In some scientific disciplines it is common for draft papers to be circulated to colleagues before publication — a practice that has become formalized with the advent of electronic preprint servers pioneered at the Los Alamos National Laboratory. Superstring theorists and others working at high energies were in the vanguard, and it may be relevant to note that these communities are, within the spectrum of disciplines, unusual in their approach to peer review — for example, many high-energy physics experimental papers are refereed extensively within the large teams of collaborating authors before being circulated.

In other areas, where the communities are larger and the variability in quality and sheer volume of preprinted material somewhat daunting, preprint servers are active but appear to be considered less useful for those reasons. And in some areas of research, life is too fraught with competition for preprints to be advisable. Nevertheless, preprint servers are, for some, part of the means of routine scientific discourse. They seem unlikely to displace the need for the selection and broader propagation of results provided by most journals.

There is no conflict between preprint circulation and submission to Nature — we only request that submitting authors inform us where and when a preprint has been placed on a server. But we do actively discourage prior exposure of results in the public media, given our conviction that the breadth and quality of coverage is strengthened by our embargo policy (for a statement of embargo policy and sanctions, see “author information” at But Nature is more than just an archive. So we reserve the right not to publish papers whose results have already been disseminated and discussed beyond the ranks of specialists. And, as with conferences, our policy is that scientists wishing to publish in Nature are required not to encourage or cooperate with media coverage before publication, while communicating freely with other researchers in whatever way seems appropriate.