Do Plants Really Communicate With Each Other? What the Chemistry Actually Shows

Do Plants Really Communicate With Each Other?

 What the Chemistry Actually Shows

Every summer, along ordinary roadside hedges in Niigata, Japan, camellia shrubs lose patches of leaf tissue to a small caterpillar in the tussock moth family, 

Euproctis pseudoedentula. Two ecologists at Niigata University spent seasons studying exactly this system, and what they eventually documented moving between damaged and undamaged bushes doesn't show up to the eye at all. A shrub growing near a chewed-up neighbor can end up harder for insects to eat afterward than one growing near an untouched neighbor  weeks later, with no roots touching, no physical link of any kind between them. Just air. Nobody planted a warning in that air. Nobody sent a message in any sense a person would recognize. And yet the outcome looks exactly like one plant passed information to another.

That's the whole mystery in a few sentences, and it's the real question hiding inside "do plants communicate with each other." Not whether plants speak. Whether one plant's ordinary chemistry can alter what happens to a plant nearby  and how far that resemblance to communication actually goes.

Biologists have used the word "communication" for cases like this since at least the early 1980s, and it causes no end of confusion outside biology departments. In animal behavior research, communication just means one organism produces something  a sound, a color, a chemical, a movement  that changes another organism's behavior or physiology, and that the trait persists partly because it does this. Nothing in that definition requires a brain, a nervous system, or an intention to be understood. A firefly's flash counts. A skunk's spray counts. By the same definition, so does a stressed plant leaking an altered blend of airborne chemicals that a neighbor's tissue happens to detect and react to. Whether a given case sits closer to a firefly's flash  built by evolution specifically to be received  or closer to a cough spreading a cold through a room, informative to anyone nearby but not evolved for that purpose, is one of the harder questions in this field. Hold onto it; it comes back later.

The physical carrier for most of this is a class of chemicals called volatile organic compounds, or VOCs: carbon-based molecules light and unstable enough to evaporate off a leaf's surface into open air at ordinary outdoor temperatures. Plants release VOCs constantly for reasons that have nothing to do with signaling anybody. Some attract pollinators. Some repel or poison insects directly. Some are metabolic overflow, the exhaust of a stressed cell running chemistry faster than it can keep tidy. What comes off a single plant is almost never one compound  it's a blend, often dozens of molecules in shifting ratios that depend on time of day, plant age, and exactly what's chewing on it. A 2024 survey of the flowering cultivar Camellia japonica 'High Fragrance' catalogued 349 separate volatile compounds in its scent alone, the mixture changing composition as the bloom aged. That figure describes floral fragrance, not injury response, but it makes the general point stick: plant odor, chemically, is rarely a single message carried by a single molecule.

Damage changes the recipe. When a caterpillar bites a leaf, the wound itself triggers one set of internal chemical responses  but the caterpillar's saliva can trigger a much larger one. T.C.J. Turlings, J.H. Tumlinson, and W.J. Lewis showed this cleanly in 1990: corn seedlings that were simply cut released almost none of the terpenoid volatiles that seedlings actually being eaten by caterpillars did, until the researchers applied caterpillar saliva to the fresh cuts, at which point the corn began emitting the same volatile blend as real herbivory. The plant, in effect, could tell an accident from an attack, and it only bothered producing the elaborate airborne chemical signature for the attack. Compounds released this way are called herbivore-induced plant volatiles, or HIPVs, and they can do more than one job simultaneously. In the corn system, the same volatiles also attracted a parasitic wasp, Cotesia marginiventris, that hunts the caterpillars responsible for the damage  the plant's own distress chemistry doubling as a recruitment cue for its attacker's enemy.

Not every compound in that cloud is a message to anybody, though. Most of a plant's volatile output likely evolved for reasons having nothing to do with other plants  deterring insects directly, screening ultraviolet radiation, venting toxic byproducts. A neighboring plant reacting to some of that cloud may simply be exploiting information leaking out incidentally, the way someone might overhear a private conversation through a thin wall. Biologists keep the categories separate on purpose: a signal is a trait shaped by evolution because it gets received; a cue is something an organism reacts to that evolved for an unrelated reason. Most documented cases of plant "communication" fall somewhere between the two, and sorting out which is which, for any specific compound, can take a research career.

This is where defense priming enters, and it's one of the more elegant ideas to come out of decades of frustrating fieldwork. In 2000, Richard Karban and colleagues clipped sagebrush (Artemisia tridentata) shrubs in California's Sierra Nevada and measured roughly a tenfold spike in methyl jasmonate  a volatile, hormone-related compound  drifting off the wounded branches. Wild tobacco (Nicotiana attenuata) transplanted nearby for the experiment ended up with more of a defensive enzyme and less subsequent damage from grasshoppers and cutworms than tobacco planted beside unclipped sagebrush. At first it read like proof that one plant's chemical alarm had switched a neighbor's defenses on directly. Karban's own follow-up work complicated the story in a way that made it more interesting, not less. A 2006 study found the amount of methyl jasmonate actually reaching field tobacco was too low to activate its defenses fully on its own. What the sagebrush volatiles seemed to be doing instead was priming the tobacco  nudging its chemistry into a more alert, faster-reacting state without switching on the full, metabolically costly defense response itself. A primed plant that is later actually bitten responds harder and faster than an unprimed one; a primed plant that's never bitten never pays for defenses it didn't need. That distinction, readiness against full activation, shows up across a wide range of induced-defense research, because plants that mount every possible defense constantly tend to grow more slowly and produce fewer seeds.

The sagebrush system also undercuts any tidy notion of a single, universal plant vocabulary. A 2016 follow-up found sagebrush responds more strongly to volatile cues from clipped sagebrush in its own local population than to cues from sagebrush roughly 230 kilometers away, populations whose terpenoid chemistry differed measurably by region. Researchers described this, half-jokingly, as something close to geographic dialects. If that pattern holds more broadly, the chemical shorthand for "I've been damaged" isn't fixed even within a single species  it drifts locally, the way accents do.

Internally, much of this runs through two overlapping hormone pathways. Jasmonate signaling generally governs the response to chewing insects and physical wounding; salicylate signaling generally governs the response to pathogens feeding on living tissue. The pathways aren't independent  activating one often partly suppresses the other, which is part of why a plant successfully fending off a fungal infection can become more vulnerable to an insect that shows up days later, and the reverse holds too. None of this needs anything resembling a nervous system. It behaves more like two overlapping thermostats than a decision being made.

Testing any of this rigorously is harder than it sounds, and the field's history explains why so many popular claims outrun the evidence. Early experiments in the 1980s connected potted plants through shared air in sealed chambers, cut a leaf on one, and measured chemical shifts in the other  and drew immediate criticism, because a sealed chamber with shared soil, root contact, and recycled air bears little resemblance to a windy hillside. Getting a clean result outdoors means ruling out root contact, fungal contact, shared insect populations, and ordinary weather, all while still capturing what real plants experience in the field.

The 2024 Camellia study handled this by working with intact plants growing exactly where they'd always grown, along roadsides where Camellia japonica and its relative C. rusticana are common ornamental and hedge trees in Japan. The researchers, Yusuke Sakurai and Satomi Ishizaki, exposed healthy branches  some on the same damaged plant, some on separate neighboring plants  to volatiles collected from artificially wounded leaves, then compared how much real insect and disease damage those branches accumulated afterward against branches exposed only to air from intact leaves. Branches that received the wounded-leaf volatiles lost less tissue, the effect held for both the plant's own undamaged branches and its neighbor's, and it persisted for roughly three months. A separate air-transfer setup, piping headspace air from a damaged branch straight to an intact one with no other physical connection between them, produced the same pattern  evidence that the effect really traveled through air, not soil. That's a genuinely careful result. It's also a narrow one: two related Camellia species, one regional herbivore community, no full identification yet of which specific compounds in the blend are doing the work. Treating it as proof that "plants communicate," full stop, is exactly the leap the data doesn't support on its own.

How does a leaf without a nose detect an airborne chemical in the first place? In broad strokes, some VOCs appear to diffuse into leaf tissue and interact with membrane proteins, triggering a rapid rise in cytosolic calcium  the same kind of ion signal cells throughout biology use to convert an outside event into an internal response. That calcium spike can set off a cascade of reactive-oxygen production and hormone signaling that eventually reaches the plant's genes, switching some on and others off. Much less settled is the receptor step itself. For most of the individual volatiles implicated in plant-to-plant effects, nobody has yet identified the specific molecule on a plant cell's surface that first grabs it.

Calling a damaged Camellia a plant that "warns" its neighbor is reasonable shorthand, and slightly misleading if taken literally. Nothing points to a shrub deciding to alert anybody. What's better supported is a plainer chain: a wound triggers a hormonal cascade, the cascade changes what volatiles a leaf releases, some of those volatiles drift to a neighbor, and the neighbor's tissue adjusts its own chemistry on contact  sometimes in a way that helps it later. Every link in that chain is real. The word "warning" smuggles in something extra, an intention to inform, that the biology doesn't require and mostly can't test for. Researchers working in this space separate the pieces deliberately: a cue or signal is whatever the damaged plant releases; information is what a receiver can extract from encountering it; a response is what the receiver's tissue then does. None of those three needs a fourth term, intention, to exist at all.

Everything described so far happens above ground, in open air. Underneath, a different system runs through soil that most people never think about. Most land plant species partner with mycorrhizal fungi, whose thread-like hyphae extend far past a root's own reach, trading water and mineral nutrients pulled from soil for sugars the plant makes through photosynthesis. Because a single fungal individual can attach to more than one plant's roots at once, sometimes many, these partnerships can knit separate root systems into a shared physical structure known as a common mycorrhizal network. That much is real and well documented.

