Since JSTOR's founding in 1994, the popular online archive of scholarly journals hadn't had a single member publisher decide to walk away -- until this month. But last week, JSTOR lost a journal -- and not just any journal, but Science, the flagship publication of the American Association for the Advancement of Science and an early adopter of the JSTOR approach.

In a statement, Science said that "our strategic planning must reflect a business environment that is in a constant state of transition, one that has recently seen dramatic technological and competitive changes." More scientific societies are "digitizing and controlling their own content, and AAAS shares the belief that it is now time to assume the full responsibility for maintaining a complete electronic archive of its flagship publication."

A spokeswoman for the AAAS said that the move did not reflect any lack of satisfaction with JSTOR's performance or any financial disagreements.

Under the agreement with JSTOR, the AAAS can end the relationship -- which it is doing effective at the end of 2007 -- but it cannot take back the material already in the JSTOR archives. So current JSTOR subscribers will continue to have access to the old issues of Science. Libraries that subscribe to JSTOR after Science's departure, however, will not have access.

Michael Spinella, executive director of JSTOR, said that he was "disappointed" to see Science leave, but that he did not see the departure as a significant problem for his organization. He said that he was concerned that the decision would deprive the journal of some of its non-specialized audience. Scientists will continue to belong to the AAAS and receive the journal, Spinella said, but he argued that one of the more significant benefits of JSTOR contracts is that they make journals available to non-specialists, so a humanities or social sciences professor who might never subscribe to Science will end up reading articles in it.

Spinella also noted that JSTOR contracts are not exclusive, so the AAAS could have continued to maintain its own archive while remaining in JSTOR. Libraries purchase packages of subscriptions from among JSTOR's approximately 900 participating journals. JSTOR, which is a nonprofit entity, then shares profits with the journals, which do not pay to participate. About 3,600 academic libraries -- half of them in the United States -- have contracts with JSTOR. Any libraries that relied on JSTOR for access to the Science archives will now have to pay for those subscriptions, which may be what the journal had in mind in mentioning in its statement that continuing with JSTOR was "not practical from either a strategic or operational perspective."

Some libraries have reported concerns related to the impact of the "moving wall" used by JSTOR now that a journal is leaving. The wall allows journals to delay inclusion in the JSTOR archives by a certain time frame, to encourage subscribers to pay for their own subscriptions and not rely on JSTOR. In the case of Science, the wall is five years behind publication, and so will end with the final issues of 2002. A few libraries, confident that the wall would continue to move, have reportedly already thrown out some of their 2003-2006 issues of Science, leaving a gap in their collections.

Spinella said he had heard from at least one such library and regretted that libraries feel so pressured on issues of space and cost that some could be caught in this situation, which he said JSTOR was designed to help. "I think libraries face these kinds of choices every day," he said.