Chapter 11 · Plants that Feel · Activity 11.3

Sound and Vibration

Does a plant respond to sound? The honest answer demands that you first rule out the biggest error — that the sound hits not the plant but your cable.

Duration 90 min Difficulty medium–high Group in pairs Prerequisite Activity 11.1 H

In a nutshell

What: You play the plant tones of different pitches and look for a reaction in the voltage curve. The focus is on a clean control: you first prove that you are measuring the plant and not the vibration of the loudspeaker coupling into the cable.

You need: the sensor from 11.1, a laptop, a phone or small loudspeaker, a free tone-generator app and — for the control — a cut-off leaf or a piece of damp kitchen paper.

What it's about

Whether plants "hear music" is a popular and delicate subject. Some studies report that flowers produce more nectar at certain frequencies; others find nothing. The reason for the dispute is almost always the same: sound is vibration, and vibration couples easily into thin cables and electrodes. A deflection on a tone therefore proves nothing about the plant reacting — perhaps only your measurement rig is wobbling in time.

So we turn the task around. The first step is not "Does the plant react to tones?" but: "How strongly does the sound couple into my rig, with no living plant at all?" Only once you know that can you judge at all whether more happens on the living leaf.

Before you start

The sensor from 11.1 must run calmly. Always keep the loudspeaker at the same distance and in the same place — otherwise you are comparing apples with pears. Not too loud: 60–70 dB is enough, that is normal room volume.

A little background

Why frequency matters. Your RC filter from 11.1 only lets slow signals below ~16 Hz through. An audible tone (100–8000 Hz) is therefore heavily damped by the filter — as an electrical signal. What does get through is a mechanical effect: when the sound wave physically sets the leaf or the cable in motion, the contact resistance of the electrode changes, and that produces a slow signal the filter lets through. Exactly this mechanical coupling is the trap.

What a real plant reaction would be. A physiological response would come delayed and persist briefly even after the tone is switched off — unlike the mechanical coupling, which begins and ends exactly with the tone. This difference in timing is your best clue.

Experiment A — the control first

This is the most important measurement of the day. It comes before the actual experiment.

Build the dead control. Instead of on a living leaf, stick the electrode on a cut-off leaf or a piece of damp kitchen paper — something that conducts but does not live. The setup, cable, loudspeaker distance: all exactly as later.
Give a tone. Play a 440 Hz tone (concert A) for 10 seconds, then 10 seconds of silence, then the tone again. Watch the curve.
Record the coupling. If the curve reacts to the tone on the dead object, you are measuring pure mechanical coupling. Note how strong. You must exceed this value later, otherwise there is "nothing new" on the living leaf.
Minimise the coupling. If the dead control reacts clearly, damp it: put the loudspeaker on a soft cloth, increase the distance, fix the cable with tape. Repeat until the dead control shows as little as possible.

Experiment B — the living plant

Back to the living leaf. Stick the electrode back on the plant, exactly the same setup as in the control. Let it settle for two minutes.
Go through the frequencies. Play 200 Hz, 440 Hz, 1000 Hz and 3000 Hz in turn, 10 seconds each with 20 seconds' pause between. Mark each tone start with the BOOT button (sketch from 11.2).
Check the timing. Notice not just whether but when the curve reacts: exactly with the tone (suspicious — coupling) or delayed and lingering (more likely physiological)?
Compare. Is the deflection on the living leaf larger than in the dead control at the same frequency? Only this excess may count as a possible plant reaction.

Tip: one frequency, one thing

Change only one quantity per run. If you turn the volume up and change the frequency at once, you won't know afterwards what the curve reacted to. One factor at a time — that is slower, but it is the only path to a clear result.

Worksheet

Separating coupling from reaction

How strongly did the dead control react to the 440 Hz tone (no deflection / small / clear)? What does that mean for your later interpretation?
At which frequency — if any — was the deflection on the living leaf larger than in the control?
Describe the timing: did the deflection begin exactly with the tone or delayed? Did it linger after the tone or break off at once?
Why is it not enough to simply turn the volume up to get "a stronger signal"?
A classmate says: "The plant likes Mozart, it deflected." Put politely which two controls he is still missing.

Show solution

1. If the dead control already shows a deflection, that is pure mechanical coupling. Then only the excess over this value on the living leaf may count as a reaction — the base deflection is an artefact, not physiology.

2. Individual. What matters is the phrasing "larger than the control at the same frequency and volume". Without this comparison the number is worthless.

3. Mechanical coupling starts and ends exactly with the tone. A physiological response sets in rather delayed and lingers — this lingering is the strongest hint of a real plant reaction.

4. Louder also makes the coupling stronger, not just a possible plant reaction. The ratio of signal to artefact does not improve — you enlarge both at once.

5. Missing are: (a) the dead control (does the sound couple in even without a living plant?) and (b) repetition/other groups (does the same come out reproducibly?). Only then does "likes Mozart" become a finding — and even then "likes" stays an interpretation.

When it sticks

Problem	Likely cause & fix
Dead control deflects just as strongly as the living leaf	Classic mechanical coupling — your result would be worthless. Decouple the loudspeaker (soft cloth), increase distance, fix the cable, turn it down.
Curve reacts only when the tone switches on and off	That is the loudspeaker "click" (on/off pulse), not the tone itself. Have the app fade the tone slowly in and out.
No reaction at all, not even mechanical	Tone too quiet or the filter damps too much. A little louder, reduce distance — but re-check the control afterwards.
Curve gets generally restless whenever music plays	The room/table is vibrating along. Put the sensor on a heavy, decoupled base (a stack of books, foam).

Food for thought

This activity teaches perhaps the most important lesson of the whole book: the greatest enemy of a measurement is the setup itself. A clean artefact looks just as convincing as a real effect — only the control separates the two.
Research works seriously on exactly this question: whether a measured reaction is physiological or just coupled-in vibration is decided by the timing and the comparison with dead controls. Here you do in miniature what labs do at scale.
Whoever takes this activity seriously will, at every headline "plants hear/feel/like X", ask first: where was the control? That question is the real learning goal.

Extension

Frequency sweep: have the app glide slowly from 100 to 5000 Hz. Is there a band in which the living plant reacts clearly more than the control? Such "resonances" would be exciting — but only robust with a dead control at each frequency.
Vibration instead of sound: put the sensor directly on the loudspeaker (only vibration, barely airborne sound) versus next to it (only airborne sound). Which route produces more coupling? That shows you physically where the artefact comes from.
Data export: copy the numbers from the Serial Monitor into a spreadsheet and plot the living plant and the control on top of each other. The gap between the two curves is your actual result.