© Stein et al., Current Biology, 2019Fossilized roots in Cairo, New York.
The fossilized web of a 385-million-year-old root network has scientists reimagining what the world's first forests might once have looked like.

The picture they have painted couldn't be more different to what now sits in its place. Near the small town of Cairo in upstate New York, under an old highway department quarry, scientists have reconstructed the remains of what was a mighty and mature old-growth forest - home to at least three of the world's earliest tree-like plants.

Some of these initial tree 'wannabes' (known as cladoxylopsids) would have looked like large stalks of celery, shooting 10 meters (32 feet) into the sky. Others resembled pine trees, but with hairy, fern-like fronds for leaves (Archaeopteris). The third long-lost plant would have taken after the palm tree, with a bulbous base and canopy of fern-like branches (Eospermatopteris).

© WikipediaA reconstruction of Archaeopteris macilenta from the Late Devonian, Walton Formation of Hancock, New York
Seven parallel cross-sections of the Cairo site have researchers thinking these primordial trees were quite old and large. As such, they were not packed densely together, but were relatively scattered across a floodplain which ebbed and flowed with the seasons.

Dry periods were a regular part of the cycle, and yet the Cairo forest, which traced the Catskill river, seemed to host primitive trees we once thought could only survive in swamps or river deltas. These tree-like plants belong to the genus Eospermatopteris, and they look sort of like tall ferns standing on bulbous stumps.

Because these towering plants have shallow roots that don't branch, they probably didn't cope well in drier conditions - so their presence in the ancient floodplains of Cairo is confusing.

Previously, scientists have only found evidence of Eospermatopteris trees in wet lowland conditions, like the prehistoric site of Gilboa, also in New York state.

Unlike the uniform swamps of Gilboa, however, the Cairo site is 2 or 3 million years older, and its landscape is quite varied. Researchers think it was once made up of an abandoned channel with banks, and a local depression that was filled with water only in certain seasons.

Yet Eospermatopteris trees seemed to thrive here nonetheless, possibly for more than 16,000 years. Their roots had adapted to the semi-arid conditions and the possibility of short-term flooding, researchers say.

Other trees in the area came more prepared for bouts of water scarcity.

At the Cairo site, researchers also found evidence of deeper root systems from extinct pine tree-like plants, belonging to the Archaeopteris genus. These are more advanced than Eospermatopteris trees, with woodier branches and true leaves that can photosynthesize; they also possess deeper roots, sometimes spreading 11 meters wide (36 feet) and 7 meters deep (23 feet).

It was these traits that were first thought to allow primitive fern-like trees to break away from lowland swamps hundreds of millions of years ago, ultimately making their way into drier areas like floodplains, where the water table can rise and fall.

But the new findings suggest even primitive Eospermatopteris trees, without true leaves or deep roots, could have left the swamp for drier conditions.

"This finding suggested that the earliest trees could colonize a range of environments and weren't limited to the wet environments," explains evolutionary ecologist Khudadad from Binghamton University, New York.

"Not only could trees tolerate drier environments, but also the harsh environments of the expansive clays that dominated Catskill plains."

So why is it that so often we find Eospermatopteris trees dominated prehistoric deltas while Archaeopteris trees dominated the floodplains? As the trees still used spores and not seeds to reproduce, surely they would have been more likely to set up near rivers or water sources that can carry their genes farther afield.

The authors of the new study think the fossil records might be tricking us. The prehistoric Cairo forest is thought to have disappeared with a long-term flood that waterlogged the trees and killed them. But the sediment that was laid down afterwards might have preserved their roots in a way that happens very rarely in floodplains and more commonly in deltas.

"It's possible that because of the ideal condition needed for the preservation of the landscapes and organisms, the fossil records were biased towards the low-lying areas, and hence led to the conclusion that Eospermatopteris were constrained by their morphologies to the deltaic environments," the authors write.

Given the sheer age of Cairo's prehistoric forest, the authors doubt its structure is an anomaly. Instead, they argue it is "very likely a representative of mature forests of the time that haven't been preserved or yet to be discovered."

The study was published in PLOS One.