of rocks across the United States. Pictures from the era show a truck made into a roving rock-sampling lab, complete with a spare tire strapped to the hood. Disconcertingly, in light of the European and Asian findings, Graham found rocks in the same layer that seemed to be pointing magnetically in different directions. Could they have reversed themselves spontaneously?
He turned to Néel. Néel, a theoretician, predicted that it was possible and set out several rare scenarios in which it could happen. Supporting his theory, Japanese scientists showed in the laboratory that some lavas from Mount Haruna were susceptible to reversing their own fields, as long as they were cooled at a specific rate and contained a specific chemical composition. Yet Jan Hospers, a graduate student at Cambridge, who examined layers of lava flows from the highly volcanic Iceland, found clear evidence of not just one or two but three reversals of the whole field over time. He concluded in 1951 that “the earth’s magnetization has suffered repeated reversals, and that rock magnetism can be used for geological correlation. . . .”
Back and forth it went. The Earth’s magnetic rock record was reliable. The Earth’s magnetic rock record was tainted. And this was the only tool the geophysicists knew of to determine whether the poles had reversed. They were flummoxed. Dueling theories about how rocks’ orientations were laid down continued for years, even as evidence mounted from around the world that the Earth’s magnetic field had reversed many times. By 1963, a poll of twenty-eight leading paleomagnetic researchers attending a meeting in Munich found only half could support the idea that the poles had switched places, but each of them believed that some rocks could switch their own magnetic field independently.
A year later, in a development that marked a shift in the research away from Europe and toward the Americas, there was more evidence than ever that reversals were part of the planet’s inner makeup. In 1964, Allan Cox, Richard Doell, and Brent Dalrymple of the US Geological Survey in Menlo Park, California, published their landmark paper “Reversals of the Earth’s Magnetic Field” in Science. They had gone looking for the ultimate proof that the rocks could tell the story of what had gone on in the Earth’s core. That meant they needed rocks showing magnetic memory from the same time periods at different places across the Earth’s crust. And that meant knowing with a high degree of precision how old the rocks were. They used a new technique involving the radioactive decay of potassium-40 to argon-40. Today, it’s known as K-Ar dating, after the chemical symbols of the elements. (K is for potassium and Ar is for argon.) By determining how much argon-40 is in a sample compared to the radioactive potassium-40 it would have started with, you can tell how long it’s been since the rock crystallized.
With the help of scientists around the world, they collected sixty-four samples of volcanic rocks from North America, including Hawaii; Europe; and Africa, and analyzed their ages using the K-Ar dating method. At the same time, they looked at the rocks’ magnetic signals. The Geological Survey gave them a small tar-paper shack, where they could work out what it all meant. Their results produced nothing less than the first global magnetic calendar going back 4 million years, describing epochs in the Earth’s history when the poles had been where they are today and others when they had been on opposite sides of the planet. Those early findings already showed some of the peculiarities that we now know characterize reversals: They last a long time in geological terms, and long enough to be captured in the rock record; they are of irregular length; sometimes, the poles try to reverse but fail.
Most intriguingly, Cox’s group pinpointed the last time the poles reversed to 780,000 years ago. That’s before our species, modern humans, was on the Earth. And they decided to name the current epoch in honor of Brunhes. The epoch that preceded this one is called the Matuyama. Others are named after Gauss and Gilbert. Nearly sixty years after his paper on the terracottas of Pont Farin, Brunhes’s contribution to the discipline of geomagnetism was formally acknowledged.
While Cox, Doell, and Dalrymple dealt at length in their paper with the idea that rocks could spontaneously reverse their polarity, they concluded that such events were rare. In fact, they were so rare that they did not negate the robust evidence from around the world supporting the switching of the poles.
The poles really do switch places sometimes.
At last, another element of the Earth’s turbulent past was starting to come into focus. Now, what the magnetic researchers wanted was to piece together reversals going even further back in time, hopefully back to the birth of the planet’s own magnetic field when the Earth was about 1 billion years old, or perhaps even younger. How often had the poles flipped, and were they conspiring to do so again?