The Search for America’s Atlantis

Illustrations by Studio Hosego

Editor’s note: This article is part of a new series called “Who Owns America’s Wilderness?”

Like apparitions, California’s Channel Islands sometimes vanish in the morning fog. Even on mist-free days, when their golden cliffs can be glimpsed from the mainland, few people seem to take much note of them. Despite their proximity, the islands are seldom visited by Californians, who mostly know them for the way their silhouettes interrupt the horizon of a Santa Barbara sunset.

Last August, I traveled to one of the largest of the Channel Islands, Santa Rosa. I joined an expedition led by the archaeologist Todd Braje, who has spent 15 of his 45 years doing fieldwork in the islands, during which he has acquired a feel for their primeval landscapes. On our first morning hike, in Arlington Canyon, an Ice Age watershed on the island’s northwestern edge, Braje walked the canyon’s lip, looking for the gentlest slope before bounding downhill. I followed close behind, wading through pale-turquoise sage and Day-Glo-yellow poppies, dodging cacti that looked like spiked Ping-Pong paddles.

When we reached the canyon’s lower terrace, Braje stopped and pointed to a length of twine strung vertically against a cliffside. Another length of twine was strung horizontally a few feet away. Stepping back, I could see others. All were remnants of a grid that archaeologists had pressed into the cliff’s sedimentary layers, so that anything lodged within them could be dated. Braje asked me not to reveal the grid’s precise location. All I’ll say is that we were close enough to the coast to smell the sea, but too far away to hear the waves.

[From the April 2017 issue: Welcome to Pleistocene Park]

Santa Rosa is an archaeologist’s dream. Its landscapes have suffered few injuries from commercial development. And now that it’s part of Channel Islands National Park, the island is uninhabited but for a small campsite, the ranger’s headquarters, and Park Service housing, where we were staying. Better still, Braje explained, its sedimentary layers have never been scrambled by gophers, because no burrowing rodents have made it across the channel. The island is a well-preserved time capsule, and the archaeologists unsealing it are already stumbling upon extraordinary finds, especially in Arlington Canyon.

Braje scraped away part of the cliff’s face, revealing bands of ancient soil, cleanly differentiated by color. He ran a finger along the darkest one, a coffee-colored signature of a distinct geological moment roughly 12,000 years ago, when a global flash freeze briefly returned the Earth to Ice Age conditions. In the older soils beneath was a small depression, unremarkable in appearance, but epochal in its significance for Paleoindian archaeology. It was from this cavity that archaeologists pulled the oldest human bone ever excavated in California, and perhaps the oldest in the entire Western Hemisphere. The 13,100-year-old human femur fragment belonged to “Arlington Man,” whose presence here may help resolve one of the final mysteries of the human origin story: the identity of the first Americans.

Bipedal primates have made epic overland treks ever since Homo erectus left Africa for Eurasia more than 1.7 million years ago. But as early as 50,000 years ago, the ancestors of Oceania’s Aboriginal peoples became the first hominids to ever make an open-water migration. Off Indonesia, a few of them slipped into simple watercraft and braved at least 90 kilometers of waves to reach Sahul, the landmass that has since split into New Guinea and Australia. Still-greater seafaring adventures lay ahead. New research suggests that the Māori made it to Antarctica 14 centuries ago. The Lapita people island-hopped from Taiwan to Samoa and Tonga by watching subtle colors on the underside of Polynesian clouds. A growing number of archaeologists now suspect that the first Americans also came by sea.

Nearly all dates in Paleoindian archaeology are contested, but there’s a relative consensus on the timing of the peopling of the Americas. The genomes of living Native Americans suggest that their ancestors first arrived in North America more than 15,000 years ago. In only a few thousand years, they ripped down from the polar cold of northeastern Eurasia through two continents of terra incognita, which encompassed every known terrestrial biome. This feat of exploration surely ranks among humanity’s greatest, but the route and identity of those who achieved it remain matters of fierce dispute.

[Read: A new study says humans were in America 130,000 years ago]

Before the 20th century, Westerners entertained several speculations about the first peoples of the Americas, but only in the 1930s did the archaeologist Edgar B. Howard find what appeared to be actual traces of them. At an arid site near Clovis, New Mexico, Howard unearthed an ancient spear tip. The well-engineered projectile was sharp enough to penetrate thick megafauna hide, but small enough for an adult man to carry in his palm. Its slumber underground had not degraded its shape, and its purpose was preserved by the manner of its internment, amid a scatter of mammoth bones.

