'The long-held limit of 12 authors on papers existed to prevent articles ending like Hollywood blockbusters...

...where everyone gets a name check'

Most areas of research are now team efforts, and the list of names attached to papers gives journals a real headache

At Harwell, the British nuclear physics centre, a less than top-notch scientist once tried to boost his unimpressive job application by an offering of fruit to the interviewing panel. The reply from the chair was uncompromising: "It takes papers here, not pears."

But even papers in the scientific literature are not always the intellectual gold standard they should be. In some fields, such as mathematics, history or literature, an individual can publish a worthwhile contribution in his or her name alone. Provided allies, advisers and lovers are thanked, everyone is happy. But most areas of research have long since become team efforts, and the resulting papers mirror the sometimes problematic academic world that produces them. The Society for Neuroscience puts it bluntly in its guidelines for authorship of journal articles: "The credit implied by authorship is often used as a measure of scientists' productivity in evaluating them for employment, promotions, grants and prizes."

With this in mind, some journals have been rethinking their rules about the authorship of the papers they publish. The issue first arose in the form of "spurious authorship", the habit of heads of department or supervisors attaching their name, sometimes as first author, to everything their department or group produced. This problem has been much reduced in recent years. Spurious authorship is specifically forbidden by codes of practice such as that published by the International Committee of Medical Journal Editors. More than 500 journals have signed up to its rules, which specifically exclude "acquisition of funding, the collection of data, or general supervision of the research group" as reasons to be credited as an author of a journal article. This looks simple enough, but it is difficult to implement.

This summer, the New England Journal of Medicine was forced to abandon its long-held limit of 12 authors on papers published in its pages. Gregory Curfman, executive editor, said the limit existed to prevent "authorship creep", the tendency of articles to end like Hollywood blockbusters, where everyone from Bruce Willis to the assistant accountant gets a name check. More tellingly, the editors' guidelines state that "each author should have participated sufficiently in the work to take public responsibility for appropriate portions of the content", a criterion the journal thought was not likely to be met by each member of a crowd more than a dozen strong.

Now, however, this thinking has collided with the journal editors' other stipulation: that "all persons designated as authors should qualify for authorship, and all those who qualify should be listed". According to Curfman, more and more biomedical breakthroughs are based on large clinical trials. Some involve dozens of investigators. Without their cooperation, the research might not get done, so an arbitrary limit of 12 authors may discourage participation in large research projects.

A gentler approach to the same problem seems to be followed by Nature , which has never had a limit on author numbers and has published papers - for example, on the human genome - with hundreds of names. Maxine Clark at Nature said that the journal operates on the basis that "the world is a civilised place". The corresponding author - the contact person for the paper - is responsible for saying who wrote it. The only exception is if something goes wrong. Every author of the original paper has to sign up to any correction published in Nature .

Clark also accepts that the world of authorship is a subtle one. Postdoctoral students may be listed on a paper because they need the reference for their next step on the career ladder. On the other hand, the head of department may spend more time jetting to conferences than in the lab, but still appear as an author. So might a provider of significant equipment. None of these would be able to defend the paper in full. Clark adds that possible authorship problems are rare but are sometimes highlighted by referees, who tend to be members of the same small community of people as the authors.

Most authorship problems have arisen in biomedical sciences because of the numbers of people and the sums of money involved. But the social sciences and the humanities should not think they are safe. Ever more papers in these fields depend on massive datasets. In the UK, the census, running back to 1801, is one such and is being mined for information with methods similar to those used to analyse clinical trial data. An archaeological excavation can involve as many people as a space mission, all with their own highly developed skills. Should they be listed as authors instead of being thanked anonymously in the acknowledgements section of the dig report?

Martin Ince is contributing editor, The THES .