Filaments of fungi intertwine with the tips of tree roots to form underground networks that seem to benefit both organisms: the filaments, called hyphae, break down minerals in the soil that trees can then take into their roots, and the fungi get a steady supply of sugar from the trees.

Research has hinted that these connections—known as mycorrhizal networks—can extend between trees, letting one tree transfer resources belowground to another. Some scientists even argue that trees are cooperating, with older trees passing resources to seedlings and nurturing them as a parent might.

This idea of forests as cooperative, caring places has caught on in both scientific literature and popular culture, notably in the 2021 book Finding the Mother Tree: Discovering the Wisdom of the Forest, by University of British Columbia forest ecologist Suzanne Simard. There is even a punny name for the phenomenon: the “wood-wide web.”

A new analysis published in Nature Ecology & Evolution, however, argues that the evidence for mycorrhizal networks facilitating tree cooperation is not as strong as the popular story would suggest. It’s not that relationships between trees and fungi don’t exist, says co-author Justine Karst, an ecologist at the University of Alberta. Rather, in many cases, suggestive lines of evidence or studies with many caveats have been taken as more definitive than they really are. “We don’t want to kill anyone’s joy or curiosity or wonder about the forest, but we want to tamp down on some of the misinformation,” Karst says.

Mycorrhizal networks are delicate: dig up a root, and you’ve destroyed the very web of fungi and wood you wanted to study. To begin to figure out if a particular fungus really connects any two forest trees, scientists can sequence the fungus’s genes and construct a map of where genetically identical fungi are growing. This is a tremendous amount of work, Karst says, and she and her co-authors could find only five such studies across two forest types, comprising only two tree species and three fungi varieties.

And fungal networks’ ephemeral nature makes these studies even more complex. Fungi can grow as individuals after their underground connections are split, says study co-author Melanie Jones, a plant biologist at the University of British Columbia—so even genetic samples provide only a snapshot and can’t reveal whether the bits of fungi collected at two different trees are still actually connected. They may have been severed by part of the fungus dying or by something taking a bite out of it. “It’s a very thorny issue,” Jones says.

These limitations raise questions about how widespread mycorrhizal networks are and how long they last. Researchers have verified that substances provided to one tree can be taken up by a neighboring tree in the forest. Researchers can test this by providing one tree with a chemical compound tagged with a certain marker. In a 2016 study in a Swiss forest, researchers sprayed some trees’ leaves with a particular isotope of carbon and found that isotope showed up in unsprayed neighbors. But it’s not clear that fungi are necessarily responsible for this transfer, Jones says. Resources can also move directly from root to root and through pores in the soil, and it’s difficult to experimentally separate those pathways without disrupting tree growth.

To test the effect of mycorrhizal networks, researchers also often set up wide-mesh barriers, allowing fungi but not tree roots through. But Karst and Jones contend that in such cases, some researchers have rarely checked to make sure a connected mycorrhizal network has actually formed. The strongest evidence for trees sending resources via fungal pathways in a forest comes from a 2008 study in which mesh allowed fungi, but not roots, to connect ponderosa pine seedlings to older pines, Karst and Jones say. Dyes applied to cuts in older pines showed up in seedlings, suggesting water transfer via fungal hyphae. But the study authors say evidence is shaky that such water transfer actually improves seedling survival. “In the really well-controlled experiments, less than 20 percent show that the seedlings performed better,” Jones says. In the remaining 80 percent, she adds, hyphae-connected seedlings performed either equivalently or worse than the ones cut off from the fungal network.

Meanwhile another idea—that trees share underground warnings about herbivorous insects or other dangers—is predicated on a single greenhouse study, the researchers say, in which researchers connected a Douglas fir and a ponderosa pine only by fungal networks. When scientists exposed the fir to insects, the pine also started pumping out defense chemicals. The effect disappeared, however, when the firs and pines were connected by both roots and fungi, as happens in the wild. “The main message is that this hasn’t been tested in a forest,” Karst says. “When you see those pictures of ancient forests, big trees ... passing signals to each other—it just hasn’t been tested.”

The main argument for cooperative forests is that trees in a healthy forest survive better than trees in a sickly one. But such instances of natural selection as a group are rare in the wild, says Kathryn Flinn, a plant community ecologist at Baldwin Wallace University in Ohio, who was not involved in the new analysis. And in forests, individual selection favors competition, with trees vying for resources in a way that would prevent group benefits. “I find this whole controversy really interesting because it’s an example of people wanting to project their own values onto nature and of them wanting to see in nature a model for human behavior,” Flinn says.

Simard, whose forest research has provided much of the basis for the assertions that trees cooperate, responded to questions with a statement that she stands by her research. “Forests provide crucial support to life on our planet. Reducing ecosystems to their individual parts hinders us from understanding and appreciating the emergent relationships and behaviors that make these complex ecological systems thrive,” she says. “For decades a compartmentalized approach has hindered us from better understanding why forests help to regulate global climate and harbor such rich biodiversity. Applying reductionist science to complex systems accelerates the exploitation and degradation of forests worldwide.” 

Karst, Jones and their study co-author Jason Hoeksema note that most of the experiments available are narrow in focus, and they have already been used to make big claims about mycorrhizal networks. The researchers chose to focus on the subset of studies conducted in real forests, they add, because those are most relevant to the real world.

Karst says that she and her colleagues hope to push research into additional forest types and encourage investigation of the most promising areas, such as water transfer between trees. For her part, Karst thinks mycorrhizal networks may be involved in at least some tree-to-tree networking, and better-designed experiments could get at that truth. “I want to have another go at it,” she says.

Editor’s Note (4/13/23): A version of this article entitled “A Tangled Web” was adapted for inclusion in the May 2023 issue of Scientific American. This text reflects that version, with the addition of some material that was abridged for print.