Long before the trees of today’s Amazon reached for the sky, scientists say, sharks weaved through salty waters there, and mantis shrimp rattled across the flooded forest floor. The researchers published their findings in the journal Science Advances.

Scientists have known for some time now that an area of the western Amazon basin was underwater millions of years ago. The exact source of that water has been the subject of some dispute. Some researchers have envisioned a wide river sweeping down from the Andes, while others say the inundation must have washed inland from the sea. But neither side had compelling evidence to support their ideas—until now.

The researchers studied two nearly-2000-foot-long sediment cores, one taken by an oil company in eastern Colombia, and the other taken across the border in northwestern Brazil by Brazilian Geological Survey. Each was packed with the natural detritus of millions of years of local life.

The bulk of each core told the story of a world on dry land, but within two thin slices—one from about 18 million years ago, another from about 12 million years ago—a glimmer of the sea appeared. The complete cores contained a total of 933 different types of pollen grains. The thin slices included types of pollen only found in salt water.

Bigger clues were yet to come: The researchers also found a fossilized shark tooth and the remains of a mantis shrimp. “It’s a lost ecosystem,” lead author Carlos Jaramillo, of the Smithsonian Tropical Research Institute, told Lizzie Wade in Science.

Lasting a few hundred thousand to a few million years, each flood period was relatively brief from a geological perspective, Jaramillo says. But they weren't so brief that they didn’t completely alter the landscape.

“The life span of a single Amazonia canopy tree is about 200-400 years,” he tells Mental Floss. “Therefore, for thousands of generations, not a single tree could occupy Amazonia. In other words, the immense forest we see today is geologically young.”

The findings came as a surprise even to Jaramillo and his colleagues. He added, “I was of the opinion that there were no floodings, but it turned out I was mistaken!”