The butterfly flap

In 1999, research suggested that genetically-modified corn might be killing off the monarch butterfly. What followed was exactly the kind of argument that is good for public health
July 19, 2003

In June the British government launched a series of public consultations about genetically-modified crops. In theory, the result of these consultations-so far, rather poorly attended and dominated by green lobby groups-will be taken into account later this year when the government makes its long postponed decision on whether to allow the sale and cultivation of GM crops in Britain. One of the most worrying claims of the anti-GM groups is that pollen from genetically-modified crops like maize and rapeseed will "contaminate" local varieties, or will harm insects, birds and animals. As evidence, Greenpeace and others point to a dramatic experiment in the cornfields of America four years ago when a young researcher found that pollen from corn genetically modified to contain a pesticide killed the larvae of America's favourite insect, the monarch butterfly.

The brilliance of an American summer would indeed be dimmed without these gaudy orange and black creatures dipping and diving in the meadows and hedgerows. And an American autumn would not be the same without the monarch's southern migration, an incredible journey when tens of millions of them fly 80 miles a day to spend the winter in a Mexican forest roost. So the news that big agriculture might be killing off America's beloved butterfly seemed a devastating blow to the case for GM crops. A closer look suggests a somewhat different story.

In the spring of 1999, as the monarchs embarked on their return flight north, a young Cornell University entomologist named John Losey reported in the journal Nature that the monarch's future appeared to be endangered; not from urban sprawl or toxic waste, but from eating the pollen of genetically-modified corn. At the time, 20m acres of American farmland, representing a quarter of the US corn crop, had been planted with seeds that included a toxin-producing gene from the common soil bacterium, Bacillus thuringiensis, or Bt. The insect-poisoning power of Bt had been known for over a century and the first commercial spray was developed in Europe during the second world war. It even became a favourite of organic farmers. Half a century later, there were 182 Bt products registered by the US Environmental Protection Agency (EPA).

Two other big crops-cotton and potatoes-had also been fitted out with the Bt gene. But in corn, the Bt toxin was designed primarily to kill the European corn borer, a caterpillar that destroys more than $1bn worth of the crop each year. The toxin punctures the delicate membranes of the caterpillar digestive tract, causing it to wither and die.

Most of the monarchs born in the midwest corn belt start life on a milkweed leaf in or around the edges of a farmer's land. When the corn sheds its pollen during July and August, pollen grains containing the Bt toxin are blown by the wind onto milkweed leaves. From earlier studies, Losey knew that Bt toxin could harm butterflies and moths, and he wondered if the monarch larvae might also suffer.

In a no-frills experiment at his laboratory at Cornell in upstate New York, he fed monarch larvae with Bt pollen. If they showed signs of harm, he intended to do more research in the field. In his lab, he misted milkweed leaves with water and sprinkled on the Bt corn pollen to a density that looked like the pollen he had observed on the milkweed in a cornfield. He then placed five three-day-old monarch larvae-caterpillars no bigger than a raindrop-on each milkweed leaf and watched them feed. The experiment was repeated five times. After four days, nearly half of the larvae were dead. Those that survived were half the weight of his control group feeding on milkweed leaves with no pollen. Larvae fed on leaves sprinkled with conventional hybrid corn pollen were still munching away, apparently no worse off.

Losey and his Cornell researchers were well aware of the uproar that the deaths of even a few monarch caterpillars could cause. Bt crops-cotton, corn and potatoes-are the green-friendly pride of the biotech industry. The idea is that farmers who plant them use less pesticide because Bt kills the corn borers. (This, however, remains in dispute. Industry estimates of the cut in pesticide use tend to be higher than independent research.)

The use of Bt varieties has increased dramatically since the first planting in 1996. They had become the industry's banker at a time, in the late 1990s, when opposition to other products was gathering force, especially in Europe. Short of some human health hazard, it was hard to think of a bigger propaganda setback than monarchs being killed by Bt corn.

The monarch's migration from Mexico to the Canadian border and back is closely monitored every year. En route, the butterflies mate in an elegant embrace and the females lay eggs-usually on the underside of milkweed leaves for protection. The monarch has its own website, Monarch Watch (, started by Orley "Chip" Taylor, an entomologist from the University of Kansas. Every year, monarch watchers record the first landing of the butterflies on to milkweed-the monarch caterpillar's food and home-as well as the first and last hatch. The butterfly's egg is a creamy dome, a natural wonder for lepidopterists.

