The competition among trees and lianas is not a battle between unrelated plants. Oddly, it grew out of an evolutionary struggle between tree species.

Coiling, woody vines, many lianas are actually descended from trees that forsook the Puritanical labor of building sturdy trunks and instead opted to slink up the trunks of other trees. This transformation to “structural parasite” has occurred many times, across a quarter of all plant families, sometimes pitting members of the same family against each other. And it truly is a transformation, because it involves much more than a slimmer trunk.

Because a liana’s stem does not need to support weight, the structure of its wood can maximize a different function: the transport of water from its roots to its leaves. Cutting through a liana stem often reveals a cross section ornately dotted, like the inside of a kiwi fruit. Those dots are water-sucking tubes, called xylem. Because they are wider than those in trees, they produce less friction with the water rising through them, vastly increasing the volume that can flow and allowing much more rapid photosynthesis and growth.

For over a century, researchers largely ignored lianas. But in the past few decades, they have finally begun to realize that lianas are important in their own right—powerful plants that can shred the social fabric of forests. It started in the late 1970s when a young graduate student named Francis Putz arrived for a fellowship at Barro Colorado Island, intent on studying how these resource thieves slither through the canopy, groping their way sightlessly from tree to tree. Using a modified .22 rifle, he would shoot fishing line over a branch 20 to 40 meters above, use that line to pull up a nylon cord, and use the cord to pull up a climbing rope. Donning a harness, he spent hundreds of hours dangling in the canopy—taking occasional afternoon naps—and familiarizing himself with the slow, bare-knuckle battles between climbing lianas and trees that did not want to be climbed.

Putz discovered that lianas are contagious. Once one managed to climb a tree, it sent shoots to grasp neighboring trees—not unlike Putz’s fishing lines. This contagion could spread surprisingly far. One liana, an entada pea vine with spiraling, meter-long pods, had spread across the tops of 49 trees.

Putz also saw how vigorously trees fought to avoid liana infestation. Some defended themselves with leaves and branches that easily peeled away, sending grasping vines crashing to the ground. Palms developed serrated sword leaves that could slice through young vines. Other trees had more violent means. In one frightening episode, Putz was 20 or 30 meters up a stout-trunked tabebuia tree when he saw a squall gathering in the distance. Having read John Muir’s account of riding out a storm high in a sequoia, he figured it was worth a try. It was, he says, “the kind of thing you do when you’re in your 20s.”

The winds arrived with the roar of an oncoming train and the screams of agitated monkeys. And for about 10 minutes, Putz jolted and swung on his ropes as the neighboring tree—a slender zanthoxylum with a thorny trunk and fern-like leaves, connected to his own by three different lianas—whiplashed, yanking his tree this way and that. When the wind subsided, he realized he had wetted himself. He also noticed that all three vines had broken. Putz came to see that the tree’s pliancy was defensive. In addition to breaking lianas, its mosh-pit thrashing stripped branches from neighboring trees—creating a buffer zone that the vines would find difficult to cross.

The stakes were clear: When lianas did manage to tether trees together, one falling individual could pull down the rest, killing them all. The lianas, though, could capitalize on the newly intense sunlight to grow rapidly over the clearing These rapidly-growing vines thrive wherever the sunlight is intense—whether at the top of the canopy, or in a clearing on the ground.

At the time Putz conducted this research, scientists generally assumed that these ecological battles were in equilibrium on a broad scale in forests that had little direct contact with industrial society. If lianas overran one spot, trees often balanced it out by dominating elsewhere. But another scientist’s work soon began to undercut this idea.

In the early 1990s, a PhD student at the Missouri Botanical Garden named Oliver Phillips was attempting to answer a longstanding question in ecology: Did forests with faster turnover—that is, faster growth, death, and replacement of trees—also have higher levels of biodiversity? To find out, Phillips had assembled records of tree growth and death from 40 tropical forest plots in Asia, Africa, Australia, and North and South America. Most had been surveyed only twice over 10 to 15 years, but collectively, they spanned from 1934 to 1993. And as Phillips tabulated the results, something strange emerged. The rate of tree turnover seemed to be increasing.

That 2002 discovery made a big splash. “No one believed it at first,” says Schnitzer, who leads the liana research team at the STRI in Panama. Some scientists wondered whether the researchers were simply getting better at counting lianas. But other studies found similar results.

