Nature, one of the world’s most-cited scientific publications, took a step toward open access on Tuesday by granting its subscribers and journalists wide authority to let outside readers view its articles at no cost.

Under the new policy, subscribers to 49 journals published by the Nature Publishing Group and collected on Nature’s website can create and share links to full-text versions of all of that content. About 100 media outlets also can include free links in news reports that reference articles in the group’s journals.

The change is a financial risk for Nature, which recognizes that it may lose money from both subscribers and nonsubscribers who buy access to a single article, Steven C. Inchcoombe, chief executive officer of the Nature Publishing Group, said in an interview outlining the decision.

"But we think the bigger risk is pretending it’s not happening," Mr. Inchcoombe said of the growing movement toward open-access formats, and Nature finding itself in a world where "that usage [is] gradually migrating elsewhere, and us being left as a glorified digital archive making available content to fuel activity everywhere else."

The move follows steadily growing pressure on publishers to permit open access to articles describing research produced with taxpayer dollars. Last year the Obama administration introduced rules requiring journals to make articles about federally financed research available to the public within 12 months. And last month the Bill & Melinda Gates Foundation said it would require that its sponsored researchers publish only in journals that allow immediate open access.

Doubts About Alternatives
Among the major science journals, Nature has been recognized by open-access advocates as especially receptive to helping scientists share their work. "It wants to be prepared for a world in which open access becomes more and more the default," said Peter Suber, director of the Harvard University Office for Scholarly Communication.

At the same time, doubts have persisted about the major alternative economic models, in which journals solicit benefactors or researchers pay to have their papers published. In a recent embarrassment for open-access advocates, the International Journal of Advanced Computer Technology accepted for publication a manuscript that consisted almost entirely of a profanity-laced seven-word phrase, repeated over and over.

High-quality journals such as Nature still must rely on subscription payments to cover their costs, Mr. Inchcoombe said. He said the publishing industry had been searching for a middle ground that meets the substance of open-access demands while giving journals enough resources to subject their published articles to a robust system of editing and peer review.

Nature’s new system falls short of open-access ideals in various ways, including that it restricts nonsubscribers to "read only" versions of articles. That prevents independent repositories from reformatting the articles for long-term storage, and it limits researchers’ ability to search or index the documents.

Also, Nature is describing the move as a trial, one that may benefit the company by helping it understand how subscribers already share content through unofficial means such as copying and pasting.

Nature hopes the experiment will help it further clarify how it can help scientists while keeping itself financially healthy, and it does not intend to rescind the change, Mr. Inchcoombe said. "Never say never, but that’s not our intention," he said. "We believe that content sharing needs to be made simple and powerful."

The Nature Publishing Group already offers about 40 percent of its content in an open-access format. It also has tested open-access supplements financed by outside sponsors, and over all has done more experimentation with open-access than any other major publisher, Mr. Suber said.

"It’s a bold move," he said of the new policy. "Nature can afford to take it because it’s so successful," but "there are other successful publishers that could also afford to take it."

Among the questions, however, is how long Nature will allow the policy to last, especially if—as expected—its articles are very widely shared, Mr. Suber said.

Mr. Inchcoombe said the company would not allow users to effectively reproduce entire issues of Nature, though he did not define exactly what behaviors it would tolerate. "We’re looking to create something that meets all of the reasonable needs," he said.