A LAB accident has revived fears about experiments that mix human and bird flu viruses and the risk that modified viruses will escape. No virus was released, but the incident has led to a call for details of accidents to be more widely publicised.

Last April, a researcher at the University of Texas, Austin, put tubes into a centrifuge to separate out their contents, which included a human flu virus modified to carry a gene from H5N1 bird flu. The centrifuge became unbalanced and stopped, and when the researcher opened it he found the lid of a safety cup holding one of the tubes had fallen off.

Fearing that the tube inside had leaked, the researcher disinfected everything and called the lab's safety officers. He was wearing a protective hood and respirator, and the whole room was at negative pressure to prevent leaks to the outside. But the researcher had made one mistake: he opened the centrifuge and removed the samples without waiting the recommended 30 minutes to allow any virus-laden aerosol to settle.

In fact, the tube was intact. But if aerosol had escaped, the consequences could have been serious, since the virus would have been able to infect humans, with unknown effects. Experiments since the accident show that the virus replicates more slowly in the lab than human flu, says Bob Krug, head of the Austin lab. But its behaviour in people might be different, and an escapee could also share its new gene with other flu viruses. Such research has been criticised for creating unpredictable viruses that may never emerge naturally (New Scientist, 28 February 2004, p 6).

"The University of Texas dodged a bullet," says Ed Hammond of the Sunshine Project, an Austin-based pressure group. He says the incident only came to light because he demanded to see the university safety committee's records. The committee was not unduly secretive, and plans to publish its minutes, but Hammond says more government oversight and public disclosure is needed from labs handling dangerous microbes.