In 1988, a French immunologist, Jacques Benveniste, and a group of colleagues published a paper in the prestigious English journal
Nature. Their data indicated that progressive dilutions prepared with water, ethanol or propanol might retain some qualities of various materials that had once been dissolved in it. In particular, they claimed to have measured effects on human immune response… Benveniste claimed that he and his colleagues found evidence that very high dilutions of anti-immunoglobulin E (essentially containing only water) had an effect on the degranulation of human basophils. He, therefore, concluded that it was the 'configuration' of molecules in water that was biologically active. This added support to the homeopathic claim that there was a quality of water that allowed for extremely high dilutions of chemicals to remain therapeutic even without any measurable evidence of the original material.
Follow-up
Nature published the article with two unprecedented conditions: first, that the results must first be confirmed by other laboratories; second, that a team selected by
Nature be allowed to investigate his laboratory following publication. Benveniste accepted these conditions; the results were replicated in Milan, Italy; in Toronto, Canada; in Tel-Aviv, Israel and in Marseille, France, and the article was accompanied by an editorial titled "When to believe the unbelievable."
After publication, a follow-up investigation was conducted by a team including the editor of Nature, Dr John Maddox, American scientific fraud investigator and chemist Walter Stewart, and "professional pseudoscience debunker" James Randi. With the cooperation of Benveniste's team, under double-blind conditions, they failed to replicate the results. Benveniste refused to withdraw his claims, and the team published in the July 1988 a detailed critique of Benveniste’s study. They claimed that the experiments were badly controlled statistically, that measurements that conflicted with the claim had been excluded, that there was insufficient avoidance of contamination, and that there were questions of undisclosed conflict of interest, as the salaries of two coauthors of the published article were paid for under a contract with the French company
Boiron et Cie.
Another group led by Benveniste has reproduced the results while others have failed to reproduce the effects. Beneveniste et al contend that the same conditions were not met in those laboratories. (
source)