During the summer, when the butterflies are busy eating and reproducing, they live for up to six weeks. Those born in September instinctively forego the debilitating rituals of courtship and sex, saving their energy for the 3,000-mile flight south. Their determination to make this arduous journey is a puzzle to entomologists. How do the monarchs know when to begin the migration? How do they find their way? What proportion of monarchs survive the journey? Such mysteries add to the creature's popularity.

To test public reaction to their experiment, Losey and his co-researchers at Cornell first shared the results with colleagues. All were in favour of publication. However, a senior entomology professor at Cornell, Anthony Shelton, warned the younger researcher that he didn't have a "story." Shelton, a believer in biotech, would become increasingly unhappy that Losey's experiment had been confined to a laboratory. The results, he would complain, were "not pertinent to the real world."

This criticism from a peer put Losey in a bind. As an assistant professor, aged 37, he was up for tenure in 2004. He needed to publish his work, but he also needed the support of senior colleagues. Finally, Losey decided to write the report and send it to journals. "It would have been irresponsible not to take the results to our peers and the public," he said later.

In search of a publisher, Losey's first stop was Science magazine, the journal of the American Association for the Advancement of Science. Its editors were not interested. Next stop was the British journal Nature, where Losey had published before. Its editors were intrigued and sent the report to two reviewers. They urged publication. After minor changes, the report-just seven paragraphs long-was published in March 1999 under the title, "Transgenic Pollen Harms Monarch Larvae." Losey warned that Bt corn pollen could have "profound implications for the conservation of monarch butterflies," a message provocative enough to make the front pages in monarch-conscious newsrooms across America. "Butterfly Deaths Linked to Altered Corn," warned the Boston Globe. "Gene Spliced Corn Imperils Butterflies," said the San Francisco Chronicle. The New York Times put a picture of a monarch on its front page with the caption "Bambi of the insect world." (The BBC and several British newspapers picked up the story too.)

The anti-biotech forces leapt into action. Greenpeace called for an immediate ban on the planting of Bt corn. Volunteers dressed up as monarchs that collapsed as they were "felled by killer corn." Green activist groups, such as the Union of Concerned Scientists (UCS) and Environmental Defense, noted that Losey's results demonstrated how the government's insecticide regulatory department, the Environmental Protection Agency (EPA), had failed to address the real risks to the monarch-and other insects-before allowing Bt corn into the environment. "One cannot help but wonder what other, perhaps less obvious, environmental impacts of genetically-engineered crops have been missed by the EPA," said the UCS biotech director, Margaret Mellon.

Before Bt products came on the market, the EPA had examined public and company reports of the effects of the toxins on a variety of organisms that might be found close to Bt crops. They included birds, fish, honey bees, ladybirds, parasitic wasps, lacewings, springtails, aquatic invertebrates and earthworms. The EPA had concluded there were "no reasonable adverse effects" to humans, the environment or any organism that Bt was not supposed to kill. Scientists knew that Bt toxins could be harmful to the larvae of butterflies, but the EPA had looked primarily at the exposure of larvae eating leaf tissue, not pollen. The agency had considered the possibility of pollen drift from Bt cornfields, but concluded the pollen was not toxic, even at relatively high doses. EPA did not specifically require tests for effects on the monarch larvae because it did not believe monarchs were likely to be present around cornfields.

Losey had identified a gap in the research, and the green groups poured through it. Mellon of UCS recalled, "We worked hard to make this a high-profile issue because without media attention we knew nothing would be done. We saw the findings as an illustration of how superficial risk assessment for GM foods was." Mellon's group reminded the public that Losey's research could spell trouble for other lepidoptera-moths and butterflies feeding in the vicinity of Bt-cornfields. (On the US endangered species list were 19 threatened butterflies and moths.) And only the year before, Swiss scientists had reported laboratory results showing Bt corn was harmful to green lacewings, insects that feed on pests, including the European corn borer. In their EPA submission, the companies had persuaded the government agency that lacewings would be immune to Bt toxins. Incredibly, the companies themselves had never assessed the risk to America's favourite butterfly-or if they had, the results had not been made public.