The “Clovis point” was not a one-off discovery. In the decades that followed, these big-game-killing spear tips were found all over North America. The oldest dated back about 13,500 years (though some researchers believe the sites stretch back only 13,000 years) and appeared, at the time of its discovery, to be evidence of the continent’s first material culture.

The ubiquity and geographic distribution of Clovis points suggest rapid colonization. For decades, researchers believed that this colonization began when the Clovis people made their exodus across the grassland steppe that then connected Siberia and Alaska. After the rollback of two ice sheets opened a new corridor east of the Rockies, they raced down into the North American interior. By then, humans had been living in Africa, Europe, Asia, and Oceania for at least 30,000 years. Apart from Antarctica, the Americas were the last of Earth’s major landmasses to be glimpsed by human eyes.

It was easy to conjure the journey of the Clovis bands trudging down from the north. Presuming they took this route, it must have seemed like providence when the frozen white wall of ice sheets to the south parted, revealing an unending series of spectacular landscapes, flush with megafauna that did not yet fear a human spear. Few, if any, prehistoric humans would have ever known such abundance. Chasing mammoths, mastodons, and giant bison, the Clovis quickly spread east to the Atlantic seaboard, west to California’s inland valleys, and south through Central America and South America. Along the way, according to one theory, they exhausted the bounty, hunting most of America’s megafauna into extinction.

[Read: A mammoth, a spear, and a new timeline for humans in the Arctic]

You may have encountered this story before in a science documentary, or etched into a museum plaque. The “Clovis first” paradigm was an axiom of Paleoindian archaeology for decades, long enough to trickle into school textbooks and dioramas. But recently, a competing theory has earned the respect of the field’s senior scientists.

In 1979, the Canadian archaeologist Knut Fladmark proposed that before the inland corridor opened for the Clovis people, humans traveled along the west coast of the Americas on small watercraft. According to Fladmark, the first Americans were not the storied big-game hunters of popular culture. They were skilled mariners who Braje thinks might have gorged themselves on otters, shellfish, and strips of campfire-dried seaweed.

Fladmark’s theory remained a fringe position for decades, but in 1997 scientists gave it a second look after archaeologists excavated Monte Verde, a coastal site that is roughly 14,500 years old—a full 1,000 years more ancient than any Clovis site. Its former inhabitants did not appear to be big-game hunters. They did, however, collect nine different types of seaweed. Strangest of all, Monte Verde is in Chile. If people were living down there 14,500 years ago, their ancestors probably began their southward trip from Beringia, the region connecting Siberia, Alaska, and the Yukon, well before the Clovis people speared their first American mastodon. And along the way, they may have stopped in the Channel Islands.

Braje and I were joined on the expedition by Kristin Hoppa, the staff archaeologist for Channel Islands National Park, and Brian Holguin, a rising star in geochemical archaeology. Holguin is a member of the Santa Ynez Band of Chumash Indians. Although they do not currently inhabit the Channel Islands, the Chumash called the area home for at least 8,000 years, and perhaps go all the way back to Arlington Man and his ancestors, who were likely the first Paleoindians on the islands. When the Spanish—no seafaring slouches themselves—first arrived in Chumash territory, in the 16th century, as many as 25,000 Chumash were living in California, and those on the coast were a decidedly waterborne people. A Spanish explorer named the first Chumash village he encountered “Pueblo de las Canoas,” on account of all the watercraft clustered on the shore.

“We believe we are the proper stewards of these lands,” Holguin told me. On top of his own archaeological work, he sometimes joins expeditions like this one because of his knowledge of the Chumash community. Holguin is the great-great-great-great-grandson of Maria Solares, the modern tribe’s cultural matriarch, who helped preserve many of its traditions. In the early 20th century, after persecution under the Spanish mission system, and the general decimation of California’s Native American population during the Gold Rush, Solares gave a series of interviews to an anthropologist from the Smithsonian. Decades later, transcripts of the interviews were used to reconstruct the Chumash language Samala, giving it new life.

Among the lore that survived is the Chumash origin myth. According to this story, the Chumash people bloomed from a seed in the soil of the Channel Islands, planted by the Earth goddess herself. After allowing them to flourish there for thousands of years, she told some to leave, to go fill the mainland, which was then empty of people.

