The area where Peru’s Manú River flows today may have once been covered by a shallow sea.

Jason Houston

Was the Amazon once an ocean?

The Amazon rainforest is a treasure trove of biodiversity, containing 10% of the planet’s species in its 6.7 million square kilometers. How it got to be that way has been fiercely disputed for decades. Now, a new study suggests that a large section of the forest was twice flooded by the Caribbean Sea more than 10 million years ago, creating a short-lived inland sea that jump-started the evolution of new species. But the new evidence still hasn’t convinced scientists on the other side of the debate.

“It’s hard to imagine a process that would cover such a large forest with an ocean,” says lead author Carlos Jaramillo, a paleontologist at the Smithsonian Tropical Research Institute in Panama City who has been in both camps.

Researchers generally agree that parts of the Amazon were once under water, but they don’t agree on where the water came from. Those in the “river camp” argue that freshwater streaming down from the rising Andes sliced up the land below, dividing plants and animals into isolated groups that later turned into new species. The fast-growing mountains also created microclimates at different elevations, sparking speciation and funneling new plants and animals into the Amazon basin. However, when marine microorganisms were discovered in Amazonian sediments in the 1990s, some scientists hypothesized that the forest was once inundated by an ocean, which created new species as forest dwellers quickly adapted to the flood.

But proving either case—the river view or the ocean view—is tough. Rocks and fossils that could paint a definitive picture are exceedingly rare. So Jaramillo and his colleagues turned to a different kind of data: cores drilled into the jungle floor. Six centimeters wide and 600 meters deep, the cylindrical cores preserve a record of the region’s past environments in the form of pollen, fossils, and sediments, going back tens of millions of years. Jaramillo used two cores: one from eastern Colombia, drilled by an oil company, and one from northeastern Brazil, taken by the Brazilian Geology Survey in the 1980s.

Jaramillo’s team went through the cores layer by layer. Most of the remains came from land-dwelling species. But in two thin layers, it found marine plankton and seashells. The Colombian core even contained the fossils of an ocean-dwelling mantis shrimp and a shark's tooth. That was enough to convince Jaramillo, who was once firmly in the river camp, that the Caribbean Sea had reached down into the western Amazon of Brazil, Ecuador, and Peru twice: once 18 million years ago, and again 14 million years ago, he writes today in Science Advances. “It’s a lost ecosystem,” he says.

These seas didn’t last for long. In northwest Brazil, the first flood endured some 200,000 years, while the second lasted 400,000 years. Colombia, which is closer to the Caribbean, was inundated for a longer period, 900,000 and 3.7 million years, respectively. Those floods could have been caused by the growing Andes, Jaramillo says. The mountains would have pushed down the rest of the continent as they thrust upward, letting seawater flow in. But that water would have been quickly displaced as freshwater and sediments flowed down the peaks and rebuilt the basin.

In geological time, these floods lasted a mere blink of the eye, Jaramillo says, “but it’s still a long time for a tree.” Even these relatively short events would have transformed the region.

The new work “makes the case [for marine flooding] much stronger, and it makes the timing more definite,” says Carina Hoorn, a geologist and palynologist at the University of Amsterdam and Ikiam Regional University of Amazonia in Tena, Ecuador, who first proposed the marine flooding theory. But Paul Baker, a geologist at Duke University in Durham, North Carolina, and Yachay Tech in Urcuquí, Ecuador, is still a firm member of the river camp. “In [Colombia], I don’t have any problem with there being a marine incursion,” Baker says. But the Brazilian core troubles him, because marine-looking plankton has turned up in other ancient freshwater lakes in Europe, he says. More convincing to Baker would be a measurement of oxygen isotopes in the shells, which could reveal whether they grew in salt- or freshwater. Jaramillo says he’s already working on it. He’d also like to find more Amazonian fossils to study species that may have gone extinct during this dynamic time.

For now, there’s only one thing Jaramillo, Hoorn, and Baker can all agree on: They will need to drill and study many more cores from across the region to solve the mystery of the Amazon’s biodiversity.