Genetically engineered plants are spreading rapidly over the fields of North America. This year farmers in the US and Canada planted an estimated 10m hectares with crops -- maize, cotton, soyabeans and oilseed rape -- that have added genes to make them resistant to insect pests or herbicide sprays.

In Europe, the equivalent transgenic crops have not got beyond small-scale trials, as environmentalists and politicians argue whether the risks of plant genetic engineering outweigh the promised benefits of improved agricultural yields and quality. The debate has generated much emotional heat but little scientific light. What, then, are the real risks?

For many consumers, the main concern is a selfish one: that eating genetically engineered foods will damage their health. Hence the campaign for all products containing genetically modified ingredients to be labelled, so people can avoid them if they want to.

Most scientists, on the other hand, are not worried by the direct health risks. They say the products of the newly inserted genes are harmless to humans -- and any unforeseen problems caused by genetic engineering can be picked up by stringent testing.

A recent example was the transfer of a gene from Brazil nuts to soyabeans. This was intended to improve the beans' nutritional quality by adding an essential amino acid, but it had the unexpected effect of triggering an allergic reaction in some people who ate them during research tests. The project was therefore dropped.

The environmental hazards of crop genetic engineering are potentially more worrying. One big risk is that terrible new weeds could be created inadvertently.

It is unlikely that many crop plants could be turned directly into weeds, simply through the addition of one or two new genes, because they are too highly bred for human purposes to thrive in the wild. But scientists are worried by the danger of the added genes spreading from crops into existing weeds and giving these a new selective advantage, through added resistance to pests, chemical sprays or bad weather.

The genes could leap into weeds through the formation of hybrids between crops and related wild plants. Just last week, French scientists reported in the journal Nature the results of experiments to test this hypothesis. They confirmed that herbicide resistance genes, added to oilseed rape, could move into hybrids between rape plants and wild radish -- and persist in the field through several generations.

Although the French experiments do not prove that this would be a problem in agricultural practice, they do give cause for concern.

Another environmental concern is the effect on insects of adding insecticidal genes to crops. Several of the first-generation transgenic crops produce a bacterial toxin, known as Bt, which kills insect pests. Biologists fear that evolutionary pressures may soon make the pests resistant to the toxin, which is also applied as a bio-pesticide. In that case, farmers would have to resume spraying chemical insecticides.

Crop engineering companies believe they can control the resistance problem by requiring farmers to plant "refuges" of conventional crops alongside transgenic Bt crops. For example, as a condition of buying Monsanto's Bt cotton seeds, US farmers must plant one acre of ordinary cotton for every 25 acres of transgenic crop. The idea is that non-resistant pests can thrive in the refuges and constantly replenish the insects' gene pool with vulnerability to Bt. But it remains to be seen whether such refuges will be large enough to work.

Then there is concern about possible knock-on damage to beneficial insects. Although insecticidal genes are aimed specifically at pests, such as the cotton bollworm, they may cause secondary damage to other insects.

A three-year project, sponsored by the EU, is now underway to study the impact of transgenic plants on bees and other pollinating insects. Preliminary results suggest that genetic engineering of rape plants slightly changes the chemical composition of nectar and pollen collected by bees -- changing their foraging pattern and increasing mortality.

Campaigners against crop engineering, such as Greenpeace, point to other specific hazards -- for example, from the "marker genes" for antibiotic resistance that are inserted into some transgenic plants for technical reasons. But their loudest alarm is about unforeseen dangers that may come from the sheer unpredictability of genetic engineering.

On the other side of the argument, the big ag-bio companies such as Monsanto, AgrEvo and Novartis point to the "risks" of abandoning crop engineering. They say it offers the only route to increasing agricultural yields enough to feed the world's growing population in the next century, without ploughing up the remaining wild parts of the planet or using unacceptable levels of chemical fertilisers, pesticides and herbicides.

In fact, it is probably already too late to stop. The agricultural and industrial momentum behind genetically engineered crops is irresistible. Too many American farmers are already convinced of their benefits. And their European counterparts will not stand being left out for long. A huge experiment in environmental genetics is under way.

11/18/97