As mentioned in the previous section, genes coding for the toxin from Bacillus thuringiensis (Bt) have been incorporated into maize (corn) to protect it from the European corn borer and into cotton to protect it from the cotton bollworm, pink bollworm and tobacco budworm. They have also been incorporated into potato to protect it from the Colorado beetle. Evidence of risks relates mainly to the possibility of pest resistance to the Bt toxin, and direct and indirect harm to non-target species.
Companies accept the possibility that the targeted pests will develop resistance. In laboratory tests, at least ten species of moths, two species of beetles and four species of flies have developed resistance to Bt toxins (Tabashnik 1994, quoted in Wolfenbarger & Phifer 2000). In the field, the diamond back moth (Plutella xylostella), a common and widespread pest of Brassica species, has developed resistance to sprays of Bt toxin. Whereas Bt sprays, which are used for pest control in organic farming, are used intermittently, the Bt toxin in Bt crops is present throughout the season. This increases pest exposure, so may increase the chances of resistance developing, especially if (as in some Bt maize, for example) the levels of Bt toxin expressed by the crop are not uniformly high or tail off towards the end of the growing season.
Companies initially assumed that the problem of pest resistance could be overcome by introducing genes coding for other types of Bt toxins into crop plants. This assumption was called into question by evidence that in some insect pests a single gene can confer resistance to four types of Bt toxin (Tabashnik et al. 1997; Fox 1997).
In response to the concerns of consumer groups and organic farmers, the US Environmental Protection Agency (EPA, responsible for the regulation of pesticidal crops) introduced a requirement for 'insect resistance management' to delay the development of pest resistance. Whereas in 1995, Bt potato was granted unconditional approval for commercial use, in 1996 Monsanto's Bt cotton was approved only on condition that farmers planted it with conventional cotton on 4% of the area, so as to provide reservoirs of Bt-susceptible pests. Subsequently, more stringent resistance management measures were imposed. In the autumn of 1996, the EPA imposed restrictions on the growing of Bt corn in cotton-growing areas of southern US to prevent additional selection pressures for Bt-resistance, since the corn earworm is also a pest of cotton (when it is known as the cotton bollworm) (Fox 1996). In 2000, the EPA increased the proportion of conventional corn that has to be planted with a GM corn crop to at least 20%, or 50% where cotton is also grown (Smith 2000; Dove 2001).
The refuge areas are intended to increase the chances that any resistant pests from the Bt crop will mate with susceptible pests from the conventional crop and produce susceptible offspring. This strategy is based on an assumption that susceptibility is the dominant trait. It also assumes that resistant and susceptible pests will mingle and will reach the reproductive stage at the same time. There is some evidence that these assumptions are not necessarily valid.
Laboratory studies showed that in the European corn borer Bt resistance may be partially dominant (Huang et al. 1999). When Bt resistant corn borer moths were mated with Bt susceptible ones, the offspring were much closer to the resistant parent than to the susceptible one in their response to the Bt toxin. Back-crossing the offspring with susceptible individuals did not result in any significant decline in resistance.
Laboratory experiments on the pink bollworm showed that although in this case resistance was recessive, resistant larvae on Bt cotton took six days longer to develop than susceptible larvae on conventional cotton (Liu et al. 1999). The researchers said that although so far Bt resistance in pink bollworm is not a problem, strategies for pest resistance management need to take into account the possibility of uneven larval development. If the slower development of resistant larvae limits the chances of crossbreeding between resistant and susceptible individuals, it could accelerate the development of resistance. If slower development reduces the numbers of resistant insects surviving over winter, it could delay the development of resistance.
Use of Bt cotton can lead to outbreaks of other pests that were previously controlled incidentally by the pesticides that the Bt toxin replaces. As a result, additional insecticidal sprays may be needed. For example, a four-year study (19961999) comparing 360 fields of Bt cotton with the same number of fields of conventional cotton in North Carolina found that although bollworm damage was more than halved in the Bt crops there was a four-fold increase in damage to the cotton bolls due to stink bugs (Bacheler 2000). Extension agents reported a considerable increase in cotton boll damage due to stink bugs in Bt cotton in southeastern US in 2000 (Hollis 2000). They recommended the use of organophosphates such as dicrotophos and methyl parathion if the stink bugs occurred in numbers above certain thresholds.
There is some evidence of harm to non-target species such as butterflies. For example, in laboratory tests at Cornell University in which larvae of the Monarch butterfly (Danaus plexippus) were fed on milkweed leaves dusted with pollen from Bt corn, nearly half the larvae died, whereas none died when fed on pollen-free leaves (Losey et al. 1999). The surviving larvae on the Bt pollen-dusted leaves consumed less than those on pollen-free leaves and grew to only half their normal size. These results attracted widespread publicity because of the Monarch butterfly's strikingly beautiful appearance. They alarmed conservationists because, although the butterfly is not endangered, it is a migratory species that overwinters in Mexico and its migratory behaviour is threatened by the loss of wooded habitat. Its larvae feed exclusively on milkweed, which is commonly found growing near corn fields. Over half its summer population occurs in the corn belt of the mid-west US.
