The West itself is dubious about the actions of its citizens. In 1997 the United States Patent Office overturned an attempt to patent the active ingredient of turmeric as an aid to wound-healing as this had long been used as a folk remedy in India. Indeed, the nation's own fingers have been burned. The enzyme used in the polymerase chain reaction comes from a bacterium collected in a hot spring in Yellowstone National Park. The Swiss company that owns the patent makes a hundred million dollars a year in royalties, while the Federal Government (the owner of the spring and presumably of the bacteria) gets not a cent. Now, the Parks Service charges a hundred thousand dollars a go (plus a guaranteed share in profits) for any company that wants to prospect for DNA on its land. Although the claims of wealth from tropical nature may be exaggerated — after all, only one of fifty thousand plants tested by the US National Cancer Institute gave a usable drug — the third world is understandably enraged. Now it is fighting back. Amazonian tribes use the skin of certain frogs as a source of poison for their darts. That substance is a pain-killer if used in minute amounts, and an American company is keen to patent it. But, counter the governments of Ecuador and Venezuela, was not the discovery made by their own people and should not at least some the profits come back to them? The Americans disagree (and are annoyed by the Venezuelans, who have put a stop to the collection of frogs by outsiders).

All this is the stuff of commerce and is as familiar to students of the history of gold-prospecting as much as to those of gene-mining. Like most of the fuss about manipulated crops, it lies outside science. However, science itself has something to say about the implications of the new technology. Not all of it is reassuring.

The most widespread fear is of the escape of engineered forms and of a new plague unleashed upon the world. Biologists have some standard defences against this concern. Manipulated creatures are likely to be less fit than those which have not been interfered with. After ail, if the gene gives its carriers an advantage it might be expected already to have evolved. Most farm animals and plants cannot survive outside farms, which is why the streets are not full of marauding sheep or potatoes. The same is true of bacteria and viruses. Children are injected with a live polio virus that has been 'attenuated' to make it less dangerous. Surveys of sewage show that this live virus is constantly escaping. That is the key to its success: even children whose parents do not allow them to be vaccinated are exposed to the viruses excreted by their treated friends. The attenuated virus has never survived in the wild, but depends on a supply of newly treated children. If all engineered organisms are as feeble there is not much to worry about.

Even so, it is worth remembering that every domestic animal is a pest somewhere. Cats wiped out many of New Zealand's birds. Goats have done worse in many places, feral pigs are everywhere in the subtropics and even horses can be a nuisance in California deserts. Plants are even more destructive. Everyone knows about the prickly pear in Australia, but a pretty yellow South African garden plant, the sour-sop, has done even more damage there.

The brash biologist can — and does — argue that we know enough not to repeat such mistakes. Biologists also point out that much of what engineering does is quite natural. Recombinant DNA is made every time sperm meets egg; species are not fixed entities as they evolve from one into another, and — often in bacteria and sometimes in plants — they even exchange genes by natural means. Huge numbers of bacteria are produced, mankind alone excreting ten with twenty-two zeroes after it of the minute creatures each day. Because of mutation many are genetically new and a few, because of the vagaries of reproduction, must include genes incorporated from other Species. None has spread and gut bacteria are, in the main, benign.

Viruses give fewer grounds for comfort. Most of the flu epidemics that cross the world each winter begin in China, when the human flu virus picks up genes from those of wild birds. Only when they pass from ducks to pigs to ourselves do the new mixtures cause trouble, but they are salutary reminder of our vulnerability to rare events in distant places.

The release of manipulated organisms was long delayed by such concerns. In California, crops are damaged by frost. As the air cools, patches of ice appear on the leaves around natural colonies of Pseudomonas bacteria. One bacterial gene causes this tiresome behaviour. Sometimes it changes by mutation to produce an 'ice-minus' strain that does less harm. Now an artificial ice-minus bacterium is sprayed onto plants and cuts down frost injury by displacing the natives. The gene was moved from a normal bacterium, sections cut out and the altered DNA reintro-duced. Although the bacteria are in some senses not engineered at all as the genes come from their own species, the plan caused an uproar. This irritated agricultural researchers. As they pointed out, legal controls would not allow DNA to be moved from a weed to a crop, which is what happened when the first wheat was made. After many battles the release was approved.

During the court cases it emerged that the military had  already played with Californian bacteria. They wanted to know how best to infect people. In the 1950s huge numbers of Serratia marcescens bacteria, then assumed to be harmless, were sprayed over San Francisco to see how they spread. Now it is known that Serratia can infect those already debilitated by disease and that a number of mysterious infections at the time were due to the bug. Even a natural bacterium which appears to have no ill effects is, it seems, dangerous when placed in unnatural circumstances.

And what if a new gene gets out of its own species and into another? Herbicide resistance genes might get from crop plants to their weedy relatives. For plants like potatoes, with no wild species in the Old World or in North America, that is unlikely, but oil-seed rape and sugar-beet in Britain, and sunflowers in the United States have plenty of local relatives with which they could hybridise. In places where wild turnip and oil-seed rape grow close together as many as one seed in a hundred is a hybrid and many of the plants that emerge are perfectly healthy. The Round-Up resistance gene has been crossed into the hybrids and works perfectly well with no apparent effects on survival. A spray-resistant wild turnip — perhaps the first of many resistant weeds — may be around the corner. Animal genes, too, may stray into unwelcome places. So many fish escape from farms that the genetic structure of North Atlantic salmon has already been damaged by crosses between farmed and local populations. Some plan to move anti-freeze genes from Antarctic fish to their warm-water relatives to farm them in colder and more productive waters. What might happen if escaped tropical fish hybridise with the natives?

To release manipulated beings is to play with the unknown and hence, inevitably, to take a risk. Some scientists suggest that it is so tiny as to be not worth considering. They are still in a phase of technological absolutism. Trust us, they say; bur like the engineers who developed nuclear power or drained the Florida Kverglades, or the Bourbons, they have forgotten nothing of the successes and learned nothing from the failures of history.