Journal Club – “The history of the Memory of Water”

January 1st, 2000 by Ben Goldacre in journal club | 2 Comments »

This is part of the Homeopathy journal club described here:

doi:10.1016/j.homp.2007.03.006    How to Cite or Link Using DOI (Opens New Window)
Copyright © 2007 Elsevier Ltd All rights reserved. The history of the Memory of Water

Yolène ThomasCorresponding Author Contact Information, a, E-mail The Corresponding Author
aInstitut Andre Lwoff IFR89, 7, rue Guy Moquet-BP8, 94 801 Villejuif Cedex, France
Received 26 March 2007;  accepted 27 March 2007.  Available online 31 July 2007.

‘Homeopathic dilutions’ and ‘Memory of Water’ are two expressions capable of turning a peaceful and intelligent person into a violently irrational one,’ as Michel Schiff points out in the introduction of his book ‘The Memory of Water’. The idea of the memory of water arose in the laboratory of Jacques Benveniste in the late 1980s and 20 years later the debate is still ongoing even though an increasing number of scientists report they have confirmed the basic results.

This paper, first provides a brief historical overview of the context of the high dilution experiments then moves on to digital biology. One working hypothesis was that molecules can communicate with each other, exchanging information without being in physical contact and that at least some biological functions can be mimicked by certain energetic modes characteristics of a given molecule. These considerations informed exploratory research which led to the speculation that biological signaling might be transmissible by electromagnetic means. Around 1991, the transfer of specific molecular signals to sensitive biological systems was achieved using an amplifier and electromagnetic coils. In 1995, a more sophisticated procedure was established to record, digitize and replay these signals using a multimedia computer. From a physical and chemical perspective, these experiments pose a riddle, since it is not clear what mechanism can sustain such ‘water memory’ of the exposure to molecular signals. From a biological perspective, the puzzle is what nature of imprinted effect (water structure) can impact biological function. Also, the far-reaching implications of these observations require numerous and repeated experimental tests to rule out overlooked artifacts. Perhaps more important is to have the experiments repeated by other groups and with other models to explore the generality of the effect. In conclusion, we will present some of this emerging independent experimental work.

Keywords: high dilution; memory; water; molecular signal; audio-frequency oscillator; computer-recorded signals

Article Outline

Historical overview: the early history of high dilution experiments
Exploring the physical nature of the biological signal
From high dilution to digital biology
The present situation

Historical overview: the early history of high dilution experiments

Presenting a brief history of what is known as the ‘Memory of Water’ is not an easy task mainly because one of the main actors, Jacques Benveniste, is no longer with us (Figure 1). There are always many controversies around cutting edge science, and especially with those whose lives have been spent pursuing unorthodox trails.

Display Full Size version of this image (55K)

Figure 1. Jacques Benveniste 1935–2004.

I first met Benveniste during a FASEB meeting in Atlanta in 1981 and joined his laboratory a few years later to set up my own Immunology team. I had the good fortune of being able to collaborate with him for over 16 years. At that time, he was at the top of his fame and gained an international reputation as a specialist on the mechanisms of allergies and inflammation with his discovery of the ‘Platelet Activating Factor’ (paf-acether) in 1970.[1] and [2] Throughout his long career, working both in the US and in France, he was responsible for the development of new ways of approaching inflammation including the patenting by the French National Institute of Health and Medical Research (INSERM) of his innovative allergy test using blood cells called basophils (FR-patent-7,520,273). Jacques’ research into allergies took him deep into the mechanisms which create such responses: understanding how the smallest amount of a substance affects the organism. The life and work of Jacques Benveniste was not only written in water.

In the early 1980s, while heading up the unit INSERM 200, Jacques took a new member onto his staff, a young medical doctor, Bernard Poitevin, whose side-interest was homeopathy. ‘He asked me if he could try my basophil degranulation test on some homeopathic preparations’, Jacques recalled, ‘and I remember distinctly saying “OK, but all you will be testing is water”.’ Thus, Jacques expressed his skepticism but accepted the proposal.

After 5 years of research they empirically observed that highly dilute (i.e., in the absence of any physical molecule) biological agents nevertheless triggered the relevant biological systems. Intrigued but cautious, Jacques was a man who adhered to the facts. He ordered a two-year long series of retests, but the same results kept recurring. Finally, Poitevin and Benveniste submitted two papers which were published in peer review journals.[3] and [4] Here, the work was treated as conventional research like many other manuscripts from peer-reviewed journals which can be found in the scientific literature on the effect of high dilutions (HD) (review in[5] and [6]).

Following accepted scientific practice, Jacques then asked other laboratories to try to replicate the findings. In 1988, scientists from six laboratories in four countries (France, Canada, Israel and Italy) co-authored an article showing that highly diluted antibodies could cause basophil degranulation. This was established under stringent experimental conditions such as blind double-coded procedures. Further, the experimental dilution (anti-IgE) and the control one (anti-IgG) were prepared in exactly the same manner, with the same number of dilution and agitation sequences. The article was submitted to Nature.7 Nature‘s referees could not fault Benveniste’s experimental procedures but could not comprehend his results. How can a biological system respond to an antigen when no molecules of it can be detected in the solution? It goes against the accepted ‘lock-and-key’ principle, which states that molecules must be in contact and structurally match before information can be exchanged. In the paper, Jacques suggested that specific information must have been transmitted during the dilution/shaking process via some molecular organization occurring in the water.

Finally, the editor of the journal, John Maddox agreed to publication, on condition that a ‘committee’ could verify Benveniste’s laboratory procedures. In July 1988, after two weeks after publication, instead of sending a committee of scientific experts, Maddox recruited—James Randi, a magician, and Walter Stewart, a fraud investigator. The three of them spent 5 days in the laboratory. Well, you all know what followed. Nature‘s attempted debunking exercise failed to find any evidence of fraud. Nevertheless, they concluded that Benveniste had failed to replicate his original study.8 This marked the beginning of the ‘Water Memory’ war, which placed him in a realm of ‘scientific heresy’. As Michel Schiff later remarked in his book: ‘INSERM scientists had performed 200 experiments (including some fifty blind experiments) before being challenged by the fraud squad. The failure to reproduce8 only concerned two negative experiments’.9 Benveniste replied to Nature10 and reacted with anger, ‘not to the fact that an inquiry had been carried out, for I had been willing that this be done… but to the way in which it had been conducted and to the implication that my team’s honesty and scientific competence were questioned. The only way definitely to establish conflicting results is to reproduce them. It may be that we are all wrong in good faith. This is not crime but science…’.

As a consequence of the controversy that ensued, Jacques became increasingly isolated. Nonetheless the team repeated the work on a larger scale, entirely designed and run under the close scrutiny of independent statistical experts, and confirmed the initial findings in Nature.11 These further experiments have been coolly received or ignored by most scientists at least partly because, given Jacques’ now-acrimonious relationship with Nature, they were published in a less renowned journal.

To date, since the Nature publication in 1988, several laboratories have attempted to repeat Benveniste’s original basophil experiments. Most importantly, a consortium of four independent research laboratories in France, Italy, Belgium, and Holland, led by M. Roberfroid at Belgium’s Catholic University of Louvain in Brussels, confirmed that HD of histamine modulate basophil activity. An independent statistician analyzed the resulting data. Histamine solutions and controls were prepared independently in three different laboratories. Basophil activation was assessed by flow-cytometric measurement of CD63 expression (expressed on cytoplasmic granules and on the external membrane after activation). All experiments were randomized and carried out under blind conditions. Not much room, therefore, for fraud or wishful thinking. Three of the four labs involved in the trial reported statistically significant inhibition of the basophil degranulation reaction by HD of histamine as compared to the controls. The fourth lab gave a result that was almost significant. Thus, the total result over all four labs was positive for histamine HD solutions.[12] and [13] ‘We are,’ the authors say in their paper, ‘unable to explain our findings and are reporting them to encourage others to investigate this phenomenon’.