Further evidence of the impact of Bt corn pollen on the Monarch butterfly came from research at Iowa State University (Hansen-Jesse & Obrycki 2000). Researchers there fed Monarch larvae on samples of milkweed that had previously been placed within a Bt corn field and at varying distances from the edge of the field at the time the corn was shedding pollen. Within 48 hours of feeding on pollen-dusted leaves from within the Bt corn, larval mortality was 19% compared with 0% for larvae on milkweed taken from a conventional corn crop and 3% on milkweed leaves with no pollen. Larvae less than twelve hours old and those fed on milkweed taken from within ten metres of a Bt crop were the most affected.
To counter concerns raised by these results, the US Environmental Protection Agency at first asked companies to ask growers to position non-Bt corn refuges between their Bt corn and any milkweed. However, subsequently the EPA decided such buffer zones were of little use, since:
during 2000 it became clear that milkweed in the cornfield, rather than outside it, is the preferred breeding place for Monarch, and the heavy pollen travels only a short distance, so that the amount on milkweed drops dramatically within a few metres from the cornfield so a buffer zone would be superfluous for protecting Monarch larvae and we did not continue our suggestion
(EPA official, interview, L. Levidov 2000, personal communication)
The research on the Monarch butterfly prompted research into possible impacts of Bt crops on other butterfly species, for example the black swallowtail (Papilio polyxenes) whose host plants are found mainly along roadsides at the edge of corn fields. Researchers at the University of Illinois placed potted host plants with black swallowtail larvae at various distances from a crop of Bt corn at the time it was shedding pollen (Wraight et al. 2000). The amount of pollen deposited on the host plants was estimated from the amount deposited on greased slides placed nearby. Large numbers of the larvae died over the seven days of the study - as happens normally in the field according to the researchers - but there was no correlation between larval mortality and proximity to the Bt corn or to the amount of pollen present on the host plant leaves. In laboratory tests over a three-day period, pollen dusted onto leaf discs of the host plants failed to kill black swallowtail larvae even at the highest pollen dose tested (10 000 grains cm-2).
How well any of these experiments reflect the conditions in commercial crops is unclear. Each has been criticised for methodological shortcomings. For example, the Cornell work was criticised for not reporting the levels of pollen deposited on the milkweed leaves, or taking into consideration a possible anti-feedant effect of the pollen. The value of the Illinois results was questioned, given the very high background level of larval mortality. The seven-day period of the Illinois study was criticised as an inadequate test of the likely impact in the field, given that black swallowtail larvae go through several generations in the same place each summer so are more likely to be exposed to corn pollen than the larvae of the Monarch butterfly.
There is conflicting evidence on the possibility that the impact of Bt toxin might pass along the food chain to affect species that feed on pests targeted by Bt crops. Predators of pests may die if they depend solely or to a large extent on pests that are effectively controlled by Bt toxin. This seems likely to be the cause of a decline in numbers of a predator of Colorado beetle noted in Bt potato (Riddick et al. 1998). Species that depend on a Bt-targeted pest to complete a stage of their life cycle may die if their host dies before that stage is completed. For example, in laboratory studies the larvae of a parasitic wasp (Cotesia plutellae) died when forced to develop on the larvae of Bt-susceptible diamond back moth (Plutella xylostella) fed on Bt oilseed rape (Schuler et al. 1999). This was hardly surprising since the moth larvae all died within five days of feeding on Bt plants, whereas the larvae of the parasitic wasp take seven days to develop. However, in the field behavioural factors may come into play. The parasitic wasps are attracted to the moth larvae by the chemicals released from the plants they are damaging. In wind tunnel studies, the researchers found that only 11% of the parasitic wasps were attracted to leaves of Bt oilseed rape on which Bt-susceptible moth larvae were feeding, while 89% were attracted to non-Bt oilseed rape leaves because those leaves suffered more feeding damage. With Bt-resistant moth larvae, there was no significant difference in feeding damage or wasp attraction between Bt and non-Bt plants. So in the field, behavioural factors might help parasitic wasps seek out and control Bt-resistant moth larvae because they cause more plant damage than Bt-susceptible moth larvae.
The possibility that Bt toxin may increase in activity when it is ingested by some species, rather than be broken down, has been raised by laboratory studies in which lacewing (Chrysoperla carnea) were reared on Bt-fed prey (Hilbeck et al. 1998; 1999). These found a significantly higher mortality (59-66%) for lacewing reared on prey fed on Bt corn than for lacewing reared on prey fed on non-Bt corn (37%). Similar results were obtained when the Bt toxin was incorporated directly into the lacewing diet (56% mortality, compared with 30% for lacewing fed on a Bt-free diet). The similarity in the results was unexpected since the level of Bt toxin in the corn leaves was less than 4 |g g-1 of fresh leaf whereas it was 25-100 |g g-1 of the artificial diet. The researchers speculated that the toxin might have been altered by biochemical processes in the prey in a way that was lethal to the lacewing predator but not to the prey.
There is little information on the possible impacts of Bt crops on soil organisms. