The Water That Lost Its Memory

SCIENCE FRICTION, acidly quipped one Paris newspaper. Across the English Channel in London, Britain's New Scientist magazine howled, NATURE SENDS IN THE GHOST BUSTERS TO SOLVE RIDDLE OF THE ANTIBODIES. After a month of heated controversy and speculation, the curtain fell last week, at least for now, on one of the strangest tales of scientific controversy in recent memory. The story became public on June 30, when the prestigious British science journal Nature published a report, hedged with "editorial reservation," on a phenomenon that defied the laws of physics and molecular biology: water apparently retained a "memory" of some molecules it once contained in solution. When such water was mixed with blood cells, that phantom memory seemingly caused a reaction.

Even Nature's editors had a hard time swallowing the results of the research, which was directed by Jacques Benveniste, a laboratory head at the French National Institute of Health and Medical Research. The initial findings were apparently reproduced by scientists in France, Canada, Israel and Italy. Nonetheless, the report was accompanied by an editorial by Editor John Maddox that was almost apologetic. "There are good and particular reasons," he wrote, "why prudent people should, for the time being, suspend judgment." Last week Nature forthrightly rejected the idea of water with a memory and relegated it to the deep freeze, along with other intriguing scientific "discoveries" that have not panned out under scrutiny.

Its demise was the work of a highly unusual investigative team that the magazine dispatched to Paris. Besides Maddox, the Nature group included James ("the Amazing") Randi, the scourge of clairvoyants, faith healers and spoon benders, and Walter Stewart, a free-lance fraud sleuth at the U.S. National Institutes of Health. Their report was merciless: "The hypothesis that water can be imprinted with a memory of past solutes is as unnecessary as it is fanciful." The behavior of the weird water was only a delusion, they concluded, based on flawed experimentation. But the matter did not end there. Nature was still smarting from the criticism that it lent credence to the whole messy business by publishing the report in the first place.

The affair strained credulity from the outset, like the proverbial little man who wasn't there. Benveniste's researchers had diluted a solution of antibodies to such a degree that there was no likelihood that even a single molecule of the antibody remained. But, voila, when human white blood cells were exposed to the superdiluted solution, they apparently responded by releasing a chemical substance, as they would have if they had encountered the initial antibody solution. The effect only worked when the solution was shaken violently. Explained Benveniste: "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." Benveniste was comfortable with his findings but openly admitted that he could not explain the strange goings-on.

Nature's editors were duly skeptical. The magazine had printed, with disclaimers, some dubious reports in the past. In 1974, for example, Nature published a paper that claimed Psychic Uri Geller, since discredited by Randi, could predict dice throws a million times as accurately as chance would predict.

Water memory was more serious. If true, it meant that water was somehow able to retain a memory of substances that had been dissolved in it. Physicists and biologists would have to drastically alter their view of matter, and pharmacologists would have to rethink conventional drug treatment. Moreover, homeopathic medicine, a fringe practice in the U.S. that is widespread in France, would get a boost. Homeopaths believe that extremely dilute solutions of some potentially harmful drugs, vigorously shaken -- a common homeopathic technique -- can treat disease.

The investigation had all the earmarks of an all-out assault on presumed chicanery. "To be frank," said Maddox, "we began by thinking that someone was playing a trick on Benveniste. Our minds were not so much closed as unready to change our whole view of how science is constructed." Notebooks were photographed, researchers videotaped, and vials juggled and secretly coded. Incredibly, the codes were wrapped in tinfoil, sealed in an envelope and stuck on the ceiling so Benveniste and his colleagues could not read them.

Even Randi was watched because of his "reputation for sleight of hand." During one crucial test, the lab suddenly rocked with laughter: Randi was enlivening things with magic tricks. "Only the constant implication that we had something to hide prevented me from stopping this masquerade," said Benveniste.

The investigators' final report debunking Benveniste's research did not imply that there had been fraud. But it did conclude that the experiments were flawed and that no substantial effort had been made to exclude systematic error, including observer bias. Reported Maddox and his team: "We believe the laboratory has fostered and then cherished a delusion about the interpretation of its data." The report expressed dismay that the salaries of two of Benveniste's colleagues had been paid by a French supplier of homeopathic medicines. The Nature investigators admitted, however, that the same firm had paid their hotel bill.

The results of the investigation infuriated Benveniste. He compared the probe to "Salem witch hunts and McCarthy-like prosecutions." Said he: "It may be that all of us are wrong in good faith. This is no crime but science as usual, and only the future knows." Maddox stuck by his final assessment, as well as by his earlier decisions to publish Benveniste's work and send the investigating team to Paris. But he added, "I'm sorry we didn't find something more interesting."