There are no genetically modified triffids stalking suburbia, and no unkillable "superweeds" choking the farm. But for nervous consumers in Australia and around the world, the spectre of transgene escape from GM crops looms larger than any real-world threat.

More than a decade of anti-GM scaremongering has persuaded many consumers that transgenic crops pose unacceptable risks to their long-term health, and to the environment, scientific and epidemiological evidence to the contrary.

In a study published in Science in 2002, Dr Mary Rieger of the Australian Weeds Cooperative Research Centre in Adelaide, showed the risks of transgene introgression into weedy crop relatives creating "superweeds" were very low.

Rieger found that hybridisation between GM herbicide tolerant canola and the wild radish (Raphanus raphanistrum) in the field is very rare - among 53 million seedlings raised from Roundup-Ready canola, she found only two herbicide-tolerant hybrids.

The public's concerns are twofold: that pollen from transgenic crops will contaminate nearby conventional and organic crops, and that GM crops will hybridise with weedy relatives, creating intractable superweeds.

GM crops have now been grown for more than a decade, with no adverse episodes for human health, no greater impact on the environment than their non-GM counterparts, and, in most cases, real environmental benefits.

But the anti-GM movement's concerns have eroded consumer confidence in GM crops and foods, and there are good commercial reasons to devise an effective way of preventing transgene escape through wind-blown or bee-borne pollen.

Physical containment of transgenic crops and their pollinators, or the use of extended buffer zones to isolate GM crops from conventional and organic crops, are costly and impractical. Gene technology itself proffers a low-cost genetic solution.

At CSIRO Plant Industry's Merbein research laboratories in north-western Victoria, Professor Steve Swain, Dr Davinder Singh and Dr Angelica Jermakow have devised an ingenious, broadly applicable strategy to prevent transgene escape from GM crops.

It exploits a combination of post-transcriptional gene silencing - RNA interference - and "gene imprinting": natural, methylation-induced suppression of particular genes, according to whether they are inherited from the female or male parent.

The CSIRO researchers have demonstrated their system in an Arabidopsis model. A basic difference from the Technology Protection System developed in the US in the late 1990s is that the transgenic plants still produce viable seed - provided similarly-engineered cultivars cross-pollinate with each other.

If GM pollen is transferred to a non-GM crop of the same species, or a weedy relative, fertilisation will fail and no viable seed will result.

The system would prevent contamination of non-GM crops, prevent transgenic crops hybridising with weedy relatives, yet still allow farmers to save GM seed for replanting the following season.

In the case of crops like maize and sunflowers, farmers have not saved seed since the 1930s because F1 hybrid seed does not "come true" to type. After initial resistance to the new-fangled hybrids, farmers gave up saving seed in return for the superior yield, disease resistance and profitability of F1 hybrids.

The Merbein research team's new system addresses two of the anti-GM movement's major objections to GM crops: there would be no "contamination" of non-GM crops, and no superweeds.

But because the solution to these objections depends on the very technology to which anti-GM activists object it seems unlikely that the anti-GM movement would endorse its use.

There would also be an issue if companies developing GM crops would have to find a way to levy fees for re-use of their proprietary technology if farmers saved GM seed from season to season.