When farmers began planting Bt cotton for the first time in 1996, many were surprised to learn they couldn't plant their whole farm in the new technology.

Growers were told they had to devote a percentage of their acres to non-Bt varieties so they could provide a “refuge” for susceptible tobacco budworm and bollworm moths that could mate with potentially resistant moths.

To most producers, the idea of coddling insects you had been spending hundreds of dollars an acre to kill seemed convoluted, to say the least. Why not just plant every acre in Bt varieties and eliminate the problem once and for all?

What may have seemed counter-intuitive to growers made sense to entomologists and to producers who had been dealing with pyrethroid-resistant tobacco budworms. Many had seen first-hand what insecticide-resistant insects could do to a cotton crop in Alabama and Mississippi the previous year.

For researchers like Mike Caprio, professor of entomology at Mississippi State University, the idea of insects developing resistance to the Bacillus thuringiensis gene in Bollgard cotton was more than scientific conjecture.

Caprio was completing his Ph.D. at the University of Hawaii in the late 1980s when scientists in the laboratory where he was working found the first field resistance to a foliar-applied Bt toxin in diamondback moths.

“DiPel is the foliar Bt product that sticks in my mind, although there was a whole slew of those products back then,” said Caprio. (Few foliar Bts were sprayed in cotton much after the introduction of varieties containing the Bacillus thuringiensis gene.)

Caprio began doing computer modeling to try to determine what was happening with resistance to pyrethroids and to foliar Bt products.

“I was doing population genetics,” he said. “I was looking at genetic variation between populations, and I used the modeling as a way to understand how those differences could arise, and how they would impact resistance.”

After receiving his Ph.D. at Hawaii, Caprio did post-doctoral research at the University of California at Berkeley, working with Marjorie Hoy on deploying natural enemies of pests such as the walnut aphid in bio-control strategies.

“She was taking bio-control agents and selecting them for resistance so that if you had to go in and spray a pest you wouldn't necessarily wipe out your natural enemies,” he said. “She had quite a bit of luck with a mite predator. I did laboratory work and some modeling on a parasitoid of the walnut aphid.”

He arrived at Mississippi State in 1993 as university researchers and Monsanto representatives were preparing for the introduction of Bollgard, the name Monsanto gave to cotton containing the Cry 1A(c) Bt gene that had proved effective in controlling tobacco budworm and bollworm.

“Things were already in gear with research leading up to the release of Bollgard cotton,” said Caprio, who immediately began working with different aspects of the technology, including the development of resistance management approaches for it.

“Originally, the hope was that some sort of seed mixture would be appropriate because then you could just sort of package it in the bag,” he said. “My predecessor, Jim Mallette, published a paper with Pat Porter, a student here, that suggested a seed mixture was not a good thing.

“Well, it was not a good thing if there was larval movement, and a lot of the research demonstrated there was larval movement in cotton pests. At that point, people switched to a refuge concept.”

Much of the initial research on Bt cotton and resistant management strategies was done at Mississippi State. “Randy Luttrell (now with the University of Arkansas) and his students, including Don Parker, who's now with the National Cotton Council, worked with insect movement,” he said.

“Theo Watson in Arizona, John Benedict in Texas, Fred Gould in North Carolina all did research in this area. There were a lot of people who worked with Monsanto to try to develop appropriate refuge requirements.”

From the beginning of the registration process for the new technology, EPA was adamant about requiring a resistance management plan for Bt cotton, in part, due to pressure from organic food producers who relied on foliar Bt products. (The latter were concerned resistance could eliminate one of their primary insect control tools.)

“EPA talked about a period of between 10 and 15 years as a minimal requirement for the longevity of the product,” said Caprio. “I think Monsanto was looking at a minimum of 10 years and researchers a minimum of 15 years. But that's just one part of resistance risk assessment. The other part would be to ask how much that costs and is it worth it?”

Because Bt toxins can have substantial sub-lethal effects that could hasten the development of resistance in tobacco budworms and bollworms, resistance management plans for those toxins has focused on a high dose-refuge strategy, according to Caprio. This strategy relies on plans expressing a sufficiently high dose of the toxin to kill greater than 95 percent of susceptible and resistant heterozygous insects, while refuges of non-transformed crops allow large numbers of susceptible insects to survive.

“The 1998 scientific advisory panel to EPA suggested that a refuge should produce 500 susceptible insects for each insect expected to survive in the transgenic crop,” he said. “The emphasis has been on the size of the refuge needed to meet this or equivalent goals.”

Under the initial resistance management plan, farmers planting Bollgard cotton were required to plant 1) 96 percent of their total acres in Bt varieties and 4 percent in non-Bt or 2) 80 percent Bt varieties and 20 percent non-Bt varieties. The 20-percent refuge option could be sprayed with foliar insecticides while the 4 percent could not.

Subsequently, EPA allowed growers to plant a 5-percent “embedded” refuge option in non-Bt cotton while planting 95 percent of their acres in Bollgard and, later, Bollgard II or the Widestrike varieties developed by Dow AgroSciences' Phytogen Brand Cottonseed.

The change allowed growers to treat the 5-percent embedded block whenever the associated Bollgard, Bollgard II or Widestrike-planted field was sprayed with an insecticide, except for a foliar Bt product.

Growers were also permitted to adopt a “community” approach that allowed them to band together to set aside a block of non-Bt cotton to serve as the refuge acres for more than one field or farm.

Still, some producers and their organizations felt the requirement to plant from 5 to 20 percent of their acreage in lower-yielding, non-Bt cotton varieties placed too great an economic burden on them. They asked EPA to reduce the refuge requirements for Bollgard cotton, but the agency declined.

Caprio, meanwhile, continued to research different resistance management strategies, how they might impact resistance evolution and, more recently, to assess the risk of resistance. One area involved the size of the refuge; that is, how big can a refuge be before it begins to yield less than the desirable number of susceptible moths.

The study found that varying the number of rows in non-Bt refuges embedded in Bt cotton had an impact on the growth rate of populations of Heliothis virescens or tobacco budworms in those refuges.

Other researchers such as John Adamczyk and Doug Sumerford, research scientists with the USDA-ARS' Southern Insect Management Research Unit at Stoneville, Miss., have studied other aspects of the issues, including the efficacy of the expression of the Bt gene in different cotton varieties.

Over time, some of the perceptions regarding refuges and varieties have changed, says Caprio, who is based on the main campus of Mississippi State.

“Rather than saying this technology will last 10 years what we're more likely to do is say we're uncertain about all these different things, but, given those uncertainties, we think there's X probability that it will last 15 years,” he notes.

“We try to incorporate some of that uncertainty into our evaluation of our assessment of the risk of resistance. And from the standpoint of the academic community, I think that's pretty much where we should limit ourselves.”

Last spring, EPA announced it was approving a natural refuge option for Bollgard II and later for Widestrike. That ruling meant growers would not have to plant a non-Bt refuge along with cotton containing the Bt gene.

Caprio says his assessment of continuing the non-Bt refuge requirements vs. an alternative or natural refuges shows there was some increased risk in simply using the natural refuge. “But is that a reason for not approving the natural refuge?” he asked. “That's up to the EPA, which is responsible to all those stakeholders.”