Braje, Holguin, Hoppa, and I hiked toward the sea from Arlington Man’s final resting place. As we sidestepped down the steep walls of the lower canyon, Braje argued that the age of the Paleoindian femur fragment supports the idea that people lived on North America’s west coast while the continent’s inner regions were still uninhabited, as the Chumash legend suggests. He told me he suspects that Arlington Man’s ancestors alighted on this island’s shores hundreds or even thousands of years before the Clovis people made their southward sprint into the interior.

Arlington Man’s remains were originally excavated by the archaeologist Phil Orr in 1960, decades before they were dated precisely enough for their significance to be understood. At the time, the Chumash were largely powerless to stop scientists from digging wherever they pleased. On the whole, that golden age of archaeology was not overly concerned with the wishes of Indigenous populations. “Some dug up cemeteries that were so fresh, they complained of the smell,” Braje said.

The Chumash still view archaeologists with suspicion. “I’ve had people sit this far from my face and call me a traitor,” Holguin told me, holding his fingers an inch apart. He said that some scientists remain sloppy in the way they deal with people, but that in general, things are getting better. “The older scientists are starting to retire, and this generation has a new perspective,” Holguin said, nodding toward Braje.

Orr originally suspected that Arlington Man hailed from the end of the Ice Age, based in part on his femur’s location, 37.5 feet beneath the present surface. But it wasn’t until 2008 that radiocarbon analysis confirmed its age as 13,100 years old.

According to the Clovis-first theory, people entered the Americas less than 500 years before Arlington Man died here. If Arlington Man’s ancestors were part of the Clovis culture, that means they would have taken only a few centuries to journey thousands of miles to California’s coast. That’s not altogether implausible, given the speed of Clovis expansion, but if they did take this route, they then promptly shed their existing hunting tools (no Clovis points have ever been found on the Channel Islands) and developed a sophisticated set of coastal technologies. This scenario strikes Braje as unlikely, because the artifacts in the Channel Islands appear to have emerged from a culture that had long lived in ecological communion with the ocean.

We get a clue as to the potential sophistication of this culture from the location of Arlington Man’s remains. During the Ice Age, when fewer carbon molecules were afloat in Earth’s atmosphere than today, more sunlight bounced off the planet’s surface and back into space. Much of Earth’s water was locked into thick ice sheets. Sea levels fell so low that some of California’s beaches stretched dozens of miles farther out to sea than they are now—and Santa Rosa was joined with three of the other Channel Islands into a super-island, called Santarosae. The channel separating the island from the mainland was narrower, but it was still too wide and violent for people to swim across. Arlington Man’s presence on Santarosae is widely considered the oldest evidence of watercraft use in the Americas.

Researchers who work in the Channel Islands say that nowhere else in the Americas has a denser collection of archaeological sites older than 8,500 years. And many more of the islands’ former human settlements may now be hidden—submerged as the Ice Age waned and meltwater from Earth’s glaciers raised seas.

Before that “big melt,” when the Clovis people would still have been penned up near the North Pole, wondering what lay beyond the great white wall to the south, the descendants of the first Americans may have been living along Santarosae’s beaches, in small settlements that have since been claimed by the waves, just like the Atlantis of myth.

You can imagine a brave few shoving off from some cove south of the Arctic Circle, heading into the great unknown. The first thousand-mile push southward would have been particularly perilous. Some of the sea surface would have been iced over, and giants would have lurked in the boats’ midst—gray whales along the coast, and bowheads, right whales, and orcas roaming the deeper waters, in abundances that defy all modern experience.

Looking over the port side of their small boats, toward Alaska, these explorers would have seen the edge of the Cordilleran ice sheet, the blue-veined white behemoth that ranged hundreds of miles inland.

Archaeologists once assumed that this sheet barred entry to North America’s interior, and perhaps forbade coastal navigation altogether.

This assumption bolstered the Clovis-first paradigm: Researchers weren’t sure how people could have navigated the giant ice blocks guarding the coast, or where they would have stopped along the way. Even if there were gaps of exposed shoreline, skyscrapers of glacial ice would have likely loomed above, threatening to rain down frozen boulders. The perceived hostility of this route made it hard to believe that anybody came to the Americas by water.

