No. 60: Nov-Dec 1988
The above summary of J. Benveniste's "infinite dilution" experiments and Nature's subsequent "investigation" was written in mid-August (1988). When we returned from a long vacation at the end of September, we found eleven new references on this novel development in the progress (?) of science. Basically, there are only two big questions: (1) Is there, despite all the furor and the machinations of Nature's "hit squad," anything of scientific value in the experiments of Benveniste's group? and (2) Should scientific journals police scientific research? The latter question should be answered by Science itself: we shall focus here on the possibility that real scientific anomalies are being concealed by all the media smoke.
What did Nature's hit squad really find? J. Maddox (Nature's editor) et al concluded that Benveniste and his colleagues did not take enough care in their work, that their data did not have errors of the right magnitude (a statistical quibble), that no serious attempt was made to eliminate systematic errors and observer bias, that the climate of the lab was "inimical to an objective evaluation of the exceptional data," and that the phenomenon was not always reproducible. (7) No evidence of fraud was found. The data originally published in Nature were not explained or shown to be invalid. (11) In fact, the Nature investigation actually confirmed some of the original findings. (5) All of the French work and that of the cooperating laboratories were attributed to "autosuggestion"! (4)
Qualifications of the Nature investigators. J. Benveniste pointed out that none of the three members of the Nature team had any experience in immunology. (4, 11) The team consisted of J. Maddox (a physicist), J. Randi (a professional magician), and W. Stewart (an organic chemist).
Curious aspects of Nature's publication and following investigation. Why did Nature accept and publish a paper when fraud and poor science were suspected? (4, 11) Why didn't Nature hold publication of the original Benveniste paper for four weeks until the investigation was completed? (4, 11) Why didn't Nature insist upon prior experiment replication by an independent laboratory? (6) Actually, replications of the experiment were completed before publication, but at labs selected by Benveniste.
Conventional explanations of Benven iste's results. Several letters to Nature have proposed reasonable explanations for the supposedly impossible results of the "infinite dilution" experiments. (8, 9) It is therefore possible that Benveniste's data are valid and not due to "autosuggestion."
Has the "infinite dilution" anomaly been exorcised? Not in our opinion. Too many unexplained data survive. We doubt, however, that many scientists will rush to their labs to explore this subject. It would be too risky in the present scientific environment. Nature has, in effect, relegated "infinite dilution" research to pseudoscience, whether deserved or not.