"A live genetically engineered virus has been used to stop mice getting pregnant. The Australian scientists who achieved the feat hope one day to use similar viruses to prevent the mouse "plagues" that periodically engulf parts of Australia, and to control other pest species.

But some scientists familiar with the work argue that if the contraceptive viruses were to infect species that are not pests, they could cause an ecological disaster. They say that the risks are so great that the modified viruses should never be released. Even the scientists responsible for the research warn that the contraceptive viruses could wreak havoc if they escaped from a country where the target species is a pest to another where the same animal is a valued part of the fauna.

"The technology has immense potential" says Lyn Hinds, deputy director of research at the Vertebrate Biocontrol Cooperative Research Centre in Canberra, run by the CSIRO, which is Australia’s national research agency, and also the Australian National University in Canberra and two agencies of the Western Australia state government. But she adds: "We are a long way from introducing it into the field and are well aware of the need for public debate and international concensus before that happens."

Research groups world wide are experimenting with contraceptive vaccines - both for controlling animal populations and for developing a human contraceptive. The vaccines work by tricking the body into treating proteins from eggs or sperm as foreign invaders. This triggers the body to make antibodies that block fertilisation.

But the technique developed by Hinds and her colleagues goes one step further by putting the gene for a protein found on the zona pellucida, the protective layer that surrounds the egg, into a virus. Infected mice not only produce antibodies against the protein, but should also pass the virus on to other mice, spreading infertility throughout the population.

Hinds’s colleagues Ron Jackson, Ian Ramshaw and Debbie Maguire inserted the gene which codes for a protein called ZP3 into the ectromelia virus, which infects laboratory mice. Thirteen females were infected with the virus. After 14 days, they were paired with males. Nine females had no offspring, and four produced two pups each. Mice infected with the virus without the gene for ZP3 all produced litters of six or more pups.

In a separate experiment, the researchers found that the mice become fertile as the levels of antibody in their bodies decreased. But after receiving a booster infection of the engineered virus, they become infertile again. The researchers have shown that, apart from preventing fertilisation, the antibodies disrupt the formation of egg cells in the ovaries, which should cause a contraceptive effect of up to six months. This would be enough to prevent mouse plagues, says Hinds, since wild mice rarely live longer than eight months.

The ectromelia virus does not infect wild mice. But Geoff Shellam and Mal Lawson of the University of Western Australia in Perth are now experimenting with mouse cytomegalovirus (MCMV), which does occur naturally in the wild.

Researchers say it should be possible to use other contraceptive viruses to target mammals other than mice that have run riot in Australia, including rabbits, feral pigs and cats. Scientists are investigating the use of the method to control the brush tail possum in New Zealand, an animal introduced from Australia. Malaysia, Indonesia and Vietnam are interested in using it to control the field rat, which destroys rice crops.

Before any of these projects go ahead, however,scientists need to assess the risks posed by releasing contraceptive viruses. In a paper given at a conference on wildlife contraception, published last week, Kent Williams of the CSIRO’s Division of Wildlife and Ecology in Canberra warns that a virus “might travel by natural processes that are unimpeded by international boundaries.”

Then there is the possibility that a virus could infect and sterilise species other than its intended target. MCMV is thought to only infect house mice, but extensive tests will be needed to show this is the case. Hinds believes that no field releases are likely before 2005.

Given the potential dangers, some scientists argue that viruses should never be released. Jay Kirkpatrick of ZooMontana in Billings, Montana, who has used zona pellucida vaccines delivered by dart gun or injection to limit the reproduction of animals including deer and elephants, argues that scientists should not be reassured by tests showing that a virus infects only target species. “What safeguards do we have that it wont mutate and infect other species?”.”