LONDON — decades of healthy profits, the publishing industry now finds itself in the throes of a revolt led by the most unlikely campus revolutionaries: the librarians.

Universities Britain to California are refusing to renew their expensive subscriptions, turning instead to “open ” publishing, an arrangement whereby material is made available free on the with few or no restrictions except for the obligation to cite it.

Paul Ayris, director of library services at University College London, describes the revolt’s goal as “the dream of every researcher — from the desktop and at the end of an Internet connection, to be able to have the world’s literature at your fingertips.”

For the moment, that dream is still a long way off. But with British universities already spending 65 percent of their library acquisition budgets on periodicals — up from 50 percent 10 years ago — and university funding cut back, the pressure for change is mounting.

“It’s not just that journal prices keep going up without any evident relationship to costs,” said Robert Kiley, head of digital services at London’s Wellcome Library. “There is also a concern that so much research, which in many cases has been funded by the taxpayer, is locked behind publishers’ pay walls.”

Asked to explain how publishers justify their rising prices, Tom Reller, a spokesman for the scholarly publisher Elsevier, based in the Netherlands, declined to comment. But in testimony before a British parliamentary select committee in 2004, Sir Crispin Davis, the company’s chief executive, said that over the previous five years “the number of new research articles we have published each year has increased by an average of 3 to 5 percent a year. Against those kinds of increases we think that the price rises of 7 to 7› percent are justified.” At the same hearing Richard Charkin, chairman of Macmillan, which publishes the journal Nature, said, “High quality like Nature have a high unit cost per paper published, because for every article published more than 10 have been reviewed and deselected.”

Like newspapers and the music business, scholarly publishing has been drastically affected by the Internet. But the differences are as striking as the parallels. Unlike journalists, most academics are paid for research or teaching, not writing. Yet all academics need to publish their work — to share and validate their research and also, crucially, to advance their careers. A scholar who does not publish regularly generally does not get promoted, making for a one-sided relationship with publishers.

Sales figures, which can often be counted in the hundreds, matter far less than impact — the number of times a publication is cited by other researchers. And most journals rely on unpaid academic peer reviewers, rather than a paid editorial staff, to select articles for publication All of which makes the economics of scholarly publishing very different from that for either newspapers, which depend on advertisers and a mass readership, or music.

“Scholars write journal articles for impact, not for money,” said Peter Suber, a fellow at Harvard University’s Berkman Center for the Internet and Society. “And while there are a lot of misunderstandings most people understand that open access is not like file sharing music or video piracy. Open access relies on the copyright holder’s consent.”

Another difference is that, at least in science, academic publishing has long had a guaranteed audience. In the past, a researcher who wanted to keep up with the latest developments in, say, biochemistry or particle physics, would have no choice but to subscribe to the relevant scholarly journal — or to ask the department or university library to take out a subscription.

The current price of Elsevier’s scientific journal BBA — Biochimica et Biophysica Acta — is more than €18,000, or $ 25,000, a year (or $ 20,980 for in the United States), which, Mr. Ayris said, “leaves less money every year in the budget to buy new books.”

Mr. Reller, the spokesman from Elsevier, said that although BBA was listed as a single journal on the company Web , the subscription actually covered nine separate titles.

The cost of print subscriptions is just one factor in the explosive growth of open access journals like PLoS One, produced by the nonprofit Public Library of Science, which has grown from publishing 138 articles in 2006 to 6,749 in 2010, making it the largest scientific journal in the world. Unlike paper publications, PLoS One has no restrictions on the number of articles it can accept.

However, according to its publishing director, Mark Patterson, it also differs from traditional journals in basing the decision to publish an article purely on whether the experiments described were conducted in a valid manner and support the results claimed.

“We don’t ask how important the work is or whether the findings are new,” said Dr. Patterson, a geneticist who worked at Oxford University and Stanford University before going into publishing. “We think there are other mechanisms that can decide those things,” he continued, adding that in addition to a much quicker submission and peer review process, PLoS One also offered researchers a chance to see how their findings were used after publication by building in tools to measure downloads, citations and even recommendations on social networks.