The "Wood Wide Web," the name popularized for forest-scale versions of this network, is where the science and the story start pulling apart. A 2023 review in Nature Ecology and Evolution led by forest ecologist Justine Karst examined decades of research behind three specific claims: that such networks are widespread through real forests, that they measurably boost the growth of seedlings tapped into them, and that older trees preferentially route resources and defense signals to their own offspring through the network. The first two claims, the review found, rest on evidence too limited or inconsistent to generalize confidently. The third claim  the most viral version, and the one behind the popular image of trees deliberately nurturing their young through underground channels  had no peer-reviewed published evidence behind it at all. The same review found that a large share of papers citing this research had described the network's structure or function inaccurately, suggesting the story had grown with each retelling. None of this makes fungal networks fictional. It means the leap from "fungi physically link plant roots" to "trees use this network the way people use a communications system" is carrying far more certainty than the underlying data support.

Fungal threads aren't the only channel below ground. Roots themselves release volatile and non-volatile compounds directly into soil, and newer research is asking whether those exudates can shift a neighboring root's growth or chemistry, or reshape the surrounding microbial community in ways that indirectly help or hurt that neighbor. This work is younger and messier than the aboveground VOC story. Soil is chemically chaotic compared to open air  diffusion is slow, uneven, shaped by moisture and microbes and particle structure  which makes isolating one plant's "message" from ordinary root secretion far harder. Most researchers in this corner of the field describe it as promising and unresolved rather than settled.

Roots may also do something odder: respond differently depending on who's growing beside them. In 2007, Susan Dudley and Amanda File grew the Great Lakes sea rocket, Cakile edentula, in shared pots  sometimes with siblings from the same mother plant, sometimes with unrelated strangers of the same species  and found that groups of strangers grew noticeably more root mass than groups of siblings, competing harder for the same soil volume. Something the plants sensed, likely through root-released chemicals, seemed to distinguish kin from non-kin. A follow-up study in Arabidopsis thaliana found a similar pattern tied to root exudates. What remains genuinely contested is why. It could reflect reduced competition between close relatives, the kind of effect kin selection predicts in animal behavior. It could also be an artifact of stranger groups being more genetically variable, and therefore more mismatched in size, which on its own can push one plant to overinvest in roots. Both explanations stay on the table, which is a more honest place to leave "plants know their family" than most headlines allow.

None of this adds up to plants helping each other out of anything resembling friendship. The same mycorrhizal fungus that supplies a seedling with phosphorus in poor soil can, in a well-fertilized garden bed where the plant barely needs the help, behave more like a low-grade parasite, pulling sugar out of the plant for less benefit in return. Ecologists describe this as a continuum running from mutualism to parasitism rather than a fixed relationship, and it applies to plant-plant interactions as much as to plant-fungus ones. A neighbor that buffers a plant against drought can turn into a resource competitor the moment rain returns and both start growing fast. Ecology stays conditional almost everywhere researchers look closely, and flattening it into a single story about cooperation erases the part that makes it worth studying.

Priming raises a further question: plant memory, a term biologists use carefully and without metaphor. A plant stressed once  drought, cold, an insect attack  can sometimes respond faster or more strongly to a similar stress weeks or months later, an effect linked partly to lingering chemical marks on chromatin, the packaging material wound around a plant's DNA, that keep certain defense genes easier to switch on. That's a real, physical form of persistence. It has nothing to do with recollection or anything resembling conscious recall. A stressed plant isn't storing an experience the way a person stores the memory of a bad afternoon. It's closer to a lock left half-turned, needing less force the second time to open the rest of the way.

So: do plants communicate with each other? The honest answer depends entirely on which half of that sentence carries the weight. If communication means one plant's chemistry can measurably change what happens to another plant nearby  its defenses, its growth, occasionally its survival  then yes, and both the Camellia and sagebrush systems show it happening under real outdoor conditions, not only in a sealed lab. If "talking to each other" implies intention, awareness, or anything resembling language, no credible research supports that, and no plant biologist studying VOCs or mycorrhizal networks claims otherwise. Both statements are true simultaneously because they answer different questions. Most of the public confusion around this topic comes from the two meanings getting swapped mid-sentence, usually somewhere in a headline.

What still stops researchers cold is the very first step in the chain: the moment a plant cell actually recognizes an airborne molecule as something worth responding to. Decades of work on animal olfaction identified specific receptor proteins tuned to specific odor molecules, receptor by receptor. For most of the volatiles implicated in plant-to-plant effects  the terpenes drifting off a wounded Camellia leaf, the methyl jasmonate lifting from clipped sagebrush  no lab has yet isolated the receptor doing that initial sensing. A few candidate membrane proteins are under investigation; none has produced a confirmed, generalizable answer. Until that gap closes, every account of how a plant detects its neighbor's chemistry describes the outcome without a confirmed mechanism for the one step that makes the whole outcome possible.

YOU MAY ALSO LIKE

Do Plants Like Music? What the Evidence Really Says

The Secret Arithmetic of Leaves How plants compute, count, and remember without a single neuron

Earth Has More Trees Than the Milky Way Has Stars