American Scientist

A group of plant studies in the early eighties generated a lot of excitement—which was quickly followed by derision. These so-called talking trees studies suggested that plants might communicate in order to warn other plants, and other parts of themselves, of imminent attack by herbivorous insects.

Back then, though, the alarm mechanisms were unclear. And after a high-profile critical review of the results was published in American Naturalist, the studies fell into relative obscurity. “People felt as if, we’ve heard these ideas, we’ve considered them, they’ve been debunked, this stuff is kind of nonsense,” says Richard Karban, an entomologist at the University of California, Davis, who studies plant-herbivore interactions.

But when Karban encountered a 1990 study by Edward Farmer and Clarence Ryan, he was intrigued. The results showed that placing a tomato plant and a damaged cutting from a sagebrush plant (Artemisia tridentata) into a vacuum jar increased the levels of compounds that appeared to be defensive against insect damage in the tomato—even though the tomato and sagebrush had no physical contact.

At the time, Karban was working in the field with wild tobacco, a relative of tomatoes. “I found myself surrounded by sagebrush,” he says. “And I thought, maybe I’ll do a quick and dirty experiment to see if clipping sagebrush makes neighboring tobacco plants more resistant.” It did—in a natural setting rather than in the lab—and Karban was able to replicate the results. “That really got me going,” he says.

He went on to show that individual sagebrush plants used airborne chemical cues for communication between their own branches. “So it got me thinking, okay, maybe these signals are largely about within-plant, within-individual communication,” he says, “and then I wondered if plants might be able to distinguish their own cues from the cues of other individuals. That’s what motivated this particular experiment.”

Many plants can use a vascular system of communication—not so for desert plants. “These desert shrubs, including sagebrush, are very poor at using vascular signals to integrate their physiological responses. So, to get systemic-induced resistance—in other words, if you damage one branch of an individual, for the other branches to respond to that damage—volatile cues were required,” Karban says.

He knew that communication in sagebrush occurred through such volatile cues: In a previous study, he clipped plants’ leaves inside plastic bags, sealed them, and then placed them in proximity to other plants. The plants nearby showed no greater resistance to herbivory—but plants next to unbagged, clipped plants evinced greater resistance.

Karban and his colleague Kaori Shiojiri, of the Center for Ecological Research at Kyoto University, wanted to see whether the effect increased for sagebrush plants responding to cues from their own branches as compared with cues from other, “nonself” sagebrush—and whether this distinction required a physical connection between the plants.

To find out, they cloned sagebrush plants from the south edge of Taylor Meadow in the Sagehen Creek Natural Reserve in California. The cuttings were grown in individual pots, then each pot was placed 5 centimeters from another sagebrush plant growing in the meadow.

In May 2007 and 2008, they placed 15 such clones next to their parent plants, and 15 next to unrelated plants. Then they chose one branch of each clone and clipped the edges of one third of the leaves on it. Two more groups were established to create a control: Some rooted plants were observed, and some were clipped in the same manner as the clones and observed.

Near the end of the summer, in September, they measured the amount of leaf damage on the rooted plants caused by herbivores (mainly grasshoppers) during the course of the season. The results were consistent over both years: Rooted parent plants with a clipped clone of themselves nearby sustained 42 percent less damage from herbivores than those plants with a nonself clone nearby. And the rooted plants that were neither clipped nor near a clipped plant at the beginning of the summer experienced the most damage of all.

Karban and Shiojiri’s study, published in the June issue of Ecology Letters, is the first to show that plants can respond more strongly to volatile cues released by themselves or clones of themselves than those released by nonself plants, with no physical contact involved. In addition, analysis of the volatile compounds released by the clipped plants revealed a large amount of variation in those compounds within individual plants, suggesting a greater facility for serving as cues.

Consuelo De Moraes, a professor of entomology at the Pennsylvania State University, finds the variations between individuals’ volatile profiles interesting. “This is an exciting study,” she says. “These findings raise fascinating questions about the nature of the mechanisms by which plants perceive volatile signals.” And, she adds, for plants to have such distinct responses to signals from their own branches and those from other plants, “fairly sophisticated” mechanisms might be required.

Karban’s next step is to try to sort out whether volatile communication is part of a genetic system. This would have implications for kin selection (when organisms adopt strategies that help their relatives’ reproductive success), which requires the ability to recognize related organisms. The demonstrated ability of sagebrush plants to respond differently to self-derived cues and cues from other plants is a start.

Some setbacks, including a fungus that infected some plants in his lab, have slowed Karban’s work; as he notes wryly, “Field work is a very stochastic process.” Even so, his studies have made him optimistic about future discoveries: “I think that we’re likely to experience a lot of surprises about what plants are capable of.”