One evening earlier this spring, German naturalist Norbert Jürgens strayed from his expedition in the Namib Desert. He walked away from his campsite beside Leopard Rock, a huge pile of schist slabs stacked like left-over roofing tiles, and into a vast plain ringed with red-burnished hills. He had 20 minutes of light left before sunset, and he intended to use them.

This next part may sound like a reenactment from a nature documentary, but trust me: This is how it went down.

Off by himself, Jürgens dropped down to his knees. He sank his well-tanned arms in the sand up to the elbows. As he rooted around, he told me later, he had a revelation.

At the time, I was watching from the top of Leopard Rock, which offered a bird’s-eye view of both Jürgens and his expedition’s quarry. Across the plain, seemingly stamped into its dry, stubbly grass, were circles of bare ground, each about the size of an aboveground pool. Jürgens, a professor at the University of Hamburg, was digging—and pondering—in one of these bare patches.

The patches are Namibia’s enigmatic fairy circles, and for decades they have drawn visitors, including our convoy, into the desert. In recent years, Jürgens and other researchers have argued bitterly over the how and why of fairy circles, disagreeing over data and theory in person and across the pages of the world’s preeminent journals.

This is more than an academic dispute over a tourist attraction, however. Fairy circles are a test case in the emerging field of biological-pattern analysis, where they may offer an encrypted message about the future of desert ecosystems—and the humans who hope to survive in them.

Fairy circles are found in parched, sandy soils on the Atlantic side of southern Africa, along a thin north-south strip that runs well over a thousand miles from South Africa’s Richtersveld mountains into southern Angola.

The smallest fairy circles are about five feet in diameter, and the further north you go, the bigger they get; the largest circles, in Angola, can sprawl across 130 feet. A single circle can persist for at least 75 years—maybe for centuries. Among their many peculiar qualities is a spooky low-level magnetism: A magnet dragged across the inside of a circle picks up far more soil than it does outside its boundary.

Since the 1970s, scientists have spitballed theories about the origin of fairy circles. The bare patches could be caused by chemical compounds emitted by Euphorbia damarana, a toxic bush. Or they could be the feeding grounds of a ravenous termite called Hodotermes mossambicus. Maybe they’re fossil termite nests; maybe the vegetation within them was killed by naturally-occurring radiation, or hydrocarbon seepage. And, I mean, there are always UFOs.

Jürgens met his first fairy circles in 1980, as a graduate student who had traveled to South Africa’s Richtersveld to study desert-adapted “living stone” plants. But the mystery didn’t really hook him until two decades later.

In 2000, he began working as the scientific coordinator for BIOTA, a sprawling network of environmental measuring stations across southern Africa. (That led to SASSCAL, the Southern African Science Service Centre for Climate Change and Adaptive Land Management, a massive EU-funded project that, along with the University of Hamburg, supports his field work.)

Jürgens began watching the fairy circles year after year, islands in the flood of data coming back from the environmental observatories. His curiosity grew. In the mid-2000s, he sampled sites throughout the fairy circles’ range, hoping to make his own sense of the issue. Jürgens excavated circle after circle, measuring soil moisture and other variables, and wrote up a table of every plant and animal he had ever found around one. “When I put them all together,” he said, “only one column was always full.”

Jürgens has spent the last five years arguing adamantly that the real culprit is not the ravenous Hodotermes termite but another termite species, a tiny, blink-and-miss-it sand dweller called Psammotermes allocerus. That assertion has placed him at the center of a controversy that still hasn’t died down.

This year, I joined him on one of his regular desert sojourns. He’s still collecting data and testing his theory, and as I watched him that evening, digging while the sun dipped behind the hills, he might have unraveled the mystery a little bit more.

Termites, you have to understand, are a very big deal. If you scooped up every animal across the global tropics and piled them on a scale, termites would account for roughly 10 percent of the total weight. Africa alone hosts over 1,000 termite species, including the genus Macrotermes, which builds cathedral-tower mounds that can stand taller than elephants.

This immense biological footprint doesn’t just interest scientists. Various indigenous groups in Africa make use of termites, according to a recent ethnobiological study. Termites, which are rich in protein, are often eaten; in parts of Rwanda, living termites are used to suture wounds; in Benin, fungus from termite nests is sometimes mixed with honey to make a memory tonic; in Côte d’Ivoire, the Agni people bury their loved ones in abandoned termite mounds.

But Psammotermes allocerus is unsung and understudied, and when I first met Jürgens, I had very little idea of what to expect.

Our expedition numbered 13. In addition to Jürgens and myself, our group included three members of Jürgens’ lab, a Namibian research assistant named Vilho Mtuleni who goes by the nickname Snake, and a gaggle of University of Hamburg masters’ students.

We gathered at the end of February in Namibia’s capital Windhoek, a city still shaking off the legacy of colonial rule by Germany and, later, apartheid South Africa.

Fairy circles in the Namib Desert (Theo Allofs / Getty)

After buying supplies, we left in a caravan of 4x4s, kicking up a cloud of dust. As we drove west from central Namibia toward the Atlantic, the land got drier and drier.

When I wasn’t busy wrestling with my truck’s manual transmission, I started noticing things. Dry river crossings. Woven bird nests. And big red Macrotermes mounds, sticking up like hitchhiker’s thumbs.

That night, we camped next to the sandy, empty bed of the Aba Huab River. The next day, Jürgens led our convoy along the riverbed, past an ornery desert elephant, and into the invitingly wide terrain, where distant hills just beg to be climbed.

You’ve seen the Namib Desert if you’ve seen Mad Max: Fury Road, where it stands in as an apocalyptic vision of the future. But the Namib has looked like it does today for at least 55 million years, maybe longer, and for all of those eons, its plants and animals have been evolving in the face of impossibility.

We kept driving. And suddenly, there they were. Jürgens, at the head of the caravan, pulled over and the rest of us followed suit. He walked out into a field at the foot of a hill, me racing to follow him and his flock of students straggling behind. “Your first fairy circle,” Jürgens said to me when we assembled. We stood in our own circle near the taller grasses at the edge of a bare patch, and Jürgens surveyed the patch like an astronomer dissecting a spiral galaxy. He pointed out the clearest fingerprint of his sand termites: a clump of grass coated in tubes of sand. The termites build the tubes around the stems they eat, he said, crumbling some of the sheeting in his fingers.

Then he toured us through his version of a fairy circle’s anatomy, a theory he first published in the journal Science in 2013. At the center is the bare patch. Surrounding that is what he calls the perennial belt, a thin ring of the tallest, healthiest grass in sight. Then there’s the halo, a thicker annulus of still-healthy grass. Between the circles are rings of sparser grass, which he calls the matrix. Jürgens looked approvingly at the sketch I was making in my journal.

His Science paper argued that the humble Psammotermes builds fairy circles to alter its environment, not unlike how beavers build dams. In this view, each circle functions as a cistern—a hydrological savings account.

Here’s the idea. When rain finally falls on these sandy patches of the Namib, the sun and the roots of plants suck the water right back out of the ground. But within the bare patch, the sand termites chew through all the plant roots, blocking growth. The absence of plants lets rainfall percolate further into the soil, to a deeper layer where it lingers in the pores between grains. According to Jürgens' measurements, the soil two feet below a fairy circle is wetter than the soil outside, roughly 5 percent water by volume. He theorizes that here, the termites can drink their fill year-round, using specialized mouthparts to slurp water from the sand.

This strategy may help the termites maintain a storehouse of food, too. By letting the grass in the perennial belt and the halo tap into their water supplies, he argues, the termites ensure that some plants persist even during drought. As for the patches’ eerie magnetism, that may arise when wind sifts through them, carrying smaller particles away and leaving a higher proportion of iron-rich, naturally magnetic grains behind.

As Jürgens has gathered evidence for his theory, he’s recorded dozens of other animals reaping the benefits of the water and the vegetation. Ants and gerbils live belowground in the bare patches, alongside the termites. Antelopes often rest there. Aardvarks dig down to eat the termites.

After leaving those first circles, we drove back through the riverbed to camp another night. As our cars slipped and slid across the sand, questions rattled around in my head. So many of the early theories explained fairy circles as scars, the result of poisoning or overharvesting. Wouldn’t it be poetic if these apparent dead spots turned out to be bulwarks of life?

There is, of course, another theory about the fairy circles. Its pedigree goes back to 1952, when British polymath Alan Turing sketched out a mathematical framework to explain patterns in nature. Just a few equations, he showed, could produce designs like the whorls of plant leaves, the stripes on a zebra, or the spots on a baby leopard.

Right around when Jürgens put forth his sand termite hypothesis, a Turing-style and termite-free explanation of fairy circles began gaining its own steam in the scientific journals. It’s called self-organization. Instead of a leopard coat, consider a lone clump of grass growing in a sandy stretch of the Namib.

There’s shade next to that clump, and more water because the grass exhales some moisture. Encouraged, another clump or two of grass grows right there. At the same time, the roots from these grasses tunnel out in all directions, trying to drink everyone else’s milkshake. Between healthy patches, the prospecting roots suck all the water away, creating an orderly patchwork of barren spots that, in computer simulations, looks an awful lot like fairy circles.

“The fairy circles are driven by water availability,” says Stephan Getzin at the Helmholtz Center for Environmental Research in Leipzig, who developed this interpretation of fairy circles along with Israeli ecologist Ehud Meron and others.

According to their models, fairy circles exist on a larger continuum. As you crank down the rainfall over a lawn of grass, you get bare patches like fairy circles, then labyrinthine patterns, then dwindling grass clumps. If you’re managing this land, the fairy circles could be a wake-up call: Act soon, because the last stage of this sequence is naked dunes.

The Namibian circles, in this view, are just snapshots along a gradient of desertification—one that unfolds not over time but in space, from the wetter east to the drier west.

To Getzin and his allies, sand termites are just one of many creatures that sometimes frequent fairy circles. “Norbert Jürgens is by training a specialized botanist,” Getzin said. “If he’s able to find these termites, then qualified or trained entomologists should be able to do that, too.”

But Walter Tschinkel, a renowned expert on social insects who has studied fairy circles for over a decade, told me that in his excavations, sand termites just aren’t that common. Another entomologist, Eugene Marais, has also failed to spot sand termites in fairy circles.

To Jürgens, “this non-seeing, non-finding of termites is a terrible thing.” He said they are easy to miss, and that they need to be excavated carefully with a leaf blower. To break this particular deadlock, he has tasked one of his Ph.D. students with creating a genetic test for sand termite DNA in sand scooped up from a fairy circle. But that test is still under development.

Meanwhile, the self-organization camp has been busy building its own case. They point to the ordered spacing of fairy circles as evidence that Turing’s math, not messy little bugs, are behind the phenomenon. They’ve also shown that water really can move across large horizontal distances in fairy-circle landscapes, a necessary step in proving their theory of long-range competition among clumps of grasses.

And in 2016, Getzin and his colleagues announced the discovery of another fairy circle–esque, ostensibly termite-free pattern about 6,000 miles away from the Namib, in the Australian outback. That announcement spawned its own ongoing sub-feud, in which Australian scientists argued that the bare patches in the outback had been previously documented and are made by termites, a charge that the self-organizers then rebutted ... and so on.

To its advocates, self-organization provides a generalized mechanism for regular patterns in vegetation all over the world. It’s a comprehensive, if less romantic, solution to the fairy-circle puzzle. “There’s such a thing in science as a story that’s too good to give up, even when all the evidence is contrary,” Tschinkel said. “I have a feeling the termite theory of fairy circles is one of them.”

So far, our caravan had made slow progress toward the actual research site. First, a stomach bug worked its way through the Hamburg students. Then a wheel loosened on Jürgens’ car. But as the sun went down on the third day of the expedition, we squinted, bounced, and jangled our way into Giribesvlakte, an enormous plain of dry, cattle-grazed grass so full of fairy circles that it looked like someone had taken a cookie cutter to it.

We set up camp. Then Jürgens, displaying his by now un-ignorable habit of starting sentences with the words “I propose,” proposed the team split itself into several groups.

Kristin Lemke, an undergraduate writing her thesis, needed help collecting data on ant nests within Giribesvlakte’s fairy circles. And Felicitas Gunter, working on her Ph.D., set out to excavate an entire corridor between two adjacent circles, carefully brushing away the sand to map out the termite tunnels.

A rhythm developed along with the permanent sensation of sand grains between our teeth. After breakfast, the team would set out. Around lunchtime, the back gate of one of the 4x4s became a buffet table, offering sausages, cheese, and crackers. After another sweaty hour or two, we would retreat to our camp at Leopard Rock, chasing fragments of shade by hugging the formation’s contours. Then, as the day began to cool, we would head back out, into the fierce wind sucked between the baking Namib and the cool Benguela current off the coast. Soon enough, the sun would drop toward the hills, and we would switch to long pants and sweatshirts.

Alicia DeWitt

One afternoon I sat down with Snake, the team’s only Namibian, as he supervised the excavation group. As we chatted, he pointed out a figure coming up the road on a donkey, leading a herd of cattle past our trucks. I asked if we could talk to the man about the circles. “We can try,” Snake said. He flagged down the herder, who stopped to wait for us while his cows strayed ahead.

Up close, I realized the herder was just a teenager, a boy from the Himba tribe. He wore sandals, salmon-colored shorts, and a pink, American-style hoodie under his grey jacket. Snake ran through his Rolodex of languages, searching for a match. Afrikaans? English? Oshiwambo? No hits. So Snake spoke Herero, the boy answering in Himba. “He just says that it’s a formation from above,” Snake said. The boy referred to the fairy circles as okarupare, a sort of antechamber or group meeting place for men in Himba compounds, Snake explained.

The next day the boy came back to the research site, this time bringing two friends. They were skeptical that a little insect could be responsible for the entire pattern. The circles were gifts from heaven, they said. But Snake showed them a small, unearthed nest, and explained a colony’s hierarchy by comparing it to Himba tribal structure. In this nest there were workers, he said, and in another nest somewhere, winged chiefs and queens were biding their time until the next rains. When the vegetation surged back to life, perhaps those would-be royals would create their own fairy circle or take over an abandoned one. The boys listened, engaged.

There was another mystery at Giribesvlakte, nested like a Russian doll inside the larger fairy circle problem. “Nobody outside our team knows about these,” Jürgens said one afternoon, when our caravan arrived a cluster of four fairy circles that were much larger than those around them. These megacircles, as Jürgens called them, occurred only in this area and only in clusters, always with an island of grass growing in their centers. Other groups, unaware of this particular observation, haven’t yet tackled these circles with their models.

“I propose we rest 10 minutes and inspect the circles,” he said. “Whoever finds the solution wins a glass of wine in the evening.”

Last year, seemingly out of nowhere, a third group joined the fairy-circle fray. In the journal Nature, a team including Princeton ecologists Rob Pringle and Corina Tarnita published a lopsided compromise: Both sides were right, they said. But one side was more right.

Their own path to the case started back in the fall of 2010, when the pair of then-postdocs met for dinner to discuss collaboration. Tarnita, who grew up acing math competitions in Romania, had switched from math to ecology halfway through graduate school, directing her mathematical firepower into a headline-grabbing critique of traditional evolutionary biology. Now she was looking for a new project.

And Pringle was looking for a mathematician. At a research site in Kenya, he had found termite mounds tiled across the savannah in an ordered pattern. Because the mounds, built by fungus-farming termites, contained more nutrients than the surrounding soil, they became little islands of plants, spiders, insects, and lizards.

The even spacing of the mounds ensured that this benefit was spread around, too. Pringle suspected that competition among the termite mounds explained the pattern, but he didn’t have the mathematical skills to test his theory.

After their dinner, Tarnita and Pringle began collaborating. (The next spring they also started dating, and are now married.)

In Kenya, as they investigated the distribution of the termite mounds, Tarnita began to glimpse evidence of self-organization—the same class of processes that could explain the Namib desert’s fairy circles—in the savannah’s clumps of knee-high grass.

Eventually, she convinced Pringle that the grasses were indeed organizing themselves in a Turing-style process, but only at small scales. Meanwhile, the jostling between termite colonies created the larger pattern of islands. In simulations, they showed, this arrangement of termite mounds would serve as refugia for life during long droughts, fortifying the ecosystem against climate change.

Tarnita and Pringle told me they hadn’t even heard about Namibia’s fairy circles until Jürgens published his sand termite idea. Curious, they followed along when the media glommed on to the debate. They felt strongly that social insects could make sprawling, landscape-scale patterns. They were baffled by suggestions otherwise. “These guys are saying it’s not theoretically possible,” Pringle said. “Of course, it is. We’ll show that it is.”

Their team dropped their own paper last year, which threw a bone to self-organization but mostly boosted the termite theory. Their computer simulations showed that even while grass arranges itself into small clumps, termites are probably responsible for the larger pattern. The simulation results appeared to match real aerial photos taken in Namibia.

Perhaps you will be unsurprised to learn that there’s a rebuttal to that Nature paper, and then a rebuttal to the rebuttal, neither of which has yet been published. But experts on both sides agree that there’s one clear way to resolve the problem: You have to go to the fairy circles, and you have to test your hypothesis in a controlled experiment. And that’s where Jürgens may already be ahead.

A fairy circle at the NamibRand Nature Reserve in southern Namibia (Tilo G / Shutterstock)

Around the fire one chilly evening at Giribesvlakte, Jürgens told us that last year, his team and collaborators went to four fairy-circle sites, ranging from South Africa to Namibia’s far northwest. In each place, they picked 10 roughly equal fairy circles, and chose five unlucky ones at random. Then, with government permission, they covered their random selections with two poisons used by farmers against termites.

Jürgens had checked the results in Giribesvlakte a few weeks earlier, he said, and it was “overwhelmingly clear” that the poisoned circles had many times more grass than the untouched ones—supporting the idea that termites were responsible for gardening the bare patches.

“We should publish that,” he said, smiling, with a slight swagger, before falling into his familiar cadence. “But perhaps, it is intelligent to keep the discussion about fairy circles alive one more year. Therefore, I propose we add some more experiments this year, hoping for a good rain.”

About that rain. Our expedition had aimed for the wettest window of Namibia’s seasons, hoping to catch plants and insects at their most active. But the skies had been relentlessly clear. One evening, after climbing up Leopard Rock, we saw a rainbow across the sky, opposite the setting sun. A few droplets fell, but that was it.

Cape Town, South Africa, famously spent much of this spring awaiting its so-called Day Zero, when the city would run out of water. Namibia has different weather patterns, but this year and most of the preceding ones have been historically dry here, too. In 2016, the country instituted emergency water-saving measures, and one high-ranking state engineer advocated a plan which might have risked conflict with neighboring Botswana by tapping the Okavango River, which runs through both countries. At the last minute, rains saved the day. “Otherwise Namibia would have been in the headlines,” Jürgens said. “Now it’s Cape Town.”

When we arrived at Giribesvlakte, the team said they had never seen it deader. According to daily pictures taken by the SASSCAL station, the prolonged drought had forced local cattle up to Leopard Rock for the first time, then pushed them farther and farther into the plains to graze. Desperate for a mouthful of living vegetation, the cows had mowed the perennial belts of tall grass around the fairy circles almost to the ground.

For Jürgens, this trend is hard to ignore. On the evening after he wandered off to dig in a circle and I watched him from afar, I asked him about the beginning of his career.

In 1980, when he went to South Africa’s Richtersveld as a student, he met some local Nama shepherds. Jürgens himself had tended sheep while growing up on farm country in northern Germany, so they understood each other.

Later that decade, an anti-apartheid group based in Cape Town called Jürgens with a request. South Africa wanted to make the Richtersveld a national park, and in the process, they wanted to kick the locals out. Jürgens agreed to testify on behalf of the shepherds at a small trial in the remote desert. His argument—that the locals were causing minimal harm to the environment—won the day, and when the Richtersveld later did become a park, the people stayed.

But now the shepherds he fought for are all gone to nearby towns. With the climate against them, agriculture has become impossible, echoing larger trends that SASSCAL is tracking across the region. When Jürgens visited the Richtersveld this February, he checked permanent plots he had been observing for 38 years. Half of them were completely bare of living plants. “So now I’m worried,” he said.

Cryptic as they are, the fairy circles may hold a tiny scrap of a larger solution. Once, on a visit to Israel’s Negev desert, Jürgens saw the ruins of the lost Nabatean civilization, which also built Petra in Jordan. In Nabatean architecture, stone structures funnel the desert’s infrequent but furious rains into cisterns, solving the problem of water scarcity much like Jürgens believes fairy circles do.

That inspired him. Perhaps drought-challenged regions could invest more in water-harvesting surfaces, for example. “I have some ideas,” he said, “but maybe I better write them down in a proper publication.”

Although Giribesvlakte was dry and dead this year, the expedition’s visit wasn’t a wash, either. Ants were surveilled, termites captured, dirt excavated. Additional fairy circles were poisoned. And then there was that moment of insight for Jürgens in the waning light, as he scrabbled in the sand of a megacircle right next to Leopard Rock.

The next day, I met him in the place he had been digging. There was a particular grass species there, a plant he had noticed when he first introduced us to the mysterious megacircles. He had me kneel and sink my hands into the sand, as he had, and feel the roots inquiring downward.

The megacircle problem, as he had framed it, was to explain why these circles were supersized, and why they had grass islands growing in the middle of what would normally just be a bare patch.

Here was the new idea: Maybe, he said, this particular grass can resist the termites that want to clear it away, which lets the grass drink from water stored under the bare patch and establish an island. Maybe then the termites, having lost the center of their water-catcher, must expand their radius outward to compensate.

It was not a proper theory but a proposition; a beginning. The megacircles had bothered him for seven years, he said: “Finally, for the first time, I have an idea.”