Different attempts have been made to substantiate the claim that serial dilution procedures are associated with changes in the water’s physical properties ([14] and [15]and see Louis Rey contribution in this issue pages 170–174). Yet, the challenge of understanding the mechanisms of how HDs work, and the role of water in them, is a difficult one to say the least. Several possible scenarios have been suggested. One proposed by Giuliano Preparata and Emilio Del Giudice, is that long range coherent domains between water molecules (quantum electrodynamics, QED) gives high dilution laser-like properties.[16] and [17] When the field matches the kinetic of the reaction, the latter becomes functional as the optimal field strength as for a radio receiver. It was to a scientific meeting in Bermuda that took place a few months before the Nature ‘affair’ erupted that these two physicists working at Milan University brought the theoretical basis for the memory of water. Another scenario predicts changes in the water structure by forming more or less permanent clusters.18 Other hypotheses will be discussed in this issue. High dilution experiments and memory water theory may be related, and may provide an explanation for the observed phenomena. As M. Schiff points out, only time and further research will tell, provided that one gives the phenomena a chance.9

Exploring the physical nature of the biological signal

Despite the difficulties after the Nature fracas, Jacques and his now-depleted research team continued to investigate the nature of the biological activity in high dilutions and aimed at understanding the physical nature of the biological signal. In his Nature paper, Jacques reasoned that the effect of dilution and agitation pointed to transmission of biological information via some molecular organization going on in the water. The importance of agitation in the transmission of information was explored by pipetting dilutions up and down ten times and comparing with the usual 10-s vortexing. Although the two processes resulted in the same dilution, basophil degranulation did not occur at HD after pipetting. So transmission of the information depended on vigorous agitation, possibly inducing a submolecular organization of water or closely related liquids (ethanol and propanol could also support the phenomenon). In contrast, dilutions in dimethylsulphoxide did not transmit the information from one dilution to the other. In addition, heating, freeze-thawing or ultrasonication suppressed the activity of highly diluted solutions, but not the activity of several active compounds at high concentrations. A striking feature was that molecules reacted to heat according to their distinctive heat sensitivity, whereas all highly diluted solutions ceased to be active between 70 and 80 °C. This result suggested a common mechanism operating in HDs, independent of the nature of the original molecule. In addition, in 1991 and in collaboration with an external team of physicists (Lab. Magnetisme C.N.R.S.-Meudon Bellevue, France), it was shown in twenty four blind experiments that the activity of highly dilute agonists was abolished by exposure to a magnetic field (50 Hz, 15×10−3 T, 15 min) which had no comparable effect on the genuine molecules. Moreover, it is worth pointing out that a growing number of observations suggest the susceptibility of biological systems or water to electric and low-frequency electromagnetic fields.[19], [20] and [21] In addition, what is suggested from the literature is a possible role of electromagnetic fields regarding informational process in cell communication.[22], [23] and [24]

At this stage, Jacques hypothesized that transmission of this ordering principle was electromagnetic in nature and move on to the idea that molecules could communicate via specific electromagnetic waves. If so, what molecule vibration modes are efficient and how can these signals be used to mimic some of the biological functions of a molecule without its physical presence?

From high dilution to digital biology

It was at the beginning of the nineties that a homeopathic physician, E. Attias convinced Jacques to try out an electrical device that he claimed transmitted chemical information. After a few positive trials with this machine, Jacques had another one built, which was used for later experiments. This second device was essentially a standard audio amplifier that, when connected to another coil, behaves as an audio-frequency oscillator. Between 1992 and 1996, we performed a number of experiments showing that we could transfer, in real time, molecular signals indirectly to water or directly to cells. Briefly, cells were placed in a 37 °C humidified incubator on one coil attached to the oscillator, while an agonist (or vehicle as control) was placed on another coil at room temperature. Here, the transfer was not a two step-process, as when water acts as an intermediary recipient of the molecular signal. In one such exploration, we showed that molecular signals associated with a common phorbol ester (phorbol-myristate-acetate) could be transmitted by physical means directly to human neutrophils to modulate reactive oxygen metabolite production. In 1996, I submitted an article about these experiments to several prestigious journals. The article was flatly rejected each time, on the grounds that we could not explain the underlying mechanism, in spite of the referees’ general opinions that our work was ‘state-of-the-art’ and was ‘provocative and intriguing and we have gone to great lengths to try to eliminate any biological variables that could bias our results.’ It was finally published in 2000.25 Appended to this article were two affidavits, one from a French laboratory (F. Russo Marie, INSERM U332, Paris, France) testifying that they supervised and blinded the experiments we did in this laboratory; the other from an US laboratory (W. Hsueh, Department of Pathology, Northwestern University, Chicago) testifying that they did some preliminary experiments similar to ours, without any physical participation on our part, and detected the same effect.

Because of the material properties of the oscillator and the limitations of the equipment used, it is most likely that the signal is carried by frequencies in the low kilohertz range.26 These considerations led to the establishment in 1995 of a more sophisticated procedure for the recording and retransmission of the molecular signals. DigiBio, a company that Jacques had set up in 1997 to finance his research, obtained in 2003 an approval for one of his French patents by the US Patent Office (6,541,978: method, system and device for producing signals from a substance biological and/or chemical activity). The characteristics of the equipment are described in Figure 2 and in.26 Briefly, the process is to first capture the electromagnetic signal from a biologically active solution using a transducer and a computer with a sound card. The digital signals are stored (Microsoft sound files *.wav). The signal is then amplified and ‘played back’, usually for 10 min, from the computer sound card to cells or organs placed within a conventional solenoid coil. The digitally recorded signals can also be played back into untreated water, which thereafter will act as if the actual substance was physically present.

Display Full Size version of this image (29K)

Figure 2. Schematic drawing of the computer-recorded signals: capture, storage and replay:

Shielded cylindrical chamber: Composed of three superposed layers: copper, soft iron, permalloy, made from sheets 1 mm thick. The chamber has an internal diameter of 65 mm, and a height of 100 mm. A shielded lid closes the chamber.

Transducers: Coil of copper wire, impedance 300 Ω, internal diameter 6 mm, external diameter 16 mm, length 6 mm, usually used for telephone receivers.

• Multimedia computer (Windows OS) equipped with a sound card (5–44 KHz in linear steps), (Sound Blaster AWE 64, CREATIVE LABS).

• HiFi amplifier 2×100 watts with an ‘in’ socket, an ‘out’ socket to the speakers, a power switch and a potentiometer. Pass band from 10 Hz to 20 kHz, gain 1–10, input sensitivity +/− V.

Solenoid coil: Conventionally wound copper wire coil with the following characteristics: internal diameter 50 mm, length 80 mm, R=3.6 Ω, three layers of 112 tums of copper wire, field on the axis to the centre 44×10−4 T/A, and on the edge 25×10−4 T/A.

All links consist of shielded cable. All the apparatus is earthed.

From 1995 to the present, several biologically active molecules (eg histamine, acetylcholine, caffeine, PMA, Melagatran… even homeopathic medicines such as Arnica montana) have been recorded, digitized and replayed to biological systems sensitive to the original molecular substance. Several biological models were used. The first one was a commonly used system by pharmacologists, called the Langendorff preparation. By injecting different vasoactive substances into the coronary artery of an isolated, perfused guinea pig heart and measuring the coronary flow, you can quantify the vasoconstricting or vasodilating effect of the agent. In typical experiments, the signal of acetylcholine (or water as control), a classical vasodilating molecule was recorded and digitized. The signal was then amplified and ‘played’ back onto water. The signal-carrying water is then injected into the isolated heart, and consequently the coronary flow increased. Interestingly, atropine, an acetylcholine inhibitor, inhibited both the effects of the molecular acetylcholine as well as the digital signal of acetylcholine. Of note, the order of the conditions and their repetitions was always randomized and blinded. Other models include: human neutrophil activation; detection of the recorded signal of bacteria (E. Coli and Streptococcus) by playing them to a biological system specific to the bacterial signal and; the inhibition of fibrinogen coagulation by a Direct Thrombin Inhibitor. Further details of three of these salient biological models have been previously described.26 Together, these results suggested that at least some biologically active molecules emit signals in the form of electromagnetic radiation at a frequency of less than 44 kHz that can be recorded, digitized and replayed directly to cells or to water, in a manner that seems specific to the source molecules.26

Assuming that we give credence to the phenomena described, one question naturally springs to mind: what do molecule vibration modes sound like? Can measurable signals been identified in the form of low frequency spectral components? Didier Guillonnet, an engineer in computer science, and at the time, a close collaborator of Jacques Benveniste admitted, ‘When we record a molecule such as caffeine, for example, we should get a spectrum, but it seems more like noise. We are only recording and replaying; at the moment we cannot recognize a pattern although the biological systems do.’ Jacques called this matching of broadcast with reception ‘co-resonance,’ and said it works like a radio set.