But new modeling suggests that the ancient coastline was not uniformly icebound. Instead, North America’s west coast appears to have been a complex fractal of microenvironments, including some ice-free zones along the mainland, and even more on offshore islands, which were outside the ice sheet’s reach. A coastal people would have known to look for bountiful ecologies where major rivers came to the sea. Ancient bear bones have been found near some Ice Age estuaries, suggesting rich food webs capable of supporting apex predators.

[Read: What happens to meat when you freeze it for 35,000 years]

At the largest rivers, small groups might have splintered off, rowing inland where the waters were calm, and hiking along the banks of the more violent stretches. Zigzagging into the continent, through river valley after river valley, they likely would have feasted on salmon runs that had never before been fished by humans.

It’s not clear how far inland these explorers could have pushed, but archaeologists have discovered a tantalizing cave site 200 miles into the Oregon interior, where they found human feces that date back 14,300 years, when the Pacific Northwest was likely inaccessible by any inland route. At another site farther northeast, in Idaho, archaeologists dug up hearthstones, charred and cracked by campfires set more than 15,000 years ago. This location predates all Clovis sites by centuries, and it too was almost certainly inaccessible from the interior, but connects to the coast by way of the Salmon, Snake, and Columbia Rivers.

While splinter groups might have traveled inland into what is now Oregon and Idaho, others likely kept moving southward, cove to cove, on what Jon Erlandson, Braje’s thesis adviser, calls the “kelp highway.” Unlike the Clovis people, who had to adapt as they moved through an astonishing variety of ecosystems, these seafarers could remain relatively set in their ways. Similar tools would work on every beach.

South of the Columbia River, the great wall of white on the mainland would have given way to a more porous fence of evergreen. These early travelers might have been astonished when they first laid eyes on the fog-shrouded redwoods that coated California’s coastal ridges. Perhaps some decided to make their home among the giant trees, while others moved on to the ice-free promised land of the Channel Islands. If you believe Braje, some sailed farther south still, along the finger of the Baja peninsula, the horn of Central America, the bulge of northwestern South America, all the way down Chile’s dry coast to Monte Verde.

Todd Surovell doesn’t believe Braje. Surovell is the head of the anthropology department at the University of Wyoming, and for nearly 20 years, he has led excavations of Ice Age hunter-gatherer sites, including one where both a mammoth and a Clovis point were found. “I get that it’s trendy to support new ideas,” he told me. “But I’m a Monte Verde skeptic.”

After the Monte Verde site was excavated, a group of the world’s most prominent experts in the peopling of the Americas flew down to Chile in 1997 to see it for themselves. “This was the Clovis mafia,” Braje says. (Many of the archaeologists had made their careers at Clovis sites.) Many of them had a stake in preserving what Braje calls the “sexiness of firstness,” which is particularly seductive to grant-makers. And yet, upon returning from Chile, nine of their names appeared on a consensus paper affirming that Monte Verde was a pre-Clovis site.

This consensus did not remain completely intact, however. C. Vance Haynes, one of Surovell’s thesis advisers, later recanted. Haynes, who is now 93, said when I called him that he told the trip’s organizers he didn’t want to participate if it meant submitting to a demand for a unanimous verdict. He told me he thought they’d agreed, “but on our last day, we got together in one of the bars in Puerto Montt, and someone said, ‘Okay, do we have a consensus?’ ” Haynes signed the consensus paper a few months after the trip, but he later wrote a paper withdrawing his support, claiming he was pressured into hasty agreement.

Haynes maintains that there isn’t sufficient evidence to support early human occupation of Monte Verde. The materials from which those artifacts are fashioned are properly dated to the end of the Ice Age, he says, but it’s at least possible that humans found them when they melted out of the permafrost, several thousand years later, and only then shaped them into tools. Haynes’s interpretation of Monte Verde isn’t widely shared, and he remains the only one of the archaeological grandees to withdraw his support for the consensus paper. But his story has been important to other skeptics in the field as they also push back against the coastal-route theory.

“I don’t believe in this model of science where the scientific elite declares for us what’s right and what’s wrong,” Surovell told me. “And one site doesn’t change everything. Monte Verde remains an anomaly. Nothing like it has been found before or since.” Ben Potter, an archaeologist at the Arctic Studies Center at Liaocheng University, in China, who believes that evidence for both the coastal-migration and overland routes is inconclusive, made the same point. “There are no other sites that are like Monte Verde.” If Monte Verde was the last stop of an epic coastal migration, there should be several pre-Clovis sites all along the Pacific shoreline. But archaeologists have not yet turned up many firm, conclusive signs of coastal communities that can be dated to earlier than 13,500 years ago.