Not all open access journals are equally open. In June the Wellcome Foundation, which in addition to housing the Wellcome Library also funds about £650 million, or $ 1 billion, worth of medical and scientific research, announced a collaboration with the Howard Hughes Medical Institute in the United States and the Max Planck Society in Germany to establish a new open access journal aimed at the top tier of researchers around the world.

“We are trying to attract papers which might previously been published in leading journals like Science, Nature, or Cell,” said Mr. Kiley, noting that the Wellcome Foundation, like the National Institutes of Health in the United States, already requires the researchers it funds to publish their results in open access journals.

“We believe that the cost of disseminating the results are part of the costs of research,” said Mr. Kiley, explaining why the Wellcome Foundation built the cost of publishing into the grants it awarded.

If open access journals represent what is know as “gold open access” — in which the final, published version of an article is made available free to the general public as soon as possible — UCL Discovery, a new venture by University College London, is an example of what advocates call “green” open access: a freely accessible repository in which researchers are invited (or required) to deposit their findings but with some possible restrictions on reuse.

Gold open access normally involves a publication fee, paid by the author or funder, Mr. Ayris said, which may be one reason why it accounts for only a fraction of scholarly publishing. However, thanks to mandates like those at the N.I.H. or the Wellcome Foundation, and repositories like UCL Discovery, or the Latindex Portal of Portals, a new project by the Autonomous University of Mexico providing open access to scientific articles from Latin America, Portugal and Spain, green open access is far more widespread.

Whoever pays the bills, publishing is not free. Under the traditional business model most of the costs were met by subscribers, though some journals do also charge contributors, making scholarly publishing one of the most consistently profitable, if least noticed, corners of the business. Elsevier, for example, reported profits of £724 million on revenues of £2 billion last year alone. According the Mr. Suber converting to open access “will involve some cost shifting, But also considerable cost savings” for libraries and university budgets.

The traditional model does have its defenders. After George Monbiot, a British academic and journalist, published an article in The Guardian newspaper last month calling academic publishers “ruthless capitalists,” Graham Taylor, director of academic publishing at the London-based Publishers’ Association, told the journal Times Higher Education that all publishers “aspire to universal access” but that it would take time to find a “sustainable, scalable, funded” way to achieve it.

Writers at the Scholarly Kitchen blog, who are mostly involved in the less commercial end of publishing, said that although subscriptions for popular titles might be expensive, the cost for each individual reader remained very low.

They also noted that many titles with high fees for American or European readers were available free or at lower costs to researchers in the developing world through the Hinari program, a partnership of the World Health Organization and several major publishers including Elsevier, John Wiley and Blackwell.

Sir John Daniel, president of the Commonwealth of Learning, an organization that helps developing countries improve access to education, said such efforts did not go nearly far enough. “One of the major obstacles to education in the developing world is the lack of high quality teaching materials,” he said. “The countries we work with can’t afford journals; they’re already paying an arm and a leg for textbooks.”

“I’ve seen it from both sides,” said Sir John, who was once briefly on the board of Blackwell. “I saw the vast industry built up from publicly funded research, and it was never clear to me what value was being added. But if you needed the material, they had you over a barrel.”

His view that open access scholarly publishing is a matter of international justice has become increasingly influential.

In June the Commonwealth of Learning will be co-host to a world conference on open educational resources with the United Nations Educational, Scientific and Cultural Organization, where Sir John was once an assistant director general. In the lead-up to that conference the Commonwealth of Learning has just been awarded a grant by the William and Flora Hewitt Foundation to conduct a survey of government attitudes to open access and to help formulate possible guidelines for government support.

“Open access is now an established part of the publishing landscape,” said Mr. Patterson of PLoS, “but there is still a long way to go before it becomes the dominant model.”