Among the various theoretical problems associated with such a signal, two appear particularly relevant. First, how is such information using water as an intermediary detected amongst much electromagnetic noise? In fact, it has been suggested that stochastic resonance is an important mechanism by which very weak signals can be amplified and emerge from random noise.27 Second, the limitations of the equipment used here, suggest that the signal is carried by frequencies in the low kilohertz range, many orders of magnitude below those generally associated with molecular spectra (located in the infrared range). However, molecules may also produce much lower ‘beat’ frequencies (Hz to kHz) specific for every different molecule. The ‘beat frequency’ phenomenon may explain this discrepancy, since a detector, for instance a receptor, will ‘see’ the sum of the components of a given complex wave.28 Clearly, more experimental and theoretical work is needed in order to unveil the physical basis of the transfer (and storage?) of specific biological information either between interacting molecules or via an electronic device.

Replicability: Although since the very beginning we have placed a great deal of emphasis on carrying out our work under the highest standards of methodology and that great effort has been made to isolate it from environmental artifacts, attempts to replicate these data in other laboratories yielded mixed results. For instance, in 1999, Brian Josephson, Nobel Laureate for Physics in 1973 invited Benveniste to the Cavendish Laboratory in Cambridge. He said, ‘We invited him to learn more about the research which seems both scientifically interesting and potentially of considerable practical importance. Jacques definitely recognized there was a problem with reproducing the effect. The situation seemed to be that in some circumstances you had reproduction and in others you didn’t; but the overall results were highly significant.’ We then realized the difficulty in ‘exporting’ a method, which is very far from conventional biology. There are many key variables that might be involved like, water purification, the container shape and material being used, the purity of chemicals, atmospheric conditions…. Only if these underlying variables are known could the experiments be reproducible. When the transfer is a two-step process using water as an intermediary support for transmitted molecular signals, it takes even more stringent conditions for the experiments to be repeatable. The digital signal is replayed onto the water, which may take or not take the signal depending, for instance, upon the local electromagnetic conditions. In this regard, it is interesting to note that the ‘informed water’ as in the HD experiments, loses its activity after heating or being exposed to magnetic fields.

More surprising and mysterious was the fact that in some cases certain individuals (not claiming special talents) consistently get digital effects and other individuals get no effects or perhaps block those effects (particularly when handling a tube containing informed water). The inhibition of fibrinogen–thrombin coagulation by a digitized thrombin inhibitor is a model particularly sensitive to experimenter effects and therefore may account for the difficulty in consistently replicating this experimental system. Despite the precautions taken to shield the information transfer equipment from magnetic or electromagnetic pollution, very little concern has been given to possible subtle human operator effects.29 We dealt with this problem in some of our own studies and also in the course of one independent replication.30

The present situation

Now that Jacques Benveniste is no longer with us, the future of the ‘digital biology’ is in the hands of those who have been convinced of the reality of the basic phenomena. It is up to them to explore with other models the generality of the effect. Most likely they will succeed if they combine full biological and physical skills to understand the nature of the biological signals.

In this regard, since June 2005, Luc Montagnier, the co-discoverer of HIV, is conducting experiments (detection of the recorded signals of various micro-organisms derived from human pathologies) which, confirm and extend the original finding. In 2006, he set up a company called Nanectis. Perhaps the most impressive emerging data is from a US group located in La Jolla, CA.

In barely four years, they have conducted novel research programs and expanded the original technology into a series of potential industrial applications. Since 2004, they have obtained several US patents (6,724,188; 6,952,652; 6,995,558; 7,081,747) and applied for International Patents (WO 06/015038: system and method for collecting, storing, processing, transmitting and presenting very low amplitude signals; WO 06/073491: system and method for producing chemical or biochemical signals). They can improve the molecular signal recording in particular by using both magnetic and electromagnetic shielding coupled to a superconducting quantum interference device (SQUID). The system records a time-series signal for a compound; the wave form is processed and optimized (selected noise amplitude, power setting…) to identify low-frequency peaks that are characteristic of the molecule being interrogated (Molecular Data Interrogation System, MIDS). The optimized signal is played back for various periods of time to sensitive biological systems. For instance, they describe one interesting model particularly relevant to the specificity of the molecular signal transmission effect. The arabinose-inducible bacterial system with a lac operon is inducible by signals from the L (+) arabinose form but not from the D (−) arabinose inactive isomer or the white noise control. Other systems include digital herbicides and plant growth regulator as well as pharmaceutical compounds such as Taxol ®, a prototype for a class of anticancer drugs. For instance, in a classic in vivo mouse xenograft model, the digital Taxol was assessed by the growth inhibitory potential of a human breast tumor. The results revealed that tumor growth, by day 36, was as statistically significantly inhibited in the group treated with the Taxol signal, as it was in the control group treated with actual molecular Taxol. If these new experimental observations can be validated, we will have added yet another valuable piece to the puzzle.

Although a theoretical explanation of how the memory of water might work must still be explored, the fact that the effective transmission of molecular signals has now been observed by independent teams using different biological systems, provides a strong additional basis to suggest that the phenomena observed by Jacques were not due simply to laboratory artefacts.

Whatever knowledge ongoing and future investigation may bring, the difficult road that Jacques travelled by opposing the automatic acceptance of received ideas, will have contributed to sustaining freedom in scientific research and putting the emphasis back where it belongs, on observable fact.


I am grateful to Drs. Isaac Behar and Anita K. Gold for critical comments on the manuscript.


1 J. Benveniste, P.M. Henson and C.G. Cochrane, Leukocyte-dependent histamine release from rabbit platelets. The role of IgE, basophils, and a platelet-activating factor, J Exp Med 136 (1972), pp. 1356–1377. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

2 J. Benveniste, Platelet-activating factor, a new mediator of anaphylaxis and immune complex deposition from rabbit and human basophils, Nature 249 (1974), pp. 581–582. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

3 E. Davenas, B. Poitevin and J. Benveniste, Effect of mouse peritoneal macrophages of orally administered very high dilutions of silica, Eur J Pharmacol 135 (1987), pp. 313–319. Abstract | Abstract + References | PDF (543 K) | View Record in Scopus | Cited By in Scopus

4 B. Poitevin, E. Davenas and J. Benveniste, In vitro immunological degranulation of human basophils is modulated by lung histamine and Apis mellifica, Br J Clin Pharmacol 25 (1988), pp. 439–444. View Record in Scopus | Cited By in Scopus

5 H. Walach, W.B. Jonas, J. Ives, R. van Wijk and O. Weingartner, Research on homeopathy: state of the art, J Altern Complement Med 11 (2005), pp. 813–829. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

6 P. Bellavite, R. Ortolani, F. Pontarollo, V. Piasere, G. Benato and A. Conforti, Immunology and Homeopathy, Evidence-based Complementary Alternative Med 2 (2005), pp. 441–452. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

7 E. Davenas, F. Beauvais and J. Amara et al., Human basophil degranulation triggered by very dilute antiserum against IgE, Nature 333 (1988), pp. 816–818. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

8 J. Maddox, J. Randi and W.W. Stewart, High-dilution’experiments a delusion, Nature 334 (1988), pp. 287–290.

9 Schiff M. The Memory of Water. UK: Ed. Thorsons, 1995.

10 J. Benveniste, Dr Jacques Benveniste replies, Nature 334 (1988), p. 291. Full Text via CrossRef

11 J. Benveniste, E. Davenas, B. Ducot, B. Cornillet, B. Poitevin and A. Spira, L’agitation de solutions hautement diluées n’induit pas d’activité biologique spécifique, CR Acad Sci Paris 312 (1991), pp. 461–466.

12 P. Belon, J. Cumps and M. Ennis et al., Inhibition of human basophil degranulation by successive histamine dilutions: results of a European multi-centre trial, Inflamm Res (Suppl 1) 48 (1999), pp. S17–S18. View Record in Scopus | Cited By in Scopus

13 P. Belon, J. Cumps and M. Ennis et al., Histamine dilutions modulate basophil activation, Inflamm Res 53 (2004), pp. 181–188. View Record in Scopus | Cited By in Scopus

14 Lobyshev VI, Tomkevitch MS. Luminescence study of homeopathic remedies. In: Priezzhev AV, Cote GL (eds). Optical Diagnostics and Sensing of Biological Fluids and Glucose and Cholesterol Monitoring, Proceedings of the SPIE, Vol 4263. MAIK “Navka/Interperiodica” (Russia), 2001, pp 1605–7422.