As spring gave way to summer in 1999, Losey's study came under scathing attack from Cornell's Anthony Shelton, who was increasingly agitated by the media fuss the report had generated. In testimony to Congress, he suggested that Losey was no better than a rumourmonger, invoking Rumour in Shakespeare's Henry IV, part two. Other scientists were also upset. A follow-up letter in Nature from a British biologist counselled the need for more "rigour" and warned that "preliminary observations should not be over-interpreted."

In a Cornell University press release, Shelton attacked Losey's experiment: "If I went to the movies and bought a hundred pounds of salted popcorn, because I like salted popcorn and then I ate those salted popcorn all at once, I'd probably die," Shelton was quoted as saying. "Eating that much salted popcorn simply is not a real-world situation, but if I died it may be reported that salted popcorn was lethal. The same thing holds true for monarch butterflies and pollen. Scientists need to make assessments that are pertinent to the real world... Few entomologists or weed scientists familiar with the butterflies or corn production give credence to the Nature article."

The biotech industry also attacked the artificial nature of Losey's data, especially the fact that the monarch larvae had eaten the pollen in the controlled environment of Losey's laboratory. It was "highly likely that in the natural setting most monarch larvae would never encounter any significant amounts of corn pollen," declared an industry press release.

But this was a bluff. The industry didn't really know what happened to monarch larvae feeding on milkweed leaves in or near Bt cornfields because the research hadn't been done. A big company involved in Bt crops, like Monsanto, might have been expected to carry out studies, but there was nothing in the public record. Even so, Monsanto joined in the attack on Losey, asserting-without proof-that the exposure of milkweed to corn pollen was "very low because only a small portion of milkweed grows in close enough proximity to cornfields for exposure to pollen."

The biotech industry public relations machine would offer the information that more monarch butterflies were killed colliding with car windscreens "than ever encounter corn pollen." In fact, the biggest hazard for monarch larvae is to be eaten by other insects. Fewer than 10 per cent survive to adulthood.

Losey does not seem to have been out to "get" the biotech companies. After Nature had accepted his report, but before it was published, Losey contacted the US-based Monsanto and the Swiss-based Novartis, the two companies involved in Bt corn, to let them know of the upcoming article. "I wanted to be above board and not blindside them," he said. Monsanto and Novartis hastily dispatched staff scientists to visit Losey and argue that the science was not "robust enough" for the generalisations he had made. "They wanted more detail," said Losey. "There was clearly a feeling that we should not publish... that we should wait until we got more data."

After the publication of his report, Losey was quoted in Monsanto's PR Newswire as one of "several academic experts who have urged caution when interpreting the results" of his own study. Losey had, indeed, always acknowledged that his experiment was a preliminary study needing extra research. He made this point again and again in media interviews.

The young Cornell researcher took the industry's flak but, in fact, he had not been the first to report that Bt corn pollen had proved lethal to monarch larvae. Another team of researchers, John Obrycki and Laura Jesse, from Iowa State University, had also fed monarch larvae with milkweed sprinkled with corn pollen and some of their larvae had died. In their three-year study, they attempted to recreate field conditions. They put potted milkweed plants in cornfields during the corn's pollen shed, then took the plants back to the lab and put larvae to feed on the leaves. Some had died. They had finished their study and reported their results to colleagues and the industry before publication of Losey's Nature paper. Their work was not published for another year, by which time it generated more alarm. They would claim to have the first evidence that transgenic Bt corn naturally deposited on milkweed in a cornfield causes significant mortality.

But the industry complained that the Iowa work was not a realistic field test, either. Some researchers agreed. Kevin Steffy, an entomologist at the University of Illinois, complained that the media had overreacted to both studies. "I'm tired of the press making an issue out of findings that don't describe the real world... I also am dismayed by some of the sweeping conclusions the authors make... If scientific evidence reveals negative impacts of transgenic crops, then let the chips fall where they may. But let's be very careful about interpretations of scientific studies."

Whatever anyone thought of the Losey or Iowa study, the researchers had clearly identified a potential hazard for the monarch butterfly. The biotech companies tried to control the damage, calling for more research and putting up funds to convene "an inclusionary process and have third parties develop the data." Several academics, including Losey and Chip Taylor of Monarch Watch, were invited to carry out studies funded 60 per cent by industry with the rest coming from government and other sources.

