When the Russian mathematician Grigori Perelman submitted his now-famous solution to the Poincaré conjecture — previously one of the longest-standing unsolved problems in mathematics — he did not publish it in a peer-reviewed print journal. He dropped it into an online archive that was “open access,” or free of charge to readers. Already a worldwide scholarly sensation, Mr. Perelman did not need the prestige and professional approbation of a peer-reviewed print journal. “But that’s not most people’s situation,” said a professor of mathematics at New York University, Sylvain Cappell.

Nevertheless, more scholars — particularly younger ones — are posting their research online before publishing it in refereed print journals, said Mr. Cappell. “It’s an opportunity to get response and reaction from other frontline researchers.”

The Internet, which has made the cost of distributing information virtually vanish, Mr. Cappell said, has also exacerbated the overlapping tensions of online versus print, open access versus paid, and refereed versus non-refereed. Formerly, editorial boards controlled the distribution of research results. Now, he said, “Electronics is bringing disintermediation to the world of research as much as to the world of travel agents.”

In the old days, scholars might mimeograph and circulate copies of their work to a handful of colleagues before it appeared in a refereed print journal. Now scholars can post research in an open-access archive and then publish it in an online or print journal, Mr. Cappell said.

The use of online peer-reviewed journals has also grown in import. “Universities now recognize the legitimacy of electronic peer-reviewed journals in the tenure process,” said Mr. Cappell. “On the other hand, few would relish being the first in his or her department to come up for tenure review with no print publications.”

But most print journals also have online editions. The publisher of 34 technical journals for the American Chemical Society, Robert Bovenschulte, said, “Some journals exist in print only, but they are a very small fraction in scientific, technical, and medical publishing.” The senior director of open-access and journal publishing at Springer, which publishes about 1,300 scientific journals, Jan Velterop, said that peer-reviewed e-journals and print journals were not essentially different. The electronic format is just “the post office,” he said, describing how that content is delivered.

The editor in chief of the American Physical Society, Martin Blume, said the nine APS journals offer a discount if a subscribing institution does not purchase the print version. He said the number of institutions subscribing to only the electronic version of the APS journals has grown, and of their subscribers in America, it is approaching 40%.

Societies and publishers have also begun to wade into open-access publishing. Two of the APS journals are available only in open-access electronic format. Mr. Blume said APS retains copyright in its journals, but gives authors of an article the right to “do what they want with it, as long as a fee isn’t charged. If a fee is charged, they need our permission.”

Online peer-reviewed journals are not without costs, such as labor involved in publishing them or server space on computers. “We’re non-profit. We’re also not for loss,” Mr. Blume quipped. Online journals support themselves through subscription fees, advertising, and “page charges,” where an author pays a fee to help cover costs.

Mr. Bovenschulte said ACS puts articles online “ASAP” — “as soon as publishable” — before they are collected for print publication. Two months ago, ACS instituted an option whereby authors can pay ACS to have their articles available free of charge online.

Many articles submitted to open-access repositories like arxiv.org, which are not peer reviewed, are ultimately published in peer-reviewed journals such as those of the APS, Mr. Blume said. The physics community sees such prior preprint publication as an extension of the tradition of widespread distribution made possible by the Xerox revolution, Mr. Blume said. But he said there are other science journals that will not accept articles that have been previously posted on open-access archives.

Still, “one could speculate that over time, the shift from print to electronic distribution of research results will shift the emphasis from refereeing to reviewing,” Mr. Cappell said. In an electronic age of instantaneous delivery, he said, “there may be less an issue of control and more an issue of reception and reaction.”

Mr. Cappell said the editorial board of journals of yore provided a common meeting ground, adjudicating tensions behind the scenes, with referees asking authors to include reference citations to other researchers’ work. It may be that reviews and surveys will come to play some of the role that referees and editors did in an earlier time, he said.

In the meantime, academic authors could always publish their research on their own web site, but they would still “have to convince their colleagues,” Mr. Velterop said. Without the imprimatur of their peers, their work is not taken as seriously.

As for refereed journals, Lee Smolin, of the Perimeter Institute for Theoretical Physics, would completely shake up the system of academic publishing and its role in tenure decisions: “The refereed journals are entirely unnecessary and continue to exist only because rules requiring faculty to be evaluated in terms of refereed publications haven’t yet changed.”

Mr. Bovenschulte said in coming years there will be more sophisticated search engines and linking. One example may be more frequent use of “web enhanced objects” where an author can embed a 3-D representation of a molecule that can be rotated. Mr. Velterop said in future decades, print may come to be seen as “a convenience product.”