A plant in a greenhouse may seem silent. But in a 2023 study, researchers recorded tomato and tobacco plants producing brief sounds at frequencies too high for most people to hear. Plants that were cut or short of water made more sounds than unstressed plants did.
The finding is intriguing not because it shows that plants speak, but because it suggests their physical condition may leave an acoustic trace in the air. If further research confirms the result in other settings and species, listening could one day help researchers monitor plant stress. For now, the work is an early step—not a ready-made diagnostic tool and not evidence of plant conversation.
What the researchers recorded
The team, led by researchers at Tel Aviv University, studied tomato and tobacco plants under controlled conditions. They recorded plants that were watered normally, plants experiencing drought stress, and plants whose stems had been cut. The sounds were captured with equipment able to detect ultrasound, including frequencies above the range of ordinary human hearing.
The stressed plants produced more detectable sounds than the control plants. The team also used machine-learning models to analyze acoustic patterns and distinguish among recordings associated with different plant conditions. That does not mean a machine can reliably diagnose any plant in a garden: the study tested a limited set of plants and conditions, and its results depend on the recording and analysis methods used.
Importantly, the researchers did not need to touch a plant with a microphone to measure the sounds. Their work examined airborne recordings, an important distinction from vibrations measured directly on a stem. The study therefore raises a question about what information might be carried by sound travelling through the surrounding air.
A possible source: water under tension
Plants move water through networks of vessels. When water is scarce, the water column inside those vessels can come under tension. One possible explanation for some of the recorded sounds is cavitation: the formation or movement of tiny air bubbles as water transport is disrupted. Such events can produce vibrations. The study’s findings are consistent with a physical process of this kind, but they do not establish a complete explanation for how each sound is made.
That caution matters. A recording can show that a sound occurred and that its pattern differed between experimental conditions. It cannot, by itself, settle the precise biological mechanism behind every click. Follow-up experiments would need to connect acoustic events to processes inside the plant, while ruling out other sources of noise.
Not a language—and not necessarily a message
Sound is often associated with communication, but the presence of a sound does not prove that it evolved to send a message. A stressed plant may make noise as a by-product of changes in its water transport or tissues. Demonstrating communication would require evidence that another organism detects the sound and responds to it in a way that benefits the sender or receiver.
The 2023 study did not show that plants intentionally make sounds, that they hear one another, or that insects and other animals use these particular recordings as cues. Those are open questions for future research, not conclusions the experiment established. The distinction keeps a striking result from becoming a more dramatic claim than the evidence supports.
Why listening could still be useful
Even without language, a sound associated with stress could be useful information. If the signal can be detected consistently under ordinary growing conditions, acoustic monitoring might complement other ways of assessing plants, such as checking soil moisture or examining leaves. In principle, a sound-based system could help researchers study when plants respond to drying or injury.
But practical use is a long way off. Greenhouses and fields contain wind, machinery, insects, and many other sounds. A tool would have to separate plant signals from that background and work across varieties, growth stages, and environments. It would also need to show that its predictions are accurate enough to guide a real decision, such as when to water. A controlled experiment is a starting point, not a field-tested product.
The broader lesson is that plant biology can be investigated through signals people do not normally perceive. Ultrasound microphones and careful analysis make it possible to ask whether a plant’s condition is reflected in the air around it. The answer so far is promising but limited: in the tested tomato and tobacco plants, stress was associated with more recorded sounds. What those sounds mean, how widely the pattern applies, and whether other organisms respond remain questions for further study.
So the next time a plant is described as “speaking,” it is worth asking what the evidence actually shows. In this case, researchers recorded ultrasonic sounds linked to experimental stress—not words, intentions, or a conversation. That is a fascinating result on its own, and a useful reminder that the boundary between silence and sound depends partly on what our instruments can detect.
