We must not confuse the facts about GM crops with bad science fiction

The triffids are coming! From the tone of some of the reactions to a mix-up at a seeds company, we might expect mutant oilseed rape plants to invade our back gardens at any moment.

The triffids are coming! From the tone of some of the reactions to a mix-up at a seeds company, we might expect mutant oilseed rape plants to invade our back gardens at any moment.

This is B-movie science fiction. The science fact is considerably less exciting. There are certainly causes for concern about the mixing of some rape seeds with a genetically modified variety, but the risks are small and theoretical. No serious scientist suggests that genetically-modified foods pose a direct threat to human health.

The worries about GM foods, therefore, are about indirect effects. One is the law of unintended consequences. Although there is no known way in which modifying the genes of plants could have any effects on animals that eat them, the lesson of the BSE disaster is that no one knew about particles called prions before, which might be implicated in the inter-species transfer of mad cow disease.

In the case of GM foods, however, it must be said that the risks of the unknown are considerably smaller than the fears. If scientists or governments had operated on the basis that new technologies might have adverse consequences which could not yet be foreseen, we would not have had the railways, the industrial revolution, television or computers.

The more serious concern about GM foods is the threat to biodiversity, the buzz word for the variety and interdependence of life. It might be observed that turning vast blocks of the British countryside into a bright and unnatural yellow in order to produce cheap vegetable oil and animal feed is hardly the best way to maintain biodiversity. This only underlines, however, the extent to which farming already represents a huge intervention by humans into "natural" eco-systems.

GM technology does not mark a significant additional intervention in scope or intensity, but it does mean a quite different type of intervention. Instead of changing eco-systems by cultivation, and altering the course of evolution by breeding, humans can now intervene directly in evolution by altering the genes of plants and animals. This means the speed with which mistakes can feed back has been accelerated - while the nature of risk has arguably become more unpredictable.

On the specific case of the oilseed rape, the principle is simple. The company, Advanta, must pay for its accident. This is something that the market and the law can handle, with the role of the state limited to ensuring that the fullest possible information is available to all - in particular to ensure that the final consumer knows which oil has been produced from GM plants, so that people have a meaningful choice. People need to be able to trust information about whether or not foods are GM or not, which means being able to trust the controls that underlie such information.

If this case provides the incentive for the faster development of tests for GM material so that products can be accurately labelled, so much the better. For that reason, too, Nick Brown, the Agriculture Minister, should be embarrassed about his failure to publicise Advanta's mistake as soon as he knew about it - a month before we did. Could it really take a month to "establish the details and to check the status of the particular genetic modification involved", as he claimed on Thursday?

Mr Brown also needs to explain how, if "there is no threat to the environment because the GM variety is sterile", Advanta's non-GM plants became cross-pollinated by a nearby "sterile" GM crop.

But these are technical details. What matters is that the triffids are not coming. Not this time, anyway.