Albert Einstein once said the reason he was able to accomplish so much was because he had "stood on the shoulders of giants" like Newton and Galileo. The Nobel Prize-winning physicist's remark was a reminder how much scientists depend on discoveries made by others. The system depends on the free and unfettered exchange of ideas, which is why the government's effort to restrict publication of research that it says could be used by terrorists has sparked a controversy over how to balance the need for openness against concerns that certain kinds of information might be misused.

Last week, the National Science Advisory Board for Biosecurity unanimously recommended that two respected scientific journals, Nature and Science, not publish certain details about experiments with a strain of bird flu virus known as H5N1 because terrorists might use the information to create a biological Armageddon. The research, which was carried out at labs in the United States and the Netherlands, involved genetically modifying the virus in ways that make it much more lethal to humans.

The investigators' goal was to find out how such genetic mutations might occur in nature so that public health officials could better prepare for an outbreak. The virus rarely infects people, but when it does it kills more than half its victims. Most documented cases have been in Asia and occurred when people came into contact with infected birds. But the researchers found that certain genetic mutations enable the virus to be transmitted through the air as an aerosol — such when people cough or sneeze. That makes it exponentially more virulent, and the government says an outbreak could kill hundreds of millions worldwide.

In the papers submitted to the journals, the researchers include a description of how they modified H5N1's genetic structure. The government wants certain details of the procedure removed prior to publication, something the authors and their editors are reluctant to do, they say, because it would hinder the ability of other researchers to develop treatments and vaccines, and because terrorists are unlikely to possess the sophisticated facilities and equipment needed to replicate their results.

Physicists have long been accustomed to dealing with questions about so-called "dual use" technologies that can be employed either in civilian power plants or in nuclear weapons programs. But this is one of the first cases in which biomedical researchers have had to confront the national security implications of their work. Even the kind of voluntary self-censorship the NASABB is requesting — the agency has no power to enforce its recommendation — impedes the free exchange of information and ideas. On the other hand, no wants to see a biological weapon fall into the hands of terrorists.

Our view is that, in principle, it's reasonable to omit some information in scientific publications whose misuse could be potentially devastating for humanity. Yet in practical terms it's hard to see how such a system might work.

The government suggests restricting the most sensitive information to a limited number of authorized institutions. But that would still involve distributing it to thousands of researchers around the world. Moreover, the long-standing practice of having scientific papers reviewed in advance by others in the field means that manuscript copies of the papers already have been widely circulated among the scientific community. If the goal was to keep the information secret, the cat has already been let out of the bag.

It's important that the scientific community, especially biologists, take the lead in thinking about where and how to draw the line between maintaining the free flow of information and government's desire to protect national security. One rule of thumb might be that even if a government suspects an avenue of research may be in danger of falling into the wrong hands, it must allow the results of the work to be disclosed while acknowledging that some sensitive information has been omitted for publication.

The National Institutes of Health, which commissioned the NSABB study, is the agency best suited to initiate such a conversation, which we suspect will only grow more urgent in the future. The goal should be establishing uniform guidelines for determining when and how the needs of biomedical research and national security conflict with each other, and how such conflicts should be resolved.

Scientific progress thrives on free and open access to information and ideas, and is stifled by its absence. But that shouldn't keep ensuring terrorists never get their hands on information that could wipe out millions of lives from being the government's highest priority.