15 V. Elia, S. Baiano, I. Duro, E. Napoli, M. Niccoli and L. Nonatelli, Permanent physico-chemical properties of extremely diluted aqueous solutions of homeopathic medicines, Homeopathy 93 (2004), pp. 144–150. SummaryPlus | Full Text + Links | PDF (154 K) | View Record in Scopus | Cited By in Scopus

16 E. Del Giudice, G. Preparata and G. Vitiello, Water as a free electric dipole laser, Phys Rev Lett 61 (1988), pp. 1085–1088. Full Text via CrossRef

17 G. Preparata, QED Coherence in Matter, World Scientific, Singapore (1995).

18 E.E. Fesenko and A.Y. Gluvstein, Changes in the state of water, induced by radiofrequency electromagnetic fields, FEBS Lett 367 (1995), pp. 53–55. Abstract | Abstract + References | PDF (294 K) | View Record in Scopus | Cited By in Scopus

19 R. Goodman and M. Blank, Initial interactions in electromagnetic field-induced biosynthesis, J Cell Physiol 199 (2004), pp. 359–363.

20 E. Ben Jacob, Y. Aharonov and Y. Shapira, Bacteria harnessing complexity, Biofilms (2004), pp. 239–263.

21 P.h. Vallée, J. Lafait, P. Mentré, M.O. Monod and Y. Thomas, Effects of pulsed low frequency electromagnetic fields on water using photoluminescence spectroscopy: role of bubble/water interface?, J Chem Phys 122 (2005), pp. 114513–114521. Full Text via CrossRef

22 G. Albrecht-Buehler, Rudimentary form of cellular ‘vision’, Proc Natl Acad Sci USA 89 (1992), pp. 8288–8292. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

23 M.W. Trushin, Studies on distant regulation of bacterial growth and light emission, Microbiology 149 (2003), pp. 363–368. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

24 B.W. Ninham and M. Boström, Building bridges between the physical and biological sciences, Cell Mol Biol 51 (2005), pp. 803–813. View Record in Scopus | Cited By in Scopus

25 Y. Thomas, M. Schiff, L. Belkadi, P. Jurgens, L. Kahhak and J. Benveniste, Activation of human neutrophils by electronically transmitted phorbol-myristate acetate, Med Hypotheses 54 (2000), pp. 33–39. Abstract | Abstract + References | PDF (188 K) | View Record in Scopus | Cited By in Scopus

26 Y. Thomas, L. Kahhak and J. Aissa, The physical nature of the biological signal, a puzzling phenomenon: the critical role of Jacques Benveniste. In: G.H. Pollack, I.L. Cameron and D.N. Wheatley, Editors, Water and the Cell, Springer, Dordrecht (2006), pp. 325–340.

27 K. Wiesenfeld and F. Moss, Stochastic resonance and the benefits of noise: from ice ages to crayfish and SQUIDS, Nature 373 (1995), pp. 33–36. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

28 C.N. Banwellk, Fundamentals of Molecular Spectroscopy, McGraw-Hill Publ., UK (1983) pp 26–28.

29 B.J. Dunne and R.G. Jahn, Consciousness, information, and living systems, Cell Mol Biol 51 (2005), pp. 703–714. View Record in Scopus | Cited By in Scopus

30 W.B. Jonas, J.A. Ives and F. Rollwagen et al., Can specific biological signals be digitized?, FASEB J 20 (2006), pp. 23–28. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

Corresponding Author Contact InformationCorrespondence: Yolène Thomas, Institut Andre Lwoff IFR89, 7, rue Guy Moquet-BP8, 94 801 Villejuif Cedex, France. Tel.: +33(0) 1 49 58 34 81.

Volume 96, Issue 3, July 2007, Pages 151-157
The Memory of Water

Journal Club – “The Memory of Water: an overview”

January 1st, 2000 by Ben Goldacre in bad science | 3 Comments »

This is part of the Homeopathy journal club described here:

doi:10.1016/j.homp.2007.05.006    How to Cite or Link Using DOI (Opens New Window)
Copyright © 2007 Elsevier Ltd All rights reserved. The Memory of Water: an overview

Martin F. ChaplinCorresponding Author Contact Information, a, E-mail The Corresponding Author
aDepartment of Applied Science, London South Bank University, Borough Road, London, SE1 0AA, UK
Received 10 May 2007;  revised 23 May 2007.  Available online 31 July 2007.


The ‘memory of water’ is a concept by which the properties of an aqueous preparation are held to depend on the previous history of the sample. Although associated with the mechanism of homeopathy, this association may mislead. There is strong evidence concerning many ways in which the mechanism of this ‘memory’ may come about. There are also mechanisms by which such solutions may possess effects on biological systems which substantially differ from plain water. This paper examines the evidence.

Keywords: memory of water; homeopathy; evidence-based medicine; water clusters

Article Outline

The evidence
Evidence against the ‘memory of water’
What is ‘water’?
Evidence for the ‘memory of water’


The ‘memory of water’ is a snappy expression that has eased its way into popular language. The term is mostly associated with Jacques Benveniste following his and others’ allergy research work.1 These research teams showed that solutes subjected to sequential physical processing and dilution demonstrate biological effects different from those apparent using just the water employed for the dilutions. The ‘memory of water’ holds within its brevity of phrase the concept that much diluted solutions appear to behave as though they contain absent solutes that had once been present.

From that beginning, its use has grown to include whether the properties of water can show distinct properties over periods of time much longer than expected and dependent on the aqueous preparation’s previous history. Originally, the term was proposed within a homeopathic context for public dissemination in the popular media. It presented an idea that was easily appreciated by non-scientists, although at a simplistic level. This caught the public imagination and, at the time, proved to be great publicity for homeopathy. More recently, however, the term ‘memory of water’ has proven to be unnecessary baggage in the homeopathy debate. Proof, lack of proof, or simple disbelief that water has, or can have, a memory has quite unnecessarily been confused with proof over whether homeopathy may or may not be efficacious.

Editorial comment in the scientific press has subsequently drawn on whether water can indeed show any ‘memory’ of its prior history as direct ‘proof’ of whether homeopathy can be successful or not. Such linkage is quite unnecessary and may easily mislead as the two areas utilize fundamentally differing and entirely independent evidence and should therefore be considered separately. One, the other, both or neither of these phenomena may represent real effects; they are not interdependent. Linking the two today such that they both stand or fall together2 is as senseless as the non-sequitur of linking the efficacy of aspirin with its bitterness. There is no need to judge homeopathy by ascertaining whether water has a memory of past events any more than we should judge conventional medicine by our level of appreciation, or ignorance, of its detailed molecular action. Thus, whether homeopathy works or not is a mostly separate issue from the content of this paper and should be judged solely on the evidence presented copiously elsewhere. It follows that simply proving that water does have a memory does not prove that homeopathic medicines work. Other considerations have to be taken into account (see later) including how any specificity of action may arise in the therapeutic effect.

This paper concerns the memory of water: to what extent past events may influence the future behaviour of aqueous solutions. Although interpreted by some as also applying to the ‘memory’ of single molecules of water3 this is a red herring just as any discussion of the human memory in terms of the properties of a single molecule of ATP. Also misleading is to discuss it solely in terms of just H2O molecules as no such material as pure liquid H2O can exist; liquid water always contains other species such as H+ ions.

Here, I discuss the memory of water in terms of the real aqueous solutions that are likely to be more generally encountered as ‘water’, including those found in a homeopathic context. Of particular relevance is that the water used in homeopathic preparations, whether distilled or deionized or both or including ethanol, may still contain many and variable solutes including nanomolar to micromolar concentrations of ions.

The evidence

All scientific hypotheses should, of course, be examined in an unbiased manner with reliance on evidence rather than belief. However, agreement or disagreement with the concept that aqueous preparations may have a memory of past events arouses great emotion and has caused careers to flounder, quite independent of any unbiased examination of the evidence and its worth. Before any evidence is examined, the initial response of most scientists and non-scientists (including both the author and indeed Jacques Benveniste) is mostly one of deep scepticism. Indeed it is often stated, by people who do not believe in such a ‘memory of water’, that the burden of proof it requires should be much greater than for other scientific hypotheses.4 Such an attitude may itself be considered unscientific: the same level of supporting evidence should be accepted for all scientific developments. If a lower level of proof is set for hypotheses that fit prior beliefs then we bias our view of science in favour of such beliefs and may be easily misled. That such a process is generally used is self-evident and has resulted in the slow uptake of new ideas and the overly long retention of fallacious concepts.

