With its flamboyant orange-and-black wings and incredible 1,000-mile migratory flights, the monarch butterfly is one of the world's best-known and most beloved insects. And like a miner's canary, it has become a kind of biological early-warning system, succumbing to environmental changes long before humans notice them. Last week the monarch sounded another alert--fanning new fears about bioengineered crops.

In a study published in Nature, Cornell entomologist John Losey and his colleagues reported that pollen from corn made pest-resistant by the addition of bacterial genes could spell trouble for monarchs. In his experiments, Losey scattered pollen from the genetically modified corn onto milkweed--the butterfly's only food during its larval or caterpillar stage--and watched what happened with alarm. Most of the caterpillars that ate these leaves either died or were stunted.

The Cornell tests set off a flutter of concern not only for the survival of the monarchs--already threatened by logging in their winter roosts in the mountains west of Mexico City and by pesticides in their Cornbelt breeding grounds--but also over our increasing dependence on high-tech, genetically engineered food crops. "This is a heads-up," warns entomologist Fred Gould of North Carolina State University.

Approved by the Environmental Protection Agency in 1996, so-called Bt corn has become enormously popular with farmers, and now accounts for up to 25% of the U.S. corn crop, or about 20 million acres. By splicing DNA from the common soil bacterium Bacillus thuringiensis into the corn's genes, scientists have created a plant that turns out the same toxin as the bug. While the toxin is deadly to the corn borer, which costs U.S. growers more than $1 billion annually, it is harmless to humans--as well as to such beneficial insects as ladybugs and honeybees. Indeed, organic farmers have long used Bt sprays as a natural pesticide.

With hundreds of millions of dollars at stake, agritech companies aren't eager to draw sweeping conclusions from the Cornell experiments. "Obviously the work is preliminary and inconclusive," says Monsanto spokesman Randy Krotz, minimizing the possibility that corn pollen could ever be blown far enough to affect monarch habitats. But it was just such a discovery--of pollen-dusted milkweed 200 ft. from the edge of cornfields--that prompted Losey's study in the first place. Says he: "We asked ourselves, 'What would happen if the milkweed would be dusted with Bt [corn pollen]?'" His experiments quickly gave an answer: within four days, 44% of monarch larvae placed on the dusted leaves were dead, while controls survived unscathed.

Losey is eager to take the experiments into the field, to measure pollen density at various distances from its source so as to determine risk to monarch larvae at each site. Says Losey: "We have to weigh the costs and benefits [of Bt corn], then decide as a society what we want." But that decision may already have been made. The Bt gene is now regularly spliced into potatoes (as protection against the Colorado potato beetle) and cotton (against the boll weevil).

Five years after U.S. regulators approved the first genetically altered food crop, the "FlavrSavr" tomato, there are all manner of brave new foods on the way: beans and grains with more protein, caffeine-less coffee beans, strawberries packed with more natural sugars, and potatoes that soak up less fat during frying. At last count, says plant ecologist Allison Snow of Ohio State University, field trials have been conducted for some 50 gene-spliced food plants, including squash, melons, carrots, onions, peppers, apples and papayas.

But such tinkering can go awry. As even their proponents concede, spliced genes, like any other genes, can be picked up by wild species. The fear is that they will create what geneticist Norm Ellstrand of the University of California at Riverside, calls "a weedier weed"--a species, such as the superweed that turned up in France when sugar beets crossed accidentally with a wild relative, that is both harder to control and more ecologically disruptive. Scientists also fear that as use of Bt crops increases, so will resistance in the very pests they're aimed at, depriving organic farmers of a natural pesticide they'd come to trust.

Measures are being considered to avert such calamities--for example, ringing cornfields with patches of plain, old-fashioned corn so that not all pests become resistant. But these efforts haven't silenced critics, especially in Britain, where a noisy debate is raging over what the London tabloids like to call "Frankenstein foods." Last week the British Medical Association called for a moratorium on commercial planting of all transgenic crops until scientists agree on their safety. In India, Monsanto is running into a p.r. buzz saw in its efforts to introduce a Bt cotton called Bollgard--even as it wrestles with continuing protests over its stalled plans to include in its new crops so-called terminator technology that would compel farmers to buy fresh seed for each planting.

Viewing the new crops as useful alternatives to pesticides, most scientists want work on them to continue, if more cautiously. The message from the monarchs, meanwhile, is that even the most well-intentioned biotechnologies are not risk-free.