The US government's restrictions on research into 'select agents' — organisms that could potentially be used as bioweapons — have rarely been popular among scientists, largely because of the monitoring and bureaucracy involved. But after the terrorist attacks of 11 September 2001 and the anthrax mailings that followed, researchers have accepted the restrictions as a fact of life. And the consequences have sometimes proved to be positive for science.

When pathogens can explode out of nowhere within months, the slow review process can do more harm than good.

This is especially true for research into agriculturally important plant diseases; pathogens that were first included on the select-agent list in 2002. The concern was that 'agroterrorists' might use these diseases to devastate crops across whole regions or countries. The result has been a much-needed boost in funding for the field — especially for detecting outbreaks of plant disease, whether the origins are natural, accidental or intentional. Before 2002, each state ran its own plant-pathology lab, creating a patchwork monitoring infrastructure with disparate practices and little communication. Now the US Department of Agriculture (USDA) operates an integrated network of plant-disease labs that spans the nation.

Nevertheless, the regulations can be inflexible. Take the case of huanglongbing, or citrus greening (see page 148), a disease from Asia that was first seen in the United States in 2005 near Miami, Florida, and has now spread to nearly half of the state's 67 counties. Scientists there can find the bacterium that causes huanglongbing right outside their windows. But because it is listed as a select agent, they cannot study infected plants in their labs for more than a week without violating the law. So in this instance the listing has become not only irrelevant, but downright harmful because it prevents or delays research that might stop the scourge.

To its credit, the law on select agents allows for situations such as this to be remedied through a biennial re-evaluation of the list. The USDA is undertaking such a review at the moment, and may remove one species of huanglongbing bacterium from the list this year. But the process is far too slow — not least because the USDA casts such a wide net in soliciting feedback for its re-evaluations. When pathogens can explode out of nowhere within months or even weeks, such well-intentioned mechanisms can do more harm than good.

To prevent this from happening again, the USDA needs to streamline its re-evaluation process to allow it quickly to de-list a plant disease once it becomes widespread in US soil. Waiting for an arbitrary deadline or excessive amounts of outside input before making such decisions undoes what good the select-agent rule has done.

Last October, the House Committee on Energy and Commerce held a hearing in which some members called for greater oversight of biosecurity research. The focus was the proliferation of laboratories studying human and animal pathogens that could be used as agents of terror. But Congress should also reconsider what oversight and coordination can do to protect people and plants. And it should pay heed to the Catch-22 of regulating research that needs to be nimble and creative to effect such protection.