The science surrounding the ‘memory of water’ is a confusion of evidence for an effect, data showing a lack of evidence for an effect, evidence and opinion that there should be no effect and lack of evidence that there is an absence of an effect; all of which may be obtained under different conditions with varying probabilities. Also, the observation of a phenomenon is usually accompanied by an explanation but the observation does not necessarily prove that the explanation is correct. Thus the explanation of his experimental observations by Jacques Benveniste as due to the ‘memory of water’ may or may not be correct, whereas the data that he published, and its correctness or otherwise, is quite independent of this explanation. Unfortunately, too often the explanation is examined more closely than the experimental data, which may lead to the data being rejected without due cause.

Evidence against the ‘memory of water’

Before describing the evidence for the memory of water, it is pertinent to discuss the evidence suggested by the many scientists who deny water its ‘memory’. Rather surprisingly, these do not concern the production or examination of experimental data showing no effect of their prior history on the properties of solutions. They mostly concern arguments involving the ease with which hydrogen bonds between water molecules may break. Individual hydrogen bonds do not last long in liquid water (about a picosecond). Based on this one fact the opinion may be proffered that the mesoscopic structure of water must change on about the same time scale.

Such arguments are completely fallacious as is easily recognized if metal hydrates or solid water (ice) are considered. In the case of ice the hydrogen bonds also only last for the briefest instant but a piece of ice sculpture can ‘remember’ its carving over extended periods. Cation hydrates exist and are commonly described with particular structure (eg the octahedral Na+(H2O)6 ion) but the individual water molecules making up such structures have but the briefest of residence times (<microseconds).

What such arguments fail to address is that the behaviour of a large population of water molecules may be retained even if that of individual molecules is constantly changing. Such behaviour is easy to observe: a sea wave may cross an ocean, remaining a wave and with dependence on its history, but its molecular content is continuously changing.

The remaining evidence presented against the memory of water concerns whether water clusters may retain their organization for time periods greater than a fraction of a second. Evidence denying the long-life of such water clusters is generally based on computer modelling but also includes NMR and diffraction data.5 There are several good reasons why such methods would not show any significant clustering properties for liquid water.

Computer simulations only operate for nanoseconds of simulated time, although taking hours or days of real time. Such short periods are insufficient to show longer temporal relationships, for example those produced by oscillating reactions.6 They also involve relatively few water molecules (of the order of 100–1000 or so) over small (nanometre) dimensions, insufficient for showing large scale (not, vert, similarmicron) effects. They utilize models for the water molecules that are inherently flawed, showing poor correspondence to the real experimental properties of water (except for those properties on which they were individually based) and hence poor at predicting known properties and likely to be highly inaccurate at predicting unknown properties.7 NMR and diffraction both determine individual water molecules as structures averaged from throughout the sample (akin to averaging the world’s population of men and women and coming up with an illusory ‘average’ person) and are incapable of detecting imprecise and mobile clusters where components may change.

It is clear, however, that in the absence of other materials or surfaces (see below), the specific hydrogen bonding pattern surrounding a solute does not persist when the solute is removed. This may be demonstrated by the change in the water density as salt is removed. If there was ‘memory’ at work here, such water would retain the effective high density that existed in the presence of salt, but it does not. Also, no long-lived clusters made up of particular water molecules can be discerned (>1 ms, >5 μM) by NMR.8 On removal of hydrogen bonds when a hydrophilic solute is removed, the space vacated must be filled by aggressively hydrogen bonding water molecules. Such water would tend to alter the residual hydrogen bonding towards that pre-existing before the solute was originally added. Clearly such a process cannot be considered a memory effect involving ‘remembering’ the state before the solute was added.

Also, it is problematic to put forward a working hypothesis as to how small quantities of just H2O could have any different quantifiable effects when confronted with large amounts of complex and confounding solution in a human subject. Other materials must be present to stabilize such structuring against immediate destruction.

A further argument proffered against water having a memory involves drawing conclusions that water molecules must in their past been in multitudinous contact with almost infinite animate and inanimate objects and therefore cannot possibly ‘remember’ this whole history. I do not dispute this argument, but it is of no relevance to the state of known samples of liquid water, where the history concerns just the sample and is not the sum of the individual memories of all the molecules since the beginning of time (indeed individual H2O molecules only have lifetimes of fractions of a second).

Too often the final argument used against the memory of water concept is simply ‘I don’t believe it’. Such unscientific rhetoric is heard from otherwise sensible scientists, with a narrow view of the subject and without any examination or appreciation of the full body of evidence, and reflects badly on them.

What is ‘water’?

In order to properly discuss the memory of water it is first necessary to note what is meant by ‘water’ in this context. Here we assume that water is a solution of various, and varying, materials in liquid water. Real pure liquid water (if it could be created) would still consist of a number of molecular species including ortho and para water molecules, water molecules with different isotopic compositions such as HDO (D=deuterium) and Click to view the MathML source. Such water molecules occur as part of weakly-bound but partially-covalently linked9 molecular clusters containing one, two, three or four hydrogen bonds, and hydrogen ion and hydroxide ion species.

Apart from such molecules there are always adventitious and self-created solutes in liquid water. Distilled and deionized water still contain significant and varying quantities of contaminating ions. Often the criterion for ‘purity’ is the conductivity, but this will not show ionic contaminants at nanomolar, or even somewhat higher, concentrations due to the relatively high conductivity of the H+ and OH ions naturally present. Other materials present will include previously dissolved solutes, dissolved gasses dependent on the laboratory atmosphere, gaseous nanobubbles, material dissolved or detached from the containing vessels, solid particles and aerosols (also dependent on the laboratory history) entering from the gas phase, and materials produced from water molecules and these other solutes on standing.

Liquid water is clearly a very complex system even before the further complexity of molecular clusters, gas–liquid and solid–liquid surfaces, reactions between these materials and the consequences of physical and electromagnetic processing are considered. It is remarkable that sceptics feel able to state with straight faces that they understand such systems to the extent that they know how they will behave in the absence of evidence on which to base their opinions. Certainly, there are plenty of scientific papers still being published which investigate the unpredictable behaviour of just parts of such complex solutions.

As applied to homeopathy, the memory of water concept should also be extended to the memory of aqueous ethanolic preparations. Addition of ethanol to water adds an important further dimension of complexity. Ethanol forms solutions in water that are far from ideal and very slow to equilibrate.10 Although usually considered a single phase, such solutions consist of a complex mixture dominated by water–water and ethanol–ethanol clusters, where hydrogen bonding is longer-lived than in water alone.11 They also favour nanobubble formation.12 Thus, the peculiar behaviour of aqueous solutions (as mostly discussed in this paper) is accentuated by the presence of ethanol.

Evidence for the ‘memory of water’

The concept of the memory of water revolves around whether the properties of such aqueous solutions change with time and/or processing and/or previous history. There are two aspects this problem. Can any memory of water effect be evidenced?, and is there satisfactory explanation for the appearance of memory in water? Clearly the first element should be sufficient. If there is evidence that the history of a sample of water affects its properties, then the ‘memory of water’ concept is proven without the need for a rationale for its action. However, it would seem that many scientists require an answer to the second part as well because the concept that water may possess a memory effect is perceived as so unlikely that simple proof that it happens is insufficient for them. In other areas of science experimental evidence is easily accepted where people ‘believe’ it to be true without a known rationale for its mechanism. An example is gravity. We believe it due to numerous observations but do not know how it exists. There is no requirement that the explanation for the memory of water is the correct explanation only that it must ‘seem’ reasonable. Of course, if it is also correct, that is a bonus!

