Unlike any other category of writer or author, scientists seldom make money from seeing their work in print. Instead, being published in a scientific journal is a means to an end, to communicate research findings with peers.

This significant difference left scientific publishers looking distinctly exposed when the rise of the internet gave scientists direct access to their readers - other scientists.

Sure enough, journal publishers have come under pressure from initiatives inspired by scientists who wanted their work to be freely available. Perhaps the best example is PubMed Central, a web site set up by the National Institutes of Health, the prestigious US biomedical research body, with the aim of creating a single repository of all published biomedical research.

Publishers were invited to make their journals available free on the site. To date very few have agreed to do so, and mostly at least six months after publication elsewhere.

PubMed Central also wants to add value, by allowing scientists a single point of entry to an online medical library, providing tools so they could search across all journals, and access related public databases such as those holding gene sequences or protein structures.

Meanwhile, the Public Library of Science (PLS), a scientists' pressure group, began a campaign to make scientific literature freely accessible to scientists and the general public, "for the benefit of scientific progress, education and the public good."

Since November 2000 more than 30,000 scientists from 177 countries have signed an open letter from PLS, urging publishers to allow research to be distributed freely by independent, online public science libraries. The PLS called on signatories to boycott journals that did not comply.

Publishers have also been subject to a moral imperative to allow free access to scientists and clinicians in developing countries. At the instigation of the Secretary General of the United

Nations, Kofi Annan, the World Health Organisation approached six of the leading journal publishers to make their research available on Health InterNetwork, a UN programme to improve health in developing countries. As a result, the Health InterNetwork Access to Research Initiative (Hinari) was launched in July 2001.

The founding members are Blackwell, Elsevier Science, Harcourt International (since acquired by Elsevier), John Wiley and Wolters Kluwer. Between them they have made 1,500 journals available free to institutions in countries with a gross national product per capita of below $1,000, and at a reduced rate in other developing countries.

Then, in February this year, George Soros, the financier and philanthropist backed a new effort to provide free and unrestricted access to the literature, launching the Budapest Open Access Initiative with a $3m grant.

Traditional scientific publishers also came into competition with new, virtual rivals, most notably, BioMedCentral, the first internet-only scientific publishing company, which arrived on the scene promising not only that there would be no hard copy version of its journals, but that all its information would be free.

BioMedCentral gets its revenue from advertising and an article processing charge of $500, payable by individual scientists or their institutions, for manuscript production and editing costs.

The company, based in London, further challenged traditional scientific publishers by saying it would use end-to-end electronic processes to cut the time taken to publish an article to no more than eight weeks. For printed journals it is not unusual for it to take six months or more for an article to be published.

These, and other moves put traditional publishers in a spin, leaving them to argue that they are not merely a conduit for disseminating scientific papers, but add significant value, not only through the intellectual input of editing them, but also in running the peer review system by which articles are subjected to the scrutiny of scientific peers.

Mary Waltham, formerly president of Nature, and managing director and publisher of The Lancet, and now an independent consultant advising publishers on their electronic strategies, says the fundamental value that publishers add is still needed on the internet.

"I see no influence on the concept of peer review, only on the process. Peer review may be more mechanised, but it is not going away. After all, formal publication in peer-reviewed journals is still what career progression depends on."

Like most old economy companies, scientific publishers had to defend themselves against the threat of virtual competition at the same time as reformulating their business models to take advantage of internet technologies to cut costs and become more efficient. The sector is very fragmented with many very small publishers without the means, or the know-how, to publish electronically.

Scientific publishing is the specialism of one of the few successful venture-funded dotcoms, Ingenta plc, which spun out of the University of Bath in May 1998. It manages and distributes research via the net and hosts e-journals for publishers.

The company's services include data conversion to electronic formats, secure online hosting, access control, collecting pay-per view payments and distribution services.

Ingenta is not in competition with established publishers says Mark Rowse, chief executive. "When the net began to evolve, publishers were confused about if they needed to control the distribution process.

"We're providing a service which is analogous to printing and distributing a printed journal."

Ingenta manages 12m articles from over 5,400 online publications, and is accessed by over 5m researchers a month via Ingenta.com, or links from other websites. The company claims nine of the top 10 leading scientific publishers as clients, including Elsevier with 1,428 publications (see article on Elsevier's ScienceDirect online service on the FT-IT website, www.ft.com/ ftit), and Blackwell with more than 600.

Adding the e-channel has increased publishers' costs, but going online is pushing up revenues. In the six months ended March 2002 Ingenta's turnover was ý7.1m, of which 20 per cent came from pay per view. "This is coming from people who aren't subscribers, so although the revenue is shared with us, it is beginning to represent significant new income for publishers."

Scientists are also demanding aggregation, so they can search across all titles from a single point.

Mr Rowse says that although publishers have resisted such moves, they are becoming more reconciled. "They see it a bit like being in the Yellow Pages if you are plumber. They want to be in it, but have to accept that they will appear alongside their competitors."

Ms Waltham agrees that in the past two years publishers have become less nervous of aggregation. "For small and medium publishers, scientific portals do allow them to reach bigger markets. And there is now a recognition that the Journal of Keyhole Surgery.com can't expect readers to flock direct to its web site."