A tomato plant suffering from thirst may produce dozens of tiny sounds that human ears cannot detect.

Scientists have discovered that these hidden ultrasonic clicks contain information about the physical condition of the plant.

Even more remarkably, artificial intelligence can distinguish certain types of plant stress by analyzing their acoustic patterns.

The discovery challenges an ordinary assumption about nature: that plants experience their surroundings in complete acoustic silence.

For decades, researchers understood that plants respond to light, temperature, moisture, chemicals, and changes in their environment.

However, the possibility that their internal struggles might produce detectable airborne sounds received relatively little scientific attention.

That changed when researchers at Tel Aviv University began investigating what happens when sensitive microphones monitor stressed plants.

Their experiments revealed an unexpected acoustic world operating just beyond the upper limits of normal human hearing.

The findings were published in the scientific journal Cell in March 2023, attracting widespread international media attention.

Headlines described plants screaming, crying, and calling for help, creating an extraordinary image of distressed vegetation everywhere.

Yet the scientific reality was more complicated, and arguably more fascinating, than those sensational descriptions initially suggested.

Researchers had not discovered evidence that plants experience pain or deliberately produce cries resembling those of animals.

Instead, they uncovered measurable sounds that reflect physiological changes and might provide valuable information about plant health.

Understanding those sounds could eventually transform agriculture, environmental monitoring, and our understanding of interactions between living organisms.

The investigation involved evolutionary biologist Lilach Hadany and zoologist Yossi Yovel, working with an interdisciplinary research team.

Yovel’s background studying bats was particularly relevant because these mammals communicate and navigate using extremely high-frequency sounds.

Specialized microphones developed for studying ultrasound offered an unusual opportunity to investigate whether plants produce similar frequencies.

The scientists began with a deceptively simple question that previous research had not fully addressed through airborne recordings.

Could plants emit sounds detectable from a distance, rather than vibrations requiring instruments attached directly to their tissues?

To investigate, the researchers placed tomato and tobacco plants inside specially designed acoustic chambers that minimized interference.

Sensitive ultrasonic microphones were positioned approximately four inches from the plants, recording signals beyond ordinary human hearing.

The team compared healthy plants receiving normal watering with plants experiencing drought stress or damage from cutting.

This experimental design allowed researchers to identify differences associated with specific conditions rather than ordinary environmental noise.

The results revealed a striking contrast between plants growing comfortably and those experiencing significant physiological disturbances or injury.

Healthy plants generally produced fewer than one detectable sound per hour under the controlled conditions of the experiment.

Tomato plants experiencing dehydration, however, produced approximately 35 ultrasonic clicking sounds per hour on average during stressed periods.

Tomato plants subjected to stem cutting generated approximately 25 sounds hourly, demonstrating another distinct response to physical damage.

Tobacco plants also produced measurable sounds under stressful conditions, although their emission rates differed from those of tomatoes.

The recordings consisted of extremely short acoustic bursts resembling sharp clicks when electronically converted into frequencies humans hear.

Without specialized equipment, an observer standing beside the plants would remain completely unaware of this remarkable acoustic activity.

The sounds generally occurred in ultrasonic frequencies, roughly between 20 and 100 kilohertz, beyond typical adult hearing.

Human hearing usually extends toward 20 kilohertz under ideal conditions, with upper-frequency sensitivity commonly decreasing throughout adulthood.

Many animals, however, possess auditory systems capable of detecting frequencies that humans cannot perceive without technological assistance.

This distinction immediately raised questions about whether the sounds might have biological significance beyond the laboratory environment.

The researchers also measured sound pressure levels that demonstrated these ultrasonic emissions were not merely microscopic internal vibrations.

Some recorded clicks reached approximately 60 to 65 decibels at close range, depending on measurement conditions and frequency.

Although comparisons with human conversation offer perspective, ultrasound behaves differently, making direct loudness comparisons potentially misleading.

Crucially, the emissions traveled through surrounding air rather than remaining restricted to the plant’s internal vascular structures.

That distinction meant organisms possessing suitable hearing might potentially detect stressed vegetation without making physical contact with it.

Scientists had previously studied acoustic emissions inside plants, especially those associated with their internal water transport systems.

However, demonstrating airborne ultrasonic emissions and extracting information about stress represented a significant advance in plant bioacoustics.

The next challenge involved determining whether the recorded clicks contained recognizable patterns or simply represented random physical disturbances.

Researchers turned to machine learning, allowing computational models to analyze subtle acoustic differences across large collections of recordings.

Artificial intelligence did not translate plant language into human words, despite interpretations suggested by some sensational headlines.

Instead, the models learned statistical distinctions between recordings from different plant species and different forms of physiological stress.

The systems achieved classification performance substantially better than chance, including roughly 70 percent accuracy in certain experimental comparisons.