There are several ways water can be shown to have a memory. As a simple example, human taste is quite capable of telling the difference between two glasses of water, processed in different ways (eg one fresh and one left undrunk for several days), where present analytical methods fail. There is a change, of course, but such a change would never be noticed by computer simulations on pure H2O. Vybíral and Voráček have shown that water changes its properties with time and its previous history.13 There is also a well-known ‘memory’ effect concerning the formation of clathrate hydrates from aqueous solutions whereby previously frozen clathrates within the solution, when subsequently melted, can predispose later to a more rapid clathrate formation.14 These examples may be explained, for example, by the presence of nanobubbles, extended chain silicates or induced clathrate initiators,15 respectively. Once an explanation is accepted, of course, the ‘memory of water’ seems no puzzle at all.

There are numerous other examples of the slow equilibration of aqueous solution. Thus, it can take several days for the effects of the addition of salts to water to finally stop oscillating16 and such solutions are still changing after several months showing a large-scale (not, vert, similar100 nm) domain structure.17 Also, water restructuring after infrared radiation persists for more than a day,18 and water photoluminescence changes over a period of days.19 Changes to the structure of water are reported to last for weeks following exposure to resonant RIC (resistance inductance capacitance) circuits.20 Conductivity oscillations (not, vert, similar0.5 Hz) at low concentrations of salts also show the weak tendency to equilibrium in such solutions.21

There is a strange occurrence, similar to the ‘memory of water’, in enzyme chemistry where an effectively non-existent material still has a major effect; enzymes prepared in buffers of known pH retain (remember) those specific pH-dependent kinetic properties even when the water is removed such that no hydrogen ions are present;22 these ions seemingly having an effect in their absence contrary to common sense at the simplistic level.

The effect of physical and electromagnetic processing is also evident from a number of studies; for example, due to changes in the amount of silica23 or redox molecules24 produced. Also manifest is that some ultra-dilute solutions, far beyond present detection by chemical analytical analysis, are known to have significant biological effects. A clear case of this is new-variant Creutzfeldt–Jakob disease, caused by infinitesimal amounts of prion protein.

There are several rational explanations as to how water may show different properties dependent on its previous history. In fact these are so obvious that it is a wonder why there is any further fuss about the ‘memory of water’. The current difficulty is choosing between the many reasonable explanations those that are the main causes for any memory effect. What is there in these solutions that depend on its history and which out of these constituents change so slowly as to still be showing effects at future times? To answer this question we need both thermodynamics and kinetics.

A water molecule in liquid water is never at a thermodynamic minimum for any appreciable (or measurable) length of time. This is because there are countless energy states in water with little difference in energy between them and the natural thermal fluctuations in liquid water are easily sufficient to allow change. Water consists of a heterogeneous mixture of states with little tendency for individual molecules to reside at a thermodynamic minimum; a process that is exacerbated by the presence of ethanol.

Solutions used in homeopathy contain many materials that may have biological effects. Of these, nanobubbles, nanoparticles and redox-active materials may separately or together cause biological responses. In terms of specificity, nanoparticles may present the most important possibility, if often overlooked. It is certainly true that such solutions show clear material differences from the diluting water used.25 The process of silica dissolution has been much studied26 ever since it was proven by Lavoisier over 200 years ago and fits with this argument. This may explain why glass is preferred over polypropylene tubes. A thorough investigation into the structural differences previously reported between homeopathically potentized (ie succussed and extremely diluted) and unpotentized nitric acid solutions showed that the effect was lost if different glassware was used.27

It should be noted that dissolved silica is capable of forming solid silicate particles with complementary structures (ie imprints) to dissolved solutes and macromolecules and such particles will ‘remember’ these complementary structures essentially forever.23 Redox-active material, such as superoxide anions and hydrogen peroxide,24 are associated with the control of many cellular processes and their presence is easily capable of giving rise to real biological responses. Nanobubbles are sub-micron scale gaseous bubbles that have long been subject of dispute as their existence does not fit with the commonly held belief in the Laplace equation relating the cavity internal pressures to the surface tension and cavity radius. However the evidence for nanobubbles is now overwhelming12 and the role of dissolved gas in water chemistry is likely to be more important than commonly realized;28 particularly involving the formation and development of these nanobubbles29 and the properties of their interfaces. Relevant to this are changes in carbon dioxide hydration, and hence pH, resulting from different hydrogen bonding effects. Nanoparticles may act by themselves or in combination with the nanobubbles to cause considerable ordering within the solution,30 thus indicating the possibility of solutions forming large-scale coherent domains.

A key feature of any difference between water before and after its use in preparing homeopathic dilutions is likely to be the shaking (succussion) between successive dilutions, and which may produce significantly increased concentrations of silicate, sodium and bicarbonate ions31 by dissolution of the glass tubes and from the atmosphere, respectively. Although not often recognized, except by microbiologists, such shaking can also produce aerosols saturating the laboratory atmosphere for extended periods and offering a route for the contamination of later dilutions.

Succussion involves the effect of pressure changes due to the shock waves produced. The magnitude of this pressure has not been well examined but may be estimated, from conservation of energy equating kinetic energy with strain energy, to be about 5–100 MPa dependent on the procedure. Equally increasing and decreasing (negative) pressure will be encountered so involving the compression and stretching of the hydrogen bonded network. Increasing pressure causes gas dissolution and decreasing pressure causes gas formation. Due to the slow kinetics of bubble initiation, it seems reasonable that such effects will mostly concern pre-existing gas nanobubbles in the bulk and at phase surfaces. Certainly bubbles could both grow and divide during such processing.

A further effect of pressure changes involves the silicate glass–liquid surface. Pressure waves would not only encourage dissolution but may dislodge nanoparticles of silicate (or other) material. It should be noted at this point that glass is not homogeneous but consists of nano-sized domains with differing structures.

Mechanically induced hydrogen bond breakage may also give rise to increased (but low) hydrogen peroxide formation24 and such effects have been reported to last for weeks,29 keeping the solution far from thermodynamic equilibrium.23 Such processes are well-known to produce long term oscillatory behaviour.[6] and [24] It may be relevant to note that the presence of hydrogen peroxide can take part in and catalyze further reactions with other reactive species such as molecular oxygen and dissolved ozone (not often recognized but also present in nanomolar amounts in ‘pure’ water) which may well vary with the number of succussion steps and their sequence and may offer an explanation for the changes in the effects of homeopathic preparations with the number of dilutions. Also of note are the known effects of low concentrations of reactive oxygen species on physiological processes such as the immune response.

That homeopathic preparations are always made using glass tubes may be indicative of the importance of silicates to these phenomena. If this is the case, there will be significant differences between using the same tube during dilution and using fresh glassware at each stage. If the former situation holds, as in the Korsakoff method, there may be marked consequences in terms of enduring changes to the glass surface and the continued presence or build up of materials at the surfaces, including the increased possibility of surface microbial contamination when ethanol is not used.

The processing of solutions also induces electric and electromagnetic effects, both of which seem to produce changes that have long lifetimes.32 The interface of solution with silica, for example, produces high localized fields (Enot, vert, similar109 V m−1) caused by the partial charges on the atoms and the small distances between the surface and first hydration layer. Moreover, the flow of polar water molecules on succussion will itself create changes in electric field.

In addition to the breakage of hydrogen bonds, electromagnetic fields may perturb gas/liquid interfaces, produce reactive oxygen species33 and increase the differences in the properties between the ortho and para forms of water.34 Together with mechanical action, they will lower the dielectric constant of the water,35 due to the resultant partial or complete destruction of the hydrogen-bonded network.

Consequently, the solubility properties of the water will change during succussion and produce changes in the concentration of dissolved gases and hydrophobic molecules at interfaces thus encouraging their reaction (eg due to singlet oxygen, 1O2, or hydroxyl radical, OH·, formation) or phase changes (eg formation of extensive surface nanobubbles36).

These processes also result in the additional production of low concentrations of hydrogen peroxide and other redox materials24 with long-lasting effects. An interesting (and possibly related) memory of water phenomenon is the effect of water, previously exposed to weak electromagnetic signals, on the distinctive patterns and handedness of colonies of certain bacteria.37 Here, the water retains the effect for at least 20 min after exposure to the field.