Around that time, Schnitzer himself was drifting toward lianas. He had just finished his PhD dissertation at the University of Pittsburgh, showing that falling trees create gaps in tropical forests that boost biological diversity, by providing space where fast-growing “pioneer” tree species can root. Eventually, through a process called succession, those pioneers give way to slower-growing species that make up the bulk of the forest. But Schnitzer noticed (just as Putz had before him) that lianas also often thrive in these sunny gaps; and what’s more, he found that lianas can actually overwhelm the pioneer trees, shading them and slowing their growth. Gaps that would normally fill within 5 to 10 years often remained sparsely treed for 25 years or more. The lianas, he says, are “holding disturbed areas in a state of arrested succession.”

Over the next few years, the pervasive effects of lianas on forests became increasingly clear. In 2005 and 2006, Geertje van der Heijden, then a PhD student working under Phillips, turned to a set of forest plots in Peru to compare the growth rates of hundreds of trees that were or were not infested with lianas. She found that the trees with lianas seemed to grow more slowly and absorb less CO2, primarily because the vines blocked their sunlight. Van der Heijden estimated that across a single hectare (about 1.4 soccer fields), lianas probably prevented trees from absorbing about 2,000 pounds (920 kilograms) of CO2 per year.

The biomass of the lianas themselves compensated for 29% of that lost CO2 storage by trees—but they do a much worse job of storing carbon. If you look at the whole forest, says van der Heijden, “lianas are driving a shift toward more carbon [stored] in leaves and less carbon in stems.” And while woody stems can lock carbon away for many decades, leaves fall to the ground and decay, returning their carbon to the atmosphere within a few months.

In 2012, van der Heijden teamed up with Schnitzer to do an actual experiment, testing whether lianas really reduce the forest’s ability to absorb and store CO2. A kilometer south of Barro Colorado Island, on the shore of Lake Gatun, workers had used machetes and branch cutters to clear lianas from several swaths of forest. They then returned to record tree growth and death in areas with and without lianas. The results, published in 2015, were dramatic: Even when the biomass of lianas was added in, these vines reduced the forests’ overall absorption of CO2 by 76 percent—or about 8,920 kilograms (19,660 pounds) per hectare per year, nearly 10 times what her previous work suggested. “We were quite amazed,” says van der Heijden, who is now a forest ecologist at the University of Nottingham.

Despite these discoveries, the causes of liana expansion remained elusive. Schnitzer’s team found that higher CO2 levels boosted liana and tree seedling growth equally, ruling out that explanation. They found the same for nutrients deposited by fossil fuel air pollution. Schnitzer and others recognized that they would need to understand how trees are experiencing the warming climate on an intimate level to get closer to an explanation. A plant physiologist at STRI named Martijn Slot turned to the topic in earnest.

During the early months of 2016, Slot spent many days at the Parque Nacional San Lorenzo, a rainforest preserve on the Caribbean coast of Panama, about 15 kilometers northeast of Barro Colorado. In the dim morning light, Slot would step from the spongy forest floor onto the steel grating of a gondola attached to a 12-story crane. The contraption hoisted him 30 meters up, past layer after layer of dense foliage; past rainbow-billed toucans that croaked as they flitted away; past an occasional dangling sloth; until he emerged into sunlight among the trees’ crowns.

Speaking Spanish into his radio, Slot directed the crane operator through a series of delicate maneuvers that brought his gondola into the fragrant embrace of a leafy branch.

For the next few hours, Slot clamped a small plastic chamber over one leaf after another to measure CO2 uptake, in order to determine each leaf’s rate of photosynthesis. Sometimes an approaching thunderstorm forced him to descend. Other times, his gondola swung in a gust of wind and he accidentally snapped the leaf from its branch, forcing him to start over.

After dozens of crane sessions and 1,700 leaves, Slot found that warm temperatures push trees beyond their comfort zone, even on seemingly mild days. While the air temperature usually stayed below 32°C (91°F), sunlit leaves often surpassed 40°C (104°F). The leaves’ photosynthetic rates peaked at about 30°C (86°F) and then plummeted, falling by 90 to 98 percent when the leaf exceeded 40°C (104°F). Slot surmised that this happened because the trees had to triage their competing needs.

During photosynthesis, leaves absorb CO2 through tiny vents on their undersides called stomata, after the Greek word for “mouth,” because they resemble tiny lips. While a leaf’s stomata are open, “a huge amount of water is evaporating,” says Slot. “For every molecule of CO2 that’s being fixed, there’s 300 molecules of water being lost.” In this way, a single large tree can exhale around 150,000 liters (40,000 gallons) of water per year.