The idea that the companies were now funding public research into a biotech hazard was a significant departure from the cosy relationship they had enjoyed thus far with US government regulators from the Food and Drug Administration (FDA) and the EPA. Both agencies had required only voluntary company research by their own in-house scientists-and the results were kept confidential whenever a company claimed trade secrets. But in agreeing to this new, more open way of doing things, the biotech industry was hardly running up a white flag. Company scientists and many outside researchers still believed that they would be proved right-that Bt corn was not a hazard to the monarch.

The ecological risk to an insect depends on the possibility of exposure to a known poison, and then the impact of that dose. Losey's brief laboratory experiment had no real information on either of these criteria. And there were no studies even on the chances of the tiny tiger-striped mite of a monarch larvae being born on a milkweed leaf close enough to a Bt cornfield. Nor on whether the birth of the monarch larvae was likely to coincide with pollen shed from the corn.

The larval stage lasts between 12 and 16 days and the corn plant sheds pollen for about seven to ten days. Even if these two events overlap, the question is still whether enough pollen would settle on the milkweed leaf-and remain there during the time the larvae is feeding-to have a harmful effect. The pollen density on the milkweed leaves used in Losey's experiment was a very rough estimate-when he was sprinkling pollen over milkweed leaves he kept going until the leaf in the lab looked roughly like the leaves he had seen in the cornfield.

In Losey's lab, the larvae were force-fed for four days on leaves constantly covered in pollen. In the field, the number of grains is not likely to remain constant. Pollen blown onto a leaf by the wind is also blown off again. Pollen can also be washed off by a shower of rain, or even a heavy dew. In addition, not all pollen from Bt corn contains the same amount of Bt toxin. At the time of Losey's study, there were many types of Bt corn made by agbiotech seed companies, including Dow AgroSciences, Monsanto and Novartis. Each type was slightly different.

Losey chose pollen from a Bt corn known as Bt11, made by Novartis, because that was the corn pollen available at Cornell at the time he began his study. But another type of Bt corn had very different characteristics-especially for a monarch caterpillar. Known to its Novartis inventors as Event 176 and sold under the trade name of "KnockOut," this type of Bt corn was modified to produce a much larger dose of toxin in the pollen. Although the corn borer does its damage to the stalk, pollen grains tend to collect at the base of the cob where the corn borer also feeds. Scientists created this extra toxin by adding a special poison-boosting promoter that increased the poison in the pollen grains. Event 176 contained up to ten times more toxin in the pollen than the other types. So, as well as the question of caterpillar exposure, the variable toxin levels of different products had to be taken into account too.

To limit the damage inflicted by Losey's Nature article, the industry needed to act swiftly to complete the research-which they hoped would prove him wrong-during the 1999 growing season. Field studies had to be in place in late July and early August when the corn shed its pollen, otherwise the work would have to wait for another year. By the autumn, the preliminary papers were ready. In November, the scientists reported to a symposium in Chicago designed by the industry to be as open as possible, including the media, outside academics not involved in the studies and critics of biotechnology.

Not one study had demonstrated the kind of mortality in monarchs that Losey had found. But the impact of that news became blurred. In their eagerness to be first to define the results, the industry's public relations people had called a few newspapers the day before with their interpretation. On the day of the symposium, several major newspapers, including the Los Angeles Times, the Chicago Tribune and the St Louis Post-Dispatch reported that the meeting would conclude that Bt pollen posed little risk.

But the industry had been over-hasty. In fact, a number of scientists at the Chicago meeting found cause for continued alarm. Of three Bt corn "events," the strongest-Novartis's Event 176-produced pollen that was found to be highly toxic to monarch caterpillars. Pollen from the two others was found to be less toxic. Some researchers pointed out that Bt toxins in pollen lose their potency after a week and the study had not made allowances for such a change since the pollen had been stored.

Becky Goldberg, of Environmental Defense, summed up the feeling of the green groups: "It appears that the real questions will only be addressed if there is funding for research on Bt corn pollen and butterflies independent of the industry." Such funding would "also help to insulate researchers from the pressures of commercial interest."

The industry still had a fight on its hands. The initial five-year registration of Bt corn-the first approval by the EPA-was running out. Losey's paper and the inconclusive results of the Chicago meeting now prompted the EPA to ask all seed companies producing Bt corn to submit new research about the toxicity of corn pollen, showing the level at which monarchs would be exposed and the potential impact of that level on monarch populations. The studies had to be completed by the spring of 2001 so that EPA could make a decision on the future of Bt corn by the autumn. Farmers had to know by then in order to purchase seed in time for 2002.

The biotech industry and the US department of agriculture each put up $100,000 to fund 26 academic and government scientists, including John Losey. They would write six papers on all aspects of Bt corn's possible threat to the monarch butterfly. Their work would be submitted to the Proceedings of the National Academy of Sciences, which requires two outside reviewers for each paper. Industry researchers were confident that the outcome this time would clear Bt corn, but they couldn't be certain. In the meantime, green groups unearthed yet another surprise hazard.

By the end of 1999, Losey's monarch butterflies had spurred the anti-biotech forces to a new offensive. Funds flowed into the green group coffers from foundations flush from a booming stock market. Friends of the Earth canvassed for members with a new campaign. "How safe is the food you eat?" asked the letter. "If deadly toxins that kill butterflies are being introduced into our food supply, what effect are these toxins having on you and your family?... The scary answer is that no one really knows."

Some scientists who had been involved in biotech research from the beginning were worried that companies like Monsanto had skipped too easily through the regulatory hoop. They were quietly pleased to see the anti-biotech forces creating a fuss, and the media taking it seriously. More research was needed, they believed. The louder the protests, the better the chances of a mid-course correction and more funds for their own inquiries. It was a good time for activists to rattle the scientific foundations of GM.

Another type of Bt corn product, traded as StarLink, made by the European seed company Aventis, provided the next opportunity. All types of Bt corn produce crystalline proteins made from the Bt gene and known by the prefix "Cry," for crystalline. When a company applied to the EPA for approval of a Bt corn, the Cry protein was put through a rather crude human allergy test. The proteins were placed in an acid solution that mimicked human stomach fluids. Usually, they readily broke down into harmless amino acids. This told researchers that there was no likelihood of the protein staying around long enough to cause an allergic reaction in humans.

Cry9c was different. It remained stable for over an hour in the acid solution, which gives the body time to react. StarLink produced the Cry9c protein. It was not poisonous to rats, nor was its biochemical structure similar to the majority of food allergens, so its stability in stomach acid did not mean it would automatically cause an allergic reaction, even in sensitive human stomachs. But when Aventis applied for a licence in 1997, the EPA could not be sure, so it approved StarLink for animal feed only. The approval implied that StarLink corn should be kept separate in the US grain system, a zero contamination standard like the standard the EU wanted for US grain imports.

After speaking with corn farmers about the difficulty of separating one corn type from another, Larry Bohlen, of Friends of the Earth, decided to test food products to see if any StarLink had crept into the food grains. In summer 2000, Bohlen went to his local supermarket store in Silver Spring, Maryland, bought a basket of corn products-cereals, chips, corn muffins and taco shells-and sent them to a lab to get their DNA fingerprints.

The results came back a few weeks later. One of the packets of Kraft Foods' taco shells tested positive for StarLink DNA. Bohlen's group announced their discovery at a press conference on 18th September, and the agbiotech industry was suddenly at the centre of another storm. Although StarLink represented less than one per cent of the US corn harvest in 2000, the StarLink discovery led the CBS evening news, and the next day stories of tainted tacos and contaminated tortillas appeared in papers across the nation.

There was no one to blame but the EPA. They had approved StarLink for animal feed only at a time when any midwest farmer could have told them it was impossible to segregate the grains. Some grains were bound to mingle, either in the mechanical harvesters, the trucks that took the grain to the granaries, the granaries themselves, or the containers that shipped the grain to the processing plant.

Over the next six months, 17 people would complain of having allergic reactions after eating taco shells. One woman went into anaphylactic shock after eating three enchiladas made with corn tortillas. The cases were investigated by the centres for disease control in Atlanta and the FDA, which said they could find no evidence that StarLink was responsible.