In homeopathy as elsewhere, dilution is never perfect, particularly at low concentrations where surface absorption may well be a major factor, so that the real degree of dilution beyond the levels that can be analytically determined will always remain unproven. Residual material may be responsible for perceived differences between calculated and actual activity. Unless great care is taken, active material may also enter from the atmosphere even at the greatest dilutions. The water used for dilution is not pure relative to the putative concentration of the ‘active’ ingredient, with even the purest water considered grossly contaminated compared with the theoretical homeopathic dilution levels. This contamination may well have a major influence, and itself be influenced by the structuring in the water it encounters. Although it does, at first sight, seem unlikely that solutes in diluted ‘homeopathic’ water should be significantly different from a proper aqueous control, it has recently been cogently argued that the concentrations of impurities can change during the dilution process by reactions initiated by the original ‘active’ material38 and this process has been modelled to show how different potencies may give rise to differing effects.39

A further consideration about ‘the memory of water’ is that the popular understanding of how homeopathic preparations may work not only requires this memory but also requires that this memory be amplified during the dilution; this amplification, necessitated by the increase in apparent efficacy with dilution, being even harder to understand and explain. Samal and Geckeler have published an interesting, if controversial, paper40 concerning the effect of dilution on various molecules. They found that some molecules form larger clusters on dilution rather than the smaller clusters thermodynamically expected. Certainly, just the presence of one such large micron-sized particle in the ‘diluted’ solution could give rise to the noticed biological action. Of course some such preparations may be totally without action, not containing such clustered particles. However, this observation still sheds light on the phenomenon, which appears to disobey the second law of thermodynamics. A possible explanation is that some biologically-active molecules, such as the fullerene C60 involved in these experiments, can cooperatively form water networks to both surround and screen them. So long as such a network structure requires the help of more than one neighbouring such cluster to stabilize its formation then, in more concentrated solution, the molecules dissolve normally. However, as they are diluted no stabilizing clusters would be available close by if the solution was homogeneous. Consequently, the clusters would stabilize each other by coalescing to form larger clusters of biologically-active molecules within their own water network (ie they form their own aqueous phase). Overall the balance is expected to be rather fine between water cluster stabilization and particle precipitation and dependent on the particle’s ability to form a strongly bound hydration shell.

Although individual molecules of water cannot retain any memory of past hydrogen bonding for periods of more than a fraction of a second, the behaviour of water clusters can be entirely different (Figure 1), as shown previously for ice and cationic hydration. Water clusters are proven entities;41 their size and lifetime dependent on their physical and chemical environment. Liquid water is made up from a mixture of such clusters forming, changing and disappearing.

Display Full Size version of this image (30K)

Figure 1. The lifetimes of clusters are independent of the lifetime of individual linkages. The figure is a two-dimensional representation of a three-dimensional phenomenon. The actual clusters of water molecules are not represented. Supposed that the star clusters (shown filled) may reform around key structures (shown as rhombuses labelled ‘r’, but closed ring oligomers of H2O in water). For each shifting cluster two units (filled circles) move to break up the existing cluster and help create a new cluster. The new clusters are identical to the old ones but only contain a proportion of the water molecules. Clusters may reform around any of the star arms.

The lifetime of a particular cluster containing specific water molecules will be not much longer than the life of individual hydrogen bonds (ie nanoseconds) but clusters can continue forever although with constant changing of their constituent water molecules. For example the icosahedral water cluster described by me42 contains one central-core structure but additionally 12 partly formed potential centre core structures on its periphery. Thus an icosahedral cluster can morph into a different but identical structure by shifting its centre; losing some molecules but gaining others. Although such complete icosahedral clusters are not thought important due to their likely low concentrations in ‘pure’ water under ambient conditions, under other conditions, with other solutes and phase interfaces and including related partial clustering such continuing structuring may well exist and be important.

This process may be considered similar to the existence of hydrogen ions in solution. There is no doubt that hydrogen ions are there, their concentration may easily be determined and they have uninterrupted and continuing effects, but individual hydrogen ions have only a fleeting existence (<nanosecond). The H+ is associated with a cluster of water molecules one moment, but in the next instant it disappears only to be replaced by a different H+ associated with an entirely different cluster of water molecules. Thus the hydrated hydrogen ion continues its existence but contains different atoms.

Water does store and transmit information concerning solutes, by means of its hydrogen-bonded network. Changes to this cluster network brought about by solutes may take some time to re-equilibrate. Succussion may also have an effect on the hydrogen bonded network (shear encouraging destructuring) and the gaseous solutes with critical effect on structuring43 and consequentially contribute to the altered heats of dilution with such materials.44

Recently, there has been some debate over ‘digital biology’; an idea originated by Jacques Benveniste that ‘specific molecular signals in the audio range’ (hypothetically the ‘beat’ frequencies of water’s infrared and far infrared vibrations) may be heard, collected, transmitted and amplified to similarly affect other water molecules at a receiver.45

This unlikely idea may be thought highly implausible but the evidence should be ignored at one’s peril. Note that as with the basic ‘memory of water’ concept, experimental confirmation of the phenomenon may not necessarily confirm the proposed mechanism.

Interestingly, however, electromagnetic emission has been detected during the freezing of supercooled water46 due to negative charging of the solid surface at the interface caused by surface ionization of water molecules followed by preferential loss of hydrogen ions. It is not unreasonable, therefore, that similar effects may occur during changes in the structuring of liquid water.

Finally, the ‘memory of water’ is considered by many to be the apparent physical result of a wider and complex holistic phenomenon.[47] and [48] Such a viewpoint lacks any mechanism for experimental testing at present.


There are a number of mechanisms for water to possess a ‘memory’. These have been described above and shown in the Table 1. The actual mechanism of action may differ between different ‘memory’ occurrences and may be the result of a combination of such phenomena. Some of these factors are clearly more likely, as described within this paper, and others can be easily eliminated or confirmed by closer examination of the procedures and/or analysis of the water.

Table 1.

Possible mechanisms by which water could achieve a ‘memory’

Specific mechanisms Non-specific mechanisms
Remaining material on surfaces Silicates, dissolved and particular
Aerosol material reintroduced Nanobubbles and their material surfaces
Bacterial material introduced Redox molecules produced from water
Imprinted silicates Natural water clustering
Remaining particle clusters Stabilized water clustering
  Ions, including from glassware
  Ethanol solution complexity

Note that, for homeopathy, ‘memory of water’ effects (if proven) not only require the solution to retain information on dilution but require this information to be amplified to negate the effect of the dilution. It is also of importance to note that non-specific mechanisms of action, such as activation of a non-specific immune response, may give rise to effects with specific health consequences. Much research work remains to be undertaken if these real and observable facts are to be completely understood.


1 E. Davenas, F. Beauvais and J. Amara et al., Human basophil degranulation triggered by very dilute antiserum against IgE, Nature 333 (1988), pp. 816–818. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

2 Wikipedia. Water memory. left angle angle bracket, accessed on 28 April 2007.

3 M.L. Cowan, B.D. Bruner and N. Huse et al., Ultrafast memory loss and energy redistribution in the hydrogen bond network of liquid H2O, Nature 434 (2005), pp. 199–202. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

4 J. Maddox, J. Randi and W.W. Stewart, ‘High dilution’ experiments a delusion, Nature 334 (1988), pp. 287–290.

5 J. Teixeira, Can water possibly have a memory? A sceptical view, Homeopathy 96 (2007), pp. 158–162. SummaryPlus | Full Text + Links | PDF (366 K)

6 I. Ferino and E. Rombi, Oscillating reactions, Catal Today 52 (1999), pp. 291–305. SummaryPlus | Full Text + Links | PDF (614 K) | View Record in Scopus | Cited By in Scopus

7 Chaplin M. Models for water. left angle angle bracket, accessed on 5 May 2007.

8 D.J. Anick, High sensitivity 1H-NMR spectroscopy of homeopathic remedies made in water, BMC Complement, Alternative Med 4 (2004), p. 15. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

9 Chaplin M. Hydrogen bonding in water. left angle angle bracket, accessed on 5 May 2007.

10 K. Takaizumi, A curious phenomenon in the freezing–thawing process of aqueous ethanol solution, J Solution Chem 34 (2005), pp. 597–612. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

11 C. Nieto-Draghi, R. Hargreaves and S.P. Bates, Structure and dynamics of water in aqueous methanol, J Phys Condens Matter 17 (2005), pp. S3265–S3272. View Record in Scopus | Cited By in Scopus

12 F. Jin, J. Ye, L. Hong, H. Lam and C. Wu, Slow relaxation mode in mixtures of water and organic molecules: supramolecular structures or nanobubbles?, J Phys Chem B 111 (2007), pp. 2255–2261. Full Text via CrossRef

13 B. Vybíral and P. Voráček, Long term structural effects in water: autothixotropy of water and its hysteresis, Homeopathy 96 (2007), pp. 183–188. SummaryPlus | Full Text + Links | PDF (279 K)

14 R. Ohmura, M. Ogawa, K. Yasuoka and Y.H. Mori, Statistical study of clathrate–hydrate nucleation in a water/hydrochlorofluorocarbon system: search for the nature of the “memory effect”, J Phys Chem B 107 (2003), pp. 5289–5293. Full Text via CrossRef

15 H. Zeng, L.D. Wilson, V.K. Walker and J.A. Ripmeester, Effect of antifreeze proteins on the nucleation, growth, and the memory effect during tetrahydrofuran clathrate hydrate formation, J Am Chem Soc 128 (2006), pp. 2844–2850. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

16 P.M. Wiggins, High and low-density water in gels, Prog Polymer Sci 20 (1995), pp. 1121–1163. SummaryPlus | Full Text + Links | PDF (3581 K) | View Record in Scopus | Cited By in Scopus

17 M. Sedlák, Large-scale supramolecular structure in solutions of low molar mass compounds and mixtures of liquids: II. Kinetics of the formation and long-time stability, J Phys Chem B 110 (2006), pp. 4339–4345. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

18 T. Yokono, S. Shimokawa, T. Mizuno, M. Yokono and T. Yokokawa, Clathrate-like ordering in liquid water induced by infrared irradiation, Jpn J Appl Phys 43 (2004), pp. L1436–L1438. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

19 V.I. Lobyshev, R.E. Shikhlinskaya and B.D. Ryzhikov, Experimental evidence for intrinsic luminescence of water, J Mol Liquids 82 (1999), pp. 73–81. Abstract | Abstract + References | PDF (531 K) | View Record in Scopus | Cited By in Scopus

20 C. Cardella, L. De Magistris, E. Florio and C.W. Smith, Permanent changes in the physico-chemical properties of water following exposure to resonant circuits, J Sci Explor 15 (2001), pp. 501–518. View Record in Scopus | Cited By in Scopus

21 S-Y. Lo and W. Li, Onsager’s formula, conductivity, and possible new phase transition, Mod Phys Lett B 13 (1999), pp. 885–893. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

22 A. Zaks and A.M. Klibanov, Enzymatic catalysis in nonaqueous solvents, J Biol Chem 263 (1988), pp. 3194–3201. View Record in Scopus | Cited By in Scopus

23 D.J. Anick and J.A. Ives, The silica hypothesis for homeopathy: physical chemistry, Homeopathy 96 (2007), pp. 189–195. SummaryPlus | Full Text + Links | PDF (242 K)

[24] V.L. Voeikov, The possible role of active oxygen in the Memory of Water, Homeopathy 96 (2007), pp. 196–201. SummaryPlus | Full Text + Links | PDF (136 K)

25 M.L. Rao, R. Roy, I.R. Bell and R. Hoover, The defining role of structure (including epitaxy) in the plausibility of homeopathy, Homeopathy 96 (2007), pp. 175–182. SummaryPlus | Full Text + Links | PDF (775 K)

26 J. Yang and E.G. Wang, Reaction of water on silica surfaces, Curr Opin Solid State Mater Sci 10 (2006), pp. 33–39. SummaryPlus | Full Text + Links | PDF (156 K) | View Record in Scopus | Cited By in Scopus

27 L.R. Milgrom, K.R. King, J. Lee and A.S. Pinkus, On the investigation of homeopathic potencies using low resolution NMR T2 relaxation times: an experimental and critical survey of the work of Roland Conte et al, Br Homeopath J 90 (2001), pp. 5–13. Abstract | PDF (150 K) | View Record in Scopus | Cited By in Scopus

28 R.M. Pashley, Effect of degassing on the formation and stability of surfactant-free emulsions and fine teflon dispersions, J Phys Chem B 107 (2003), pp. 1714–1720. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

29 L. Rey, Can low-temperature thermoluminescence cast some light on the nature of ultra-high dilutions?, Homeopathy 96 (2007), pp. 170–174. SummaryPlus | Full Text + Links | PDF (267 K)

30 Y. Katsir, L. Miller, Y. Aharonov and E. Ben-Jacob, The effect of rf-irradiation on electrochemical deposition and its stabilization by nanoparticle doping, J Am Electrochem Soc 154 (2007), pp. D249–D259. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

31 V. Elia and M. Niccoli, New physico-chemical properties of extremely diluted aqueous solutions, J Therm Anal Calorim 75 (2004), pp. 815–836. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

32 M. Yamashita, C.A. Duffield and W.A. Tiller, Direct current magnetic field and electromagnetic field effects on the pH and oxidation–reduction potential equilibration rates of water. 1. Purified water, Langmuir 19 (2003), pp. 6851–6856. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

33 M. Colic and D. Morse, The elusive mechanism of the magnetic ‘memory’ of water, Colloids Surfaces A: Physiochem Eng Asp 154 (1999), pp. 167–174. SummaryPlus | Full Text + Links | PDF (79 K) | View Record in Scopus | Cited By in Scopus

34 Andreev SN, Makarov VP, Tikhonov VI, Volkov AA. Ortho and para molecules of water in electric field. 2007; arXiv:physics/0703038v1 [physics.chem.-ph].

35 I. Danielewicz-Ferchmin and A.R. Ferchmin, Water at ions, biomolecules and charged surfaces, Phys Chem Liquids 42 (2004), pp. 1–36. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

36 P. Attard, Nanobubbles and the hydrophobic attraction, Adv Coll Interface Sci 104 (2003), pp. 75–91. SummaryPlus | Full Text + Links | PDF (448 K) | View Record in Scopus | Cited By in Scopus

37 E. Ben Jacob, Y. Aharonov and Y. Shapira, Bacteria harnessing complexity, Biofilms 1 (2004), pp. 239–263.

38 A. Morozov, Avogadro’s number and homeopathy, Homœopathic Links 16 (2003), pp. 97–100.

39 D.J. Anick, The octave potencies convention: a mathematical model of dilution and succussion, Homeopathy 96 (2007), pp. 202–208. SummaryPlus | Full Text + Links | PDF (187 K)

40 S. Samal and K.E. Geckeler, Unexpected solute aggregation in water on dilution, Chem Commun 21 (2001), pp. 2224–2225. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus

41 Chaplin M. The structure of liquid water: overview. left angle angle bracket, accessed on 5 May 2007.

42 M.F. Chaplin, A proposal for the structuring of water, Biophys Chem 83 (2000), pp. 211–221. SummaryPlus | Full Text + Links | PDF (1850 K) | View Record in Scopus | Cited By in Scopus

43 A.V. Kondrachuk, V.V. Krasnoholovets, A.I. Ovcharenko and E.D. Chesnokov, Determination of the water structuring by the pulsed NMR method, Khim Fiz 12 (1993), pp. 1006–1010 (translated in Sov J Chem Phys 1994; 12: 1485–1492).

44 V. Elia, E. Napoli and R. Germano, The ‘Memory of Water’: an almost deciphered enigma. Dissipative structures in the extremely diluted aqueous solutions, Homeopathy 96 (2007), pp. 163–169. SummaryPlus | Full Text + Links | PDF (338 K)

45 Y. Thomas, The history of the Memory of Water, Homeopathy 96 (2007), pp. 151–157. SummaryPlus | Full Text + Links | PDF (400 K)

46 A.A. Shibkov, Y.I. Golovin, M.A. Zheltov, A.A. Korolev and A.A. Leonov, In situ monitoring of growth of ice from supercooled water by a new electromagnetic method, J Cryst Growth 236 (2002), pp. 434–440. SummaryPlus | Full Text + Links | PDF (192 K) | View Record in Scopus | Cited By in Scopus

47 L.R. Milgrom, Conspicuous by its absence: the Memory of Water, macro-entanglement, and the possibility of homeopathy, Homeopathy 96 (2007), pp. 209–219. SummaryPlus | Full Text + Links | PDF (736 K)

48 A.O. Weingärtner, The nature of the active ingredient in ultramolecular dilutions, Homeopathy 96 (2007), pp. 220–226.

Corresponding Author Contact InformationCorrespondence: Martin F Chaplin, Department of Applied Science, London South Bank University, Borough Road, London, SE1 0AA, UK.

Volume 96, Issue 3, July 2007, Pages 143-150
The Memory of Water