As the temperature climbs, so too does the rate of evaporation. At some point the leaf closes its stomata—its mouths—and holds its breath. This stops water loss, but also halts photosynthesis, because the leaf can no longer inhale CO2. Leaves regularly stop photosynthesis during the hottest hours of the day. But rising temperatures could force them to do this more frequently and for longer periods, reducing the trees’ food supply.

Many tropical trees “are already operating at their peak temperature,” says Kenneth Feeley, a tropical forest ecologist at the University of Miami in Florida. “As it gets hotter, they’re becoming less efficient and it’s going to lead to slower growth.” Amazonia has warmed by 0.9°C (1.6°F) since 1950, with another 1 to 4°C (1.8 to 7.2°F) of warming predicted by 2100, depending on the volume of greenhouse gasses that humans emit. Even if rising CO2 levels have caused forests to grow faster, as Phillips has found, rising temperatures could eventually erode and reverse the growth-enhancing effects of CO2. Meanwhile, changes in rainfall are already stressing the Amazon rainforest. The four-month dry season has lengthened by roughly 20 days since 1980, and the region experienced major droughts related to heat waves in 2005, 2010, and 2015-16, killing billions of trees.

This nexus between drought and heat could further boost lianas as trees falter, both now and in the future. Indeed, Schnitzer has found that lianas in Amazonia are more abundant in locales with longer and drier dry seasons. And he and van der Heijden have found that lianas on the south shore of Gatun Lake grow roughly three to four times as quickly as trees during the four-month dry season, even though they grow at similar rates in wet times. These differences especially stood out during the unusually hot and dry 2015-16 El Niño drought, when trees stopped growing entirely, but lianas continued to grow at normal rates. If that happens every five to seven years, whenever there’s an El Niño, says Schnitzer, it will give lianas “a huge advantage.”

The vines also have other abilities that may compound this one. In November 2015, Medina-Vega, Schnitzer’s former postdoc, began an experiment at the Parque Natural Metropolitano, a forest preserve bordering Panama City, which has a pronounced dry season, similar to Barro Colorado Island. There, he selected several dozen tree and liana branches, and meticulously labeled each of their 6,861 leaves with a sharpie. Over the next 17 months, he measured the growth of each branch and recorded every leaf that fell or sprouted. He found that at this site, lianas searched for sunlight more efficiently than trees did. The skinny liana branches lengthened by up to 38 meters—about 15 times more than any tree branch did. Liana leaves, like the stems, were also thinner than those of trees, making them “cheap” and “easily replaceable,” says Medina-Vega. When a liana’s leaf ended up in the shade, the liana simply shed it and replaced it with a new one, somewhere in sunlight. Having cheap leaves also allowed lianas to respond more nimbly to the seasons. Both lianas and trees lose their leaves during the dry season in this location, as in many places in the tropics, but Medina-Vega found that when the rains returned, lianas grew theirs back about a month earlier than the trees did.

So not only do lianas siphon water more quickly than trees, they also punch above their weight in the fight for sunlight. In forests with strong dry seasons, they account for less than 5 percent of the stem biomass—but produce 15 to 40 percent of the leaf mass. Climate change could amplify this advantage. “Let’s say these forests are getting drier and drier,” says Medina-Vega. “The performance of lianas will improve.” And tree growth could suffer.

Despite all the evidence of liana increase, it’s still uncertain what it means. Forest changes involve multiple complicated factors and can take centuries to play out, and the lives of human observers are short by comparison.

Tropical ecologist Flavia Costa of the Instituto Nacional de Pesquisas da Amazônia in Manaus believes that site history may have more to do with liana findings than any global phenomenon. Over a 10-year survey of the Adolpho Ducke Reserve in northern Brazil, she and her colleagues found that overall liana abundance remained the same, even though numbers at different plots fluctuated. “I think that data is not as strong as people try to picture,” she says.

Barro Colorado’s increase, for example, could have come from the fact that it didn’t become an island until 1914, when engineers building the Panama Canal dammed a nearby river. The rising waters led to the formation of Gatun Lake, and Barro Colorado, once a hill surrounded by swampland, became isolated at its center. This isolation could have triggered a longterm ecological cascade, if say, populations of large animals such as peccaries or howler monkeys declined as a result, limiting their dispersal of large tree seeds and giving small-seeded lianas a vacuum to fill.