Emoto's Water Crystals: What the Photographs Actually Show

The famous images of "thanked" and "insulted" water are real photographs of real ice. What they are not is evidence that words change water — and the reason is a single step in the method that almost no retelling mentions. Here is how the pictures were made, what the one double-blind test found, where the older "memory of water" claim came from, and what is left that is worth keeping.

What the claim is

Masaru Emoto (1943–2014) was a Japanese author and businessman who ran the IHM Corporation in Tokyo. Beginning with Messages from Water in 1999 and reaching a very large American audience with The Hidden Messages in Water in 2004, he presented a simple and extraordinarily appealing proposition: expose water to a word, a piece of music, or a prayer, freeze it, and the ice crystals that form will be beautiful and symmetrical if the exposure was kind — and malformed if it was not.

The doctorate Emoto used professionally came from the Open International University for Alternative Medicine, an institution that is not accredited as a degree-granting university. That does not settle whether the claim is true, but it matters when the claim is presented as laboratory science.

How the photographs were made

The procedure, as described by Emoto's own laboratory, ran roughly like this:

1. Water is placed in something on the order of fifty to a hundred small petri dishes.
2. The dishes are frozen — in the published descriptions, at around −25 °C for a few hours.
3. They are moved to a cold room near −5 °C and examined under a microscope with a camera attached.
4. A photographer looks through the batch and photographs the crystals that stand out.
5. The selected images are published next to the word the sample was exposed to.

Step four is where the whole thing turns. In any batch of frozen droplets you will find a wide spread of forms — some near-perfect six-fold crystals, many irregular blobs, plenty of partial structures. That spread exists in every batch, regardless of what anyone said to it. Given fifty dishes and a person who knows this sample is the "thank you" sample, finding a beautiful crystal in it is not a discovery. It is a search.

This is not an accusation of fraud, and it does not need to be. It is the reason blinding was invented. Unblinded selection from a variable population will produce the expected result reliably, in good faith, every time.

What actually decides the shape of an ice crystal

The physics here is old and well settled, which is what makes the claim so hard to sustain.

In the 1930s and 40s the Japanese physicist Nakaya Ukichiro grew snow crystals under controlled laboratory conditions and mapped which shapes appear at which combinations of temperature and humidity. The resulting Nakaya diagram is still the foundation of the field: plates near −2 °C, needles near −5 °C, columns near −10 °C, the classic branched dendrites around −15 °C. Change the temperature by a couple of degrees and you change the shape of the crystal completely.

Variable	Effect on the crystal	Controlled in Emoto's photographs?
Temperature at formation	Decides the basic habit — plate, needle, column, dendrite	Not reported per dish
Supersaturation / humidity	Decides how branched and elaborate it becomes	Not reported
Freezing rate	Fast freezing gives disordered ice; slow gives structure	Not reported
Dust and nucleation sites	Determine where and how crystals start at all	Not reported
Dissolved minerals	Alter growth substantially; tap, spring and distilled water differ	Source water varies between comparisons
Time under the microscope lamp	Crystals sublimate and degrade visibly within minutes	Not reported
Which crystal gets photographed	Decides the published result entirely	No — chosen by an unblinded observer

Six known, powerful variables were left uncontrolled, and the seventh — selection — was performed by someone who knew the answer. Against that, "the word on the label" does not need to be invoked to explain anything.

The one double-blind test

To Emoto's credit, he took part in an attempt to do this properly. The result was published in 2006 in Explore: The Journal of Science and Healing by Dean Radin, Nancy Hayssen, Masaru Emoto and Takashige Kizu.

The design: water was sealed in an electromagnetically shielded room in California. Around two thousand people in Austria, who had never seen the samples, directed intention toward them. Control water sat elsewhere in the same building. The samples were then frozen and photographed, and — the important part — independent analysts who did not know which was which rated the crystals for aesthetic appeal.

The reported outcome was a small but statistically significant difference: crystals from treated water were rated slightly more appealing than controls. That is a genuinely better experiment than anything in the books, and it should be described accurately rather than waved away. Three things limit what it can carry:

• The measured quantity is beauty. Not lattice geometry, not a spectroscopic signature — how pleasing a panel found the picture. That is a legitimate thing to measure and a weak thing to build a physical claim on.
• It has not been independently replicated. A follow-up by the same group reported mixed results. No separate laboratory has reproduced it.
• Selection still enters upstream. Blinding the raters does not blind the step where crystals are chosen for photography in the first place.

A single small unreplicated positive in a complementary-medicine journal is where a hypothesis begins, not where it ends.

The older claim underneath: the memory of water

Emoto's work sits on top of a much-discussed earlier episode. In 1988 the French immunologist Jacques Benveniste published a paper in Nature reporting that antibody solutions diluted so far that no molecule of the original substance could remain still produced a biological effect — water, the interpretation went, had retained an imprint.

