EDITORIAL The Memory of Water: a scientific heresy? This special issue of Homeopathy is devoted to the ‘memory of water’, a concept forever linked to the name of the late Jacques Benveniste, although not coined by him. The term first appeared in the French newspaper Le Monde, commenting on a fierce controversy which blew up in the pages of the leading scientific journal Nature in 1988. In June of that year, Nature published a paper by a large international group led by Benveniste which made the sensational claim that the antibody anti-IgE in dilutions up to 10 À120 molar, far into the ‘ultramolecular’ range, triggers degranulation of human basophils in vitro. 1 Nature had resisted publishing the paper, and the then editor, John Maddox, agreed to do so only on the condition that Benveniste allowed an inspection team, nominated by Maddox, to visit his laboratory after publication. The team duly visited, and, a month later, published its report denouncing Benveniste’s work as ‘pseudoscience’, but nevertheless justifying its decision to publish. 2 Two subsequent attempts to reproduce Benveniste’s results failed, 3,4 although he remained defiant until his death in October 2004. Yole` ne Thomas, a long-term collaborator of Benveniste, recounts that episode and the subsequent history of the memory of water in this issue, 5 and Michel Schiff has given a detailed insider’s account of the treatment Benveniste suffered for his heresy. 6 A bad memory Yet, the memory of water is a bad memory: it casts a long shadow over homeopathy and is just about all that many scientists recall about the scientific investi- gation of homeopathy, equating it with poor or even fraudulent science. So why revive it now? The reason of course is the claims made by homeopathy for the action of ultramolecular (also called ultra high) dilutions. Although the basic idea of homeopathy is similarity, the most controversial and, for many, implausible claim concerns the properties of the ultramolecular dilutions characteristic of it. Avoga- dro’s constant, the number of particles (atoms or molecules) in a gram mole of a substance, is of the order of 10 23 . The inescapable corollary is that dilutions of substances above this level are unlikely to contain a single molecule of the starting substance, whose name appears on the label. In homeopathic terminology, 10 23 corresponds to a 23x/dH or 12c dilution. In fact, for reasons including the concentra- tion of the starting substance(s) the ultramolecular limit is often passed well before 23x/12c. In any case, it is only a statistical probability and many homeopathic starting materials of biological origin are complex mixtures of many chemicals in varying concentrations. It is this problem that links Benveniste’s work to homeopathy: he claimed to have discovered that aqueous dilutions of a protein retained the essential properties of that protein many 1:100 dilution stages after it had been diluted out. The water diluent ‘remembered’ the anti-IgE long after it was gone. The underlying hypothesis can be stated as follows: ‘Under appropriate circumstances, water retains in- formation about substances with which it has pre- viously been in contact and may then transmit that information to presensitised biosystems’. Note that this hypothesis has two parts: retention of information and transmission of information. It is now generally accepted that Benveniste’s original method does not yield reproducible results, so why has the idea of memory of water not faded away? Competing hypotheses In fact, there are competing theories for the effects of homeopathy. The most widespread is that no explana- tion is required: homeopathy has no specific effects, and its outcomes are attributable to purely placebo effects: psychological phenomena, including expecta- tion of benefit in which the homeopathic medicine plays no role except to convince the patient that they are receiving a genuine medical treatment. Among the counterarguments to this position is that homeopathic medicines and treatment regimes seem, from what is known about the factors which increase placebo effects, designed to minimise it! 7 They are small and unimpressive, and often administered at low frequencies. Of course the main counterargument is the steadily growing body of evidence from both clinical and bench science that homeopathy and homeopathic ultramole- cular dilutions have effects which cannot be discounted in this way. Other hypotheses which accept that there is something to be explained have emerged, most notably a group involving ‘macroscopic quantum entangle- ment’. These are represented in this issue in the papers by Weinga¨rtner 8 and Milgrom. 9 Yet, among those hypotheses which accept that there is something to explain about the properties of ARTICLE IN PRESS Homeopathy (2007) 96, 141–142 r 2007 The Faculty of Homeopathy doi:10.1016/j.homp.2007.05.008, available online at http://www.sciencedirect.com