[Read: Ancient infant’s DNA reveals new clues to how the Americas were peopled]

The human feces in the Oregon cave and the hearthstone site in Idaho aren’t strictly speaking “coastal” sites, and are too few to hang a whole theory on. Other evidence has been largely circumstantial. While we were surveying on Santa Rosa, Braje spotted a stemmed-point projectile, an ancient hunting technology found on the Channel Islands and other sites along the Pacific coast. Stemmed points are smaller and more delicate than Clovis points, and may have been fastened onto long sticks for spearfishing. Braje thinks they bear some resemblance to projectiles used on the Japanese archipelago, where there is evidence of watercraft going back 30,000 years. He has even suggested that a common cultural heritage might link these two coastal peoples of the Pacific Rim.

Potter is, again, skeptical. “We have some data from the intervening areas north of Japan—in the Kuril Islands, in Kamchatka, and in the Aleutian Islands,” he told me. These are all places that would have connected any ancient Japanese and American seafaring people. “And there are no stemmed points.”

And here we come to the factor that limits all of these debates. According to Braje, those areas have not been fully surveyed, and some areas of interest might not be accessible by land: The big melt at the end of the Ice Age could have hidden many potential coastal sites from generations of researchers.

Like others who aren’t yet convinced of the coastal-route theory, Potter sometimes grew impatient when I mentioned this moving-coastlines argument, as though the coastal-route folks invoked it as a get-out-of-jail-free card. “So all this argument of ‘The coastline is underwater’ is bullshit,” he once told me flatly.

It’s certainly true that not every shoreline on Earth moved miles inland when the Ice Age ended. The sheer heft of the ice sheets weighed down the northern portions of some continents. When they were unburdened by the big melt, they bobbed upward, canceling out the sea-level rise in select spots. Potter estimates that half of the coast that runs from Alaska to British Columbia stayed roughly in the same place. Much of that coastline has proved difficult to survey for pre-Clovis sites, though, because it’s relatively inaccessible by boat, or has since been swallowed by rain forest. And the few excavations in the region have already borne fruit for the coastal-migration theory: A few years ago, archaeologists found 13,000-year-old footprints beneath a beach in British Columbia.

Meanwhile, in the Channel Islands, the shorelines have certainly moved, and there is no rain-forest cover. Braje and his colleagues can put their coastal-route theory to a real test. They can search for signs of the first Americans. They just have to look underwater.

At a small dock on Santa Cruz, another of the Channel Islands, we met the archaeologist Amy Gusick, a curator at the Natural History Museum of Los Angeles County who is leading the offshore search for pre-Clovis sites. Her work is part of a wave of interest in Ice Age settlements among underwater archaeologists, who have found sites on Australia’s continental shelf and in Europe’s North Sea. Gusick, Braje, Holguin, Hoppa, and I traveled over land to a location where all their work came together. We piled into a pair of trucks and drove through eucalyptus groves up into the hill country in the interior of the island, where oak stands strafed my truck’s side, sending acorns rattling into the bed.

Several times along the way, we had to stop and wait for an island fox to cross. Small, silver-orange, and charismatic, the fox was miniaturized by the evolutionary pressures that shrink many animals on islands. Similar pressures resized a Columbian mammoth population that swam to the Channel Islands more than 150,000 years ago. The world’s only pygmy mammoths, they appear to have gone extinct around the time that humans first arrived on the islands.

Eventually, we reached the south shore of Santa Cruz and hiked down a gentle slope of green coyote scrub and pink buckwheat blossoms to a golden-sand beach. The sea was dark navy and turquoise, marbled with foamy breaks, and the sky was clear enough to see the spray edge of a white cap hundreds of yards out. We trekked two miles down the coast and crossed a field of tide pools, each a rocky urn bejeweled with purple urchins and neon-green anemones. I told Braje that I admired his commitment to doing research amid spectacular coastal scenery. “No wonder the Clovis researchers in the interior resent you,” I said.

“It’s certainly easier to recruit grad students,” he replied.

Turning back toward the bluff, I asked him how long until we reached our site. “We’ve already arrived,” he said. Sure enough, the soil below us had gone dark, as it does in many of the places where ancient coastal humans settled. Generations upon generations of organic material—campfire charcoal, bone chips from hunted dolphins, guts of fish pulled up from kelp forests—had all seeped into the ground, forming a night-sky-colored layer made starry by the shining of tiny, pearlescent seashell shards.