Nature published it with an unusual editorial note reserving judgment, then sent an investigating team to the laboratory. Under properly blinded conditions, with the sample codes concealed, the effect did not appear. Benveniste's results were not reproduced by other laboratories, and his subsequent claims that the imprint could be recorded and transmitted electronically fared no better.

The physical objection has never gone away. Liquid water is a network of hydrogen bonds that continuously breaks and reforms on a timescale of picoseconds — trillionths of a second. There is no known structure in which a pattern could persist for the hours or days these claims require.

The rice experiment — why it convinces, and why it shouldn't

This is the version most people have actually tried: three sealed jars of cooked rice, labeled thank you, you fool, and one ignored entirely. After a few weeks the thanked jar is said to ferment pleasantly while the insulted one turns black.

It is a persuasive demonstration because the person running it does the work themselves and sees the result with their own eyes. It also has no controls at all.

• The jars are not handled identically. The thanked jar gets opened, spoken into, breathed on, moved into the light. Each of those introduces different microbes and different moisture.
• Whether rice ferments or rots depends on which organisms get established first — a matter of initial contamination, salt, moisture, sealing and temperature, all of which vary between three jars on a shelf.
• The judgment is subjective and unblinded. "Pleasant" versus "spoiled" is assessed by the person who wants the result.
• Negative runs are rarely reported. People whose thanked jar rotted mostly conclude they did it wrong.

If you want to run it as an actual experiment: have someone else label the jars in a code you cannot read, keep them sealed and untouched in identical conditions, speak to them via a recording that plays to all three, and have a third party who does not know the code judge the outcome. Almost no one who reports success has done this.

Why the images are still powerful

It would be easy to stop at "debunked," and that would miss the interesting question — why this particular claim spread so far when many stranger ones did not.

Three reasons, and none of them is stupidity. The photographs are genuinely beautiful, and beauty reads as evidence. The claim encodes something people already know to be true — that the words directed at a person change them — and offers it back in a form that appears to have been measured. And water is the right symbolic substance for it: we are largely made of it, we share it, it takes the shape of whatever holds it.

That last point is often stated as an argument: the body is 70 percent water, therefore words must affect us the way they affect the dishes. The figure is closer to 60 percent in adults, and the inference does not follow in any case. But the intuition it is reaching for — that language lands in the body — is not superstition. It is simply evidence about people, and it should be argued from studies of people.

What to keep

We do not claim any of the following changes the water. Each of them changes the person, which is the only mechanism here that has ever been demonstrated.

• A glass of water as an attention marker. One glass, one minute, once a day, with nothing else happening. This is a small piece of contemplative practice wearing a costume.
• Saying the thing out loud. Speaking an intention over an object is an old and effective way of making it concrete to yourself — and it is the same mechanism as the full moon and new moon practices.
• Noticing what you say to yourself. The most defensible part of Emoto's message had nothing to do with ice, and you can act on it without any equipment.

What we will not tell you is that charged water heals anything, that it is structurally different, or that it should replace medical care. If you want the related practice of putting stones in water, read gem water and how to make it safely first — several popular stones are genuinely unsafe in water, and that is a fact about chemistry rather than a fact about belief.

Frequently asked questions

So the photographs are fake?

No. They are real photographs of real ice crystals. The step that fails is choosing which of the fifty-odd dishes to publish, done by someone who knows what the answer is supposed to be.

Didn't a double-blind study confirm it?

One study (2006) reported a small effect on how appealing blinded raters found the crystals. It has not been independently replicated, and aesthetic rating is not a measurement of the water.

Is "structured water" the same idea?

It is a commercial relative of it. Products sold as structured, hexagonal or energized water rest on the same premise — a durable arrangement of water molecules — and there is no accepted evidence that such an arrangement survives in liquid water.

What about homeopathy?

Homeopathy requires the same water-memory premise, which is what made the 1988 Nature episode significant. That premise remains unsupported.

My rice experiment worked. What does that mean?

Most likely that the jars were handled differently and colonized by different organisms. Run it with a third party coding the labels and judging the result, and see whether it survives.

Is it wrong to speak kindly to my water?

Not at all. It is a small, harmless act of attention, and there is a reasonable case for it as practice. The only thing worth avoiding is believing the water is what changed.

Related reading

• Gem water, made safely — which stones must never go in water
• Crystals that cannot go in water
• Real versus treated crystals — another case where the honest answer is testable
• How to make moon water

Sources

• Radin, D., Hayssen, G., Emoto, M. & Kizu, T. (2006). Double-blind test of the effects of distant intention on water crystal formation. Explore: The Journal of Science and Healing, 2(5), 408–411.
• Davenas, E. et al. (1988). Human basophil degranulation triggered by very dilute antiserum against IgE. Nature, 333, 816–818 — and the accompanying Nature editorial reservation and subsequent site investigation.
• Nakaya, U. (1954). Snow Crystals: Natural and Artificial. Harvard University Press — the temperature/supersaturation morphology diagram.
• Emoto, M. (1999). Messages from Water; (2004) The Hidden Messages in Water.

Questions about an order or a product? Write to 28@templeofdesire.com.