Water Memory - The Nature Controversy

The Nature Controversy

In the original study, French immunologist Jacques Benveniste's team at the French National Institute of Health and Medical Research (INSERM) diluted a solution of human antibodies to such a degree that there was virtually no possibility that a single molecule remained. Nonetheless, they reported, human basophils responded to the solutions just as though they had encountered the original antibody (part of the allergic reaction). The effect was reported only when the solution was shaken violently during dilution. Benveniste stated: "It's like agitating a car key in the river, going miles downstream, extracting a few drops of water, and then starting one's car with the water." At the time, Benveniste offered no theoretical explanation for the effect.

Benveniste submitted the research to the prominent science journal Nature for publication. There was concern on the part of Nature's editorial oversight board that the material, if published, would lend credibility to homeopathic practitioners even if the effects were not replicable. There was equal concern that the research was simply wrong, given the changes that it would demand of the known laws of physics and chemistry. The editor of Nature, John Maddox, stated that, "Our minds were not so much closed as unready to change our whole view of how science is constructed." Rejecting the paper on any objective grounds was deemed unsupportable, as there were no methodological flaws apparent at the time.

In the end, a compromise was reached. The paper was published in Nature Vol. 333 on 30 June 1988, but it was accompanied with an editorial by Maddox that noted "There are good and particular reasons why prudent people should, for the time being, suspend judgment" and described some of the fundamental laws of chemistry and physics which it would violate, if shown to be true. Additionally, Maddox demanded that the experiments be re-run under the supervision of a hand-picked group of what became known as "ghostbusters", including Maddox, famed magician-cum-paranormal researcher James Randi, and Walter W. Stewart, a physicist and freelance debunker at the U.S. National Institutes of Health.

In the first series of supervised experiments, the original experimental procedure was followed as it had been when the paper was first submitted for publication. The experiments were successful, matching the published data quite closely. However, Maddox noted that during the procedure the experimenters were aware of which test tubes originally contained the antibodies and which did not. A second experimental series was started with Maddox and his team in charge of the double-blinding; notebooks were photographed, the lab videotaped, and vials juggled and secretly coded. Randi went so far as to wrap the labels in newspaper, seal them in an envelope, and then stick them on the ceiling so Benveniste and his colleagues could not read them. No memory effect was observed in the blinded experiments.

Nature published a follow-up report in the next issue: "We conclude that there is no substantial basis for the claim that anti-IgE at high dilution (by factors as great as 10) retains its biological effectiveness, and that the hypothesis that water can be imprinted with the memory of past solutes is as unnecessary as it is fanciful." Nevertheless, there was no suggestion of fraud; Maddox and his team initially speculated that someone in the lab "was playing a trick on Benveniste," but later concluded, "We believe the laboratory has fostered and then cherished a delusion about the interpretation of its data." Maddox also pointed out that two of Benveniste's researchers were being paid for by the French homeopathic company Boiron.

In a response letter published in the same issue of the journal, Benveniste lashed out at Maddox and complained about the "ordeal" he endured at the hands of the Nature team, comparing it to "Salem witchhunts or McCarthy-like prosecutions." In both the Nature response and a following Quirks and Quarks episode, Benveniste especially complained about Stewart, who he stated acted as if they were all frauds and treated them with disdain, complaining about his "typical know-it-all attitude". In his Nature letter, Benveniste also implied that Randi was attempting to hoodwink the experimental run by doing magic tricks, "distracting the technician in charge of its supervision!" He was more apologetic on Quirks and Quarks, re-phrasing his mention of Randi to imply that he had kept the team amused with his tricks and that his presence was generally welcomed. He also pointed out that although it was true two of his team-members were being paid for by a homeopathic company, the same company had paid for Maddox's team's hotel bill.

Maddox was unapologetic, stating "I'm sorry we didn't find something more interesting." On the same Quirks and Quarks show he dismissed Benveniste's complaints, stating that the possibility that the results would be unduly promoted by the homeopathy community demanded an immediate re-test. In failing, the tests demonstrated that the initial results were likely due to the experimenter effect. He also pointed out that the entire test procedure that Benveniste later complained about was one that had been agreed upon in advance by all parties. It was only when the test then failed that Benveniste disputed its appropriateness.

The debate continued in the letters section of Nature for several issues before being ended by the editorial board. It continued in the French press for some time. For all of the arguing over the retests, it has done nothing to stop what Maddox worried about; even in the light of their failure they are still used to claim that the experiments "prove" that homeopathy works. One of Benveniste's co-authors on the Nature paper, Francis Beauvais, later stated that while unblinded experimental trials usually yielded "correct" results (i.e. ultradiluted samples were biologically active, controls were not), "the results of blinded samples were almost always at random and did not fit the expected results: some 'controls' were active and some 'active' samples were without effect on the biological system."