As we walked farther into the site, these fragments increased in size to large chips and then whole shells. Between them were splinters of ironwood, a material used by the site’s inhabitants to pry abalone from slick rock outcrops. The abalone husks littered across the ground had each yielded a steak-size dinner.

On the mainland, archaeologists researching Paleoindians often search for cave shelters and debris from stone toolmaking. But in the Channel Islands, they more often hunt for shell middens like these. Middens tend to be near beaches where prehistoric free divers could collect shellfish at low tide. After shucking them, the divers piled the discards into high mounds that likely served as windbreaks for lean-to shelters, housing bands of about 20 people. After a few millennia, the elements flattened the mounds into compact, datable layers of soil, shells, and other artifacts.

[Read: Reconstructing lost worlds with poop]

Securing Holguin’s tribe’s blessing to dig at a midden can require years of bureaucracy and tricky in-person politics. “Like any group of people, Chumash members hold a wide range of opinions on archaeology,” Braje told me. “Some don’t want anything dug up.”

“We want to know what the new dig will do for us,” Holguin, who was standing nearby, said. “Not just what it will do for the scientist’s career.”

With so many middens on the shore, the researchers expect that they might find similar sites offshore, too. Any such finds would help confirm the suspicion that most of the evidence for the coastal-migration theory is underwater. After leaving the shell midden, we drove to the island’s west side and hiked out to Black Point, a sheer cliff that dropped 200 feet to the water. Gusick and I sat on its edge, legs dangling. Gusick pointed to a kilometer-by-kilometer swath of sea, straight out from where we were sitting. Models suggest that beneath its waters was once an estuary, the sort of place where ancient humans loved to settle. It’s in those waters, not on the current islands, where archaeologists may find the most salient evidence for the coastal-route theory.

The summer before the coronavirus pandemic began, Gusick took a high-tech ship out to that site, and four others favored by her models. At each one, the ship’s captain programmed a route that guided it through hundreds of laps, until it passed over every bit of the square kilometer. Meanwhile, a machine towed behind the boat beamed out sound waves that were then processed into a map of the ocean floor, helping Gusick pick places to take core samples.

“It was a 24-hour operation,” Gusick told me. “There was something beautiful about being out on the back deck in the dark, with these huge floodlights illuminating the black sea.”

Early analysis of the soils in Gusick’s samples suggests that she has indeed located an ancient estuary, but they require further testing to be sure. She’s also working with new technology to try to locate resources that humans may have exploited, including tar seeps where people could have used pitch for arrowhead fastening. She might even be able to locate the distinctive shell layers that make up middens in the undersea sediment.

If Gusick finds Ice Age middens on the sea bottom, she will help vindicate the coastal-migration theory. She may pull up a core sample that’s full of shell bits, and perhaps a stone tool or two. Or with cutting-edge technology at a European lab, she could even search for telltale strings of humanity’s genetic code. If one of Gusick’s core samples contains definitive evidence of a pre-Clovis settlement, it will be the biggest find in Paleoindian archaeology since Edgar Howard plucked his first projectile out of the New Mexican desert. It would tell us, perhaps once and for all, that the first Americans were people of the sea.

Of course, Gusick may not find a pre-Clovis site here, and others along America’s Pacific coast may not turn up. The inland-route theory could make a comeback. America’s Atlantis could end up being dismissed as a myth, like Plato’s Atlantis. The Chumash could lose their patience with this academic feud altogether, and bar further excavations on their ancestral lands.

On this last possibility, Holguin has a more optimistic view. He told me he expects the tribe to warm up to archaeology. “I think they’ll realize how beautiful it is to be a part of a story that starts with the evolution of humans and runs all the way to our modern tribes.”

Sitting up on the cliff at Black Point, I found it easy to tap into the grandeur of that story. I could imagine America’s lost Atlanteans arriving on the old shoreline, right at the beginning of the big melt, when the super-island was still covered in giant pines. Maybe they circled Santarosae for a few days, eyeballing the places where rivers came to the sea. Maybe they marveled at pygmy mammoths drinking from streams. Maybe, weary from a great journey, they dragged their small boats up to a bluff and began making a midden that, like so many human homes—past, present, and future—wouldn’t stay above water for long.

This article appears in the October 2021 print edition with the headline “America’s Atlantis.”