That performance demonstrated that acoustic characteristics could carry useful information about which conditions the monitored plants were experiencing.

A dehydrated tomato plant did not necessarily sound identical to a tomato plant whose stem had been cut.

Similarly, recordings from tomato plants contained characteristics that could help distinguish them from recordings obtained from tobacco plants.

These differences suggested that stress-related acoustic emissions reflect underlying physical processes affected by species and environmental conditions.

Although the accuracy was imperfect, the experiment established a potentially valuable new approach to monitoring plant health remotely.

Researchers then investigated whether artificial intelligence could distinguish plant sounds from complicated background noise in practical growing environments.

A greenhouse presents challenges absent from acoustic chambers, including ventilation equipment, irrigation systems, movement, and surrounding mechanical activity.

The team recorded environmental sounds and trained computational systems to identify plant-related emissions amid those competing acoustic signals.

Their results showed that plant sounds could be detected and analyzed even under conditions resembling agricultural production.

The researchers also explored how acoustic activity changed as tomato plants experienced progressively longer periods without sufficient water.

Interestingly, the number of recorded clicks did not simply increase continuously until the plants became completely dried out.

Sound production rose during developing stress, reached a peak, and subsequently declined as dehydration became more severe.

This changing pattern suggested acoustic monitoring might reveal not only stress presence but aspects of its progression.

Such information could become particularly important when farmers must decide exactly when and where irrigation is necessary.

Traditional monitoring methods frequently rely on visible symptoms, soil moisture measurements, environmental conditions, or specialized plant sensors.

An acoustic system might offer another source of evidence before dehydration produces obvious damage visible throughout the plant.

However, researchers must establish reliability across different crops, climates, developmental stages, and real-world environmental noise before widespread adoption.

Understanding the origin of these sounds is equally important because their production does not necessarily imply deliberate communication.

The leading explanation involves cavitation, a physical phenomenon associated with the movement of water through plant vascular tissues.

Plants transport water upward through narrow structures called xylem, which connect roots with stems, branches, and leaves.

As moisture evaporates through leaf surfaces, it creates tension that helps pull additional water upward through those structures.

During severe dehydration, that tension can become strong enough to disrupt the continuous water columns inside xylem vessels.

Gas bubbles may form or expand within the water transport system, producing sudden physical changes and associated vibrations.

These events can generate acoustic emissions, offering a plausible explanation for the clicking recorded from stressed plants.

Stem cutting can also disturb water transport, providing another possible mechanism behind sounds associated with physical injury.

Nevertheless, additional investigation is necessary to establish precisely how particular airborne clicks originate under different experimental conditions.

The central scientific distinction concerns whether an emitted sound is simply informative or intentionally produced as a signal.

A plant can reveal its physical condition through sound without possessing awareness, emotions, or an intention to communicate.

Yet information does not need deliberate intention to become useful to another organism capable of detecting it.

A predator can recognize footsteps without those footsteps being deliberately produced to reveal the location of its prey.

Similarly, animals might benefit from recognizing acoustic clues about vegetation, even when plants produce those clues unintentionally.

This possibility inspired a remarkable follow-up investigation involving moths and their decisions about where to lay eggs.

Researchers studied female Egyptian cotton leafworm moths, whose offspring depend upon suitable vegetation for food after hatching.

Choosing a healthy host plant can influence whether newly hatched caterpillars encounter favorable feeding conditions during early development.

The researchers investigated whether ultrasonic sounds associated with dehydration could influence the moths’ decisions about potential egg-laying locations.

Their findings, publicly reported in 2025, provided evidence that moths can use acoustic information associated with stressed plants.

In an experimental arrangement, two healthy tomato plants were presented while drought-associated sounds played beside only one plant.

Female moths tended to prefer the silent plant, avoiding the location associated with recordings of dehydrated vegetation.

The experiment was particularly revealing because the plants themselves were healthy, isolating the influence of the recorded sounds.

When researchers interfered with the moths’ hearing, the previously observed preference disappeared, supporting an auditory explanation for behavior.

This provided compelling experimental evidence that insect decisions can be influenced by ultrasonic emissions originating from stressed plants.

It did not prove that plants consciously warn animals, but it demonstrated a meaningful ecological interaction involving sound.

The discovery encouraged scientists to reconsider how much important environmental information travels through frequencies humans ordinarily cannot hear.

Other studies have also examined whether plants can respond to vibrations produced by insects, water movement, and pollinators.

One investigation involving Arabidopsis plants explored their responses to vibrations associated with caterpillars feeding on nearby leaf tissue.

Researchers Heidi Appel and Rex Cocroft studied whether these characteristic vibrations could influence subsequent chemical defenses against herbivores.

Their experiments indicated that plants previously exposed to feeding vibrations increased defensive chemical responses during later insect attacks.