Meanwhile, Aventis voluntarily withdrew the product and started to pay out millions in compensation. The US department of agriculture bought hundreds of thousands of bags of corn seed that contained traces of Cry9c, at a cost of $20m. Japan and Korea halted imports of StarLink. In a new evaluation, the EPA decided that Cry9c had a "medium likelihood" to be a human allergen even though there was a "low probability" of sufficient quantities entering human food.

But on 7th March 2001, the EPA said that it would no longer split registration for human and animal food. In future, whatever was fit for Daisy to chew had to be fit for humans too. The taco scare was over. The industry could return to the monarch butterfly.

By September 2001, drafts of the six papers that the EPA had asked for on Bt corn and the monarch were ready for distribution to the media. The scientific reports concluded that there was no immediate significant risk to the monarch from the two most commonly grown types, Bt11 and Monsanto's Mon810, supporting the earlier results rushed through in 1999. Moreover, the studies showed that monarch caterpillars would have to be exposed to pollen levels on milkweed leaves greater than 1,000 grains/cm2 before they would show any toxic effects. And, although caterpillars were indeed found on milkweed during the one or two weeks when pollen is shed by corn, the pollen levels on milkweed leaves were found to average only about 170 pollen grains/cm2 in cornfields.

Reports from several of the studies showed much lower concentrations, even within the cornfield. Overall, the researchers estimated that "fewer than one per cent of all North American monarchs would be affected by doses of Bt pollen high enough and at the right time to even see a subtle growth effect." A New York Times headline declared, "Data on Genetically Modified Corn Reports Say Threat to Monarch Butterflies is 'Negligible'." However, KnockOut Bt corn-Event 176-was judged to be harmful to monarch larvae at concentrations of only 10 grains/cm2. Novartis, the producers of KnockOut, had never managed more than a 2 per cent market share, and announced that it would be phased out.

Losey and Obrycki were not about to withdraw, however. In their joint paper for EPA, they reported that monarch larvae might still be in danger from eating pollen mixed with the corn plant's anthers-the male organ that produces the pollen. The anthers, it turned out, were much more toxic than pollen grains. They urged the EPA to grant only a one-year extension for Bt corn-until further research could throw more light on the anther issue.

Another researcher, Mark Sears of the University of Guelph in Canada, disagreed. His study found that the anther parts on milkweed leaves would be too big for the caterpillar to eat. "To a caterpillar, an anther is about as big as a city bus," he said. "Maybe some of the larger caterpillars eat them but we haven't seen any evidence of that." That debate would continue.

Chip Taylor of Monarch Watch continued to worry about the long-term effects of the Bt pollen. He warned that the monarchs might survive but be harmed in some other way-possibly suffering weakened digestive systems. They might be unable to fly the long migration route, or perhaps not be able to reproduce in the spring. In the end, however, Taylor conceded that Bt corn was "probably not" the monarch's greatest hazard; it was more likely the weather. Monarch populations take a severe dip during droughts. He estimated, for example, that the population had fluctuated from 28m wintering in Mexico in 2000 to nearly 100m a year later.

In October 2001, the EPA reapproved the five Bt corn types on the market for five more years. Losey and co had lost a battle but, at the same time, the EPA asked for more research on the long-term effects. Green groups complained that the EPA was at fault for not having paid more attention to the monarch before approving Bt corn. The EPA replied that regulating pesticides is a hazardous business. As one EPA official put it bluntly, "You can't test everything."

But most of those involved-academics, industry and other environmentalists-thought the monarch case was a "blueprint" for how to do research in the public interest. Margaret Mellon of the UCS agreed. "It brought scientists, environmental and government folks together with industry, found a pot of money, set a research agenda and got it done."

The fear of harming wild life and the contamination of traditional crops is only one concern about the biotech revolution. There are plenty of other things the public should continue to demand besides better government regulation before the new products reach the market. They include proper labelling of the new foods; ensuring old seeds are properly preserved in public seeds banks and providing appropriate GM crops to developing countries at a price they can afford. Enlightened new private and public partnerships, such as the one that happened by accident in the case of the monarch, are needed in testing the new products and in sharing the tools of innovation.

But in this case, at least, the story of the monarch butterfly ended as an example of how co-operation between publicly-funded researchers and private industry can help us to negotiate the real perils and possible promise of the biotech harvest.