The responses differed from those produced by certain other environmental vibrations, suggesting that vibration characteristics could matter biologically.

These findings should not be interpreted as evidence that plants recognize danger through conscious thought or emotional anticipation.

Instead, they demonstrate that mechanical stimulation can influence physiological pathways involved in protecting plants against potential injury.

Another intriguing study investigated whether flowers respond to acoustic vibrations resembling the buzzing produced by visiting pollinating insects.

Researchers working with evening primrose flowers found that certain low-frequency sounds could rapidly influence sugar concentrations within nectar.

In their experiments, nectar became sweeter within minutes after exposure to relevant sounds, potentially affecting pollinator interactions.

Flower petals also vibrated in response to some frequencies, suggesting a possible mechanism for detecting surrounding acoustic activity.

Such findings raise interesting evolutionary questions about how flowering plants might respond to environmental cues associated with pollination.

Nevertheless, results from individual species cannot automatically establish that all flowers respond similarly or possess equivalent sensory abilities.

Experiments involving pea plants have also suggested that roots may respond to vibrations associated with moving water.

These studies proposed that acoustic information could influence root growth direction under certain controlled experimental conditions and constraints.

However, the mechanisms and ecological importance of such responses require continued testing, particularly outside artificial laboratory arrangements.

Together, these investigations reveal that mechanical vibrations and acoustic information deserve serious attention within modern plant biology.

The scientific investigation of plant responsiveness actually began long before artificial intelligence or modern ultrasonic recording equipment existed.

Indian scientist Jagadish Chandra Bose pioneered experiments examining plant responses to physical stimulation during the early twentieth century.

He developed sensitive instruments capable of measuring subtle growth changes and electrical reactions in living plant tissues.

His research challenged assumptions that plants were passive organisms responding only slowly and mechanically to their surroundings.

Modern researchers have since demonstrated sophisticated electrical and chemical signaling systems operating throughout different parts of individual plants.

These systems help coordinate reactions to injury, drought, temperature changes, and the activities of herbivorous insects and other threats.

In 2018, researchers studying Arabidopsis demonstrated how damage to one leaf could trigger signaling throughout the plant.

They used genetically modified plants displaying fluorescent changes associated with calcium concentrations inside their cells during signaling events.

When leaves were wounded, visible calcium waves spread from damaged regions into other parts of the plant.

These signals helped illustrate how a localized disturbance can activate broader physiological responses far from the original injury.

Glutamate signaling plays an important role in this process, contributing to the propagation of calcium-based defensive responses.

Although glutamate also functions in animal nervous systems, this similarity does not mean plants possess animal-like brains.

It demonstrates that living organisms can employ certain related molecules within very different biological structures and signaling systems.

Further research has shown that plants can detect airborne chemical compounds released by neighboring vegetation experiencing injury or attack.

Experiments using fluorescent indicators have visualized rapid cellular responses when healthy plants encounter certain airborne compounds from damaged neighbors.

Such findings provide direct evidence that plants gather environmental information through channels extending beyond their immediate physical contact.

Chemical signaling between plants has interested scientists for decades, including researchers investigating how vegetation responds to insect feeding.

Some damaged plants release volatile compounds that can influence defensive reactions in neighboring plants exposed to the same atmosphere.

However, the strength and ecological significance of these interactions depend on species, environmental conditions, and experimental circumstances involved.

The popular expression talking trees captures public imagination, but it can oversimplify the underlying processes involved in chemical signaling.

Plants exchanging chemical information are not necessarily engaging in conversations resembling communication between humans or other cognitively complex animals.

Another widely discussed concept involves underground fungal networks connecting plant roots and influencing the movement of resources between organisms.

Forest ecologist Suzanne Simard helped advance research investigating carbon movement between trees associated with shared mycorrhizal fungal networks.

Her work contributed to growing interest in how underground biological connections might influence forest ecosystems and plant interactions.

The resulting wood-wide web metaphor became popular, suggesting forests function through extensive networks of interconnected roots and fungi.

Although fungal associations are well established, several popular claims about deliberate resource sharing remain scientifically disputed or insufficiently demonstrated.

Researchers continue examining whether network connections consistently improve seedling survival or transfer significant resources under natural forest conditions.

Distinguishing documented physiological mechanisms from appealing metaphors remains essential when evaluating claims about intelligence and cooperation in plants.

Questions about plant learning and memory introduce another challenging dimension to the study of behavior without nervous systems.

Experiments involving Mimosa pudica, commonly called the sensitive plant, have examined changes in responses to repeated harmless disturbances.

This species rapidly folds its leaves when mechanically stimulated, producing one of the most recognizable plant movements in nature.

Researchers repeatedly subjected these plants to controlled dropping experiments and observed that leaf-folding responses sometimes decreased over time.

The investigators interpreted some results as possible habituation, a basic form of learning involving reduced reactions to repeated stimulation.

However, interpretations remain debated, and independent replication is important for understanding the mechanisms responsible for changing plant responses.

A more firmly established example of stimulus integration can be observed in the remarkable behavior of Venus flytraps.

These carnivorous plants possess specialized trigger hairs that respond when small animals contact their sensitive trapping surfaces.

Two closely spaced mechanical stimulations can cause the trap to close, reducing unnecessary reactions to isolated environmental disturbances.

Additional trigger events associated with struggling prey help regulate digestive processes and the plant’s investment in nutrient extraction.

Scientists sometimes describe this activity as counting, although the underlying mechanism involves electrical signals rather than conscious arithmetic.

The flytrap demonstrates how plants can integrate successive stimuli into coordinated responses without developing an animal nervous system.

Across these diverse experiments, a consistent theme emerges: plants possess more complex environmental response mechanisms than appearances suggest.

That complexity becomes especially important when considering climate change, agricultural water demand, and increasingly unpredictable environmental conditions worldwide.

Farmers often struggle to balance irrigation costs against the potentially devastating consequences of allowing valuable crops to dry.

An acoustic monitoring system could theoretically identify emerging water stress by interpreting ultrasonic signals before visible symptoms become severe.

Artificial intelligence might then combine those recordings with soil moisture, temperature, and other measurements to improve irrigation decisions.

Such technology could potentially reduce water waste while helping farmers protect yields during periods of severe environmental pressure.

But the original experiments alone do not establish that acoustic irrigation systems are already reliable at commercial scale.

Researchers must still address competing background noise, sensor placement, crop diversity, maintenance requirements, and overall operational costs.

They must also determine whether acoustic measurements consistently outperform or complement existing agricultural monitoring technologies in practical settings.

Potential applications extend beyond commercial agriculture into conservation programs dedicated to protecting vulnerable and environmentally important plant species.

The Wollemi pine offers one example of a rare tree whose survival depends upon careful environmental protection and monitoring.

Discovered living in Australia in 1994, this remarkable conifer belongs to an ancient evolutionary lineage once known through fossils.

Its surviving wild population is extremely limited, and conservation authorities protect its habitat from disturbance and unnecessary exposure.

During the devastating Australian bushfires of 2019 and 2020, extraordinary firefighting measures helped protect its remaining wild groves.

Scientists could someday investigate whether acoustic monitoring offers useful early indicators of stress in vulnerable trees like these.

For now, applying recordings from greenhouse tomatoes directly to ancient forest conifers would go beyond available experimental evidence.

Different species possess different vascular structures and environmental responses, so monitoring techniques require careful development and scientific validation.

The broader importance of plant bioacoustics lies in revealing biological information that conventional human senses cannot directly access.

Humans have always watched plants for changes in color, shape, growth, movement, and other visible signs of stress.

Ultrasonic microphones introduce another observational dimension, allowing researchers to investigate physical processes previously hidden from ordinary perception.

Machine learning can make these measurements more useful by detecting complicated patterns across enormous quantities of recorded data.

But interpreting those patterns requires restraint, especially when public attention encourages sensational comparisons between plants and suffering animals.

Current evidence does not establish that plants experience human-like pain, emotions, awareness, or intentions when their tissues produce clicks.

What researchers have demonstrated is that physiological disturbances can generate sounds carrying information about plant species and condition.

Experiments involving moths additionally demonstrate that other organisms can respond meaningfully to this previously overlooked acoustic information source.

Future studies may reveal additional interactions involving insects, mammals, plants, and the broader ultrasonic environment surrounding natural vegetation.

They may also clarify which sounds are unavoidable physical byproducts and which, if any, evolved specifically for signaling.

That distinction matters because communication in biology can involve different mechanisms, including unintended cues and evolved signals between organisms.

The research also reminds us that scientific discoveries often begin by questioning assumptions about seemingly ordinary features of nature.

A tomato plant may appear motionless and silent, yet complex physical processes continuously operate beneath its familiar green surface.

Those processes regulate water movement, detect environmental changes, activate defenses, and respond to conditions influencing the plant’s survival.

The extraordinary achievement was not teaching plants to speak, but developing instruments capable of revealing information already present.

Scientists combined sensitive recording technology, carefully controlled experiments, and artificial intelligence to uncover patterns hidden beyond human hearing.

Their findings encourage a more nuanced understanding of plants as responsive organisms governed by intricate physical and biological systems.

The next challenge is establishing how widely those acoustic phenomena occur and whether they can support practical environmental decisions.

Until then, the central discovery remains extraordinary without embellishment: stressed plants emit detectable sounds that reveal something about their condition.

The world around us was never truly silent; our ability to listen was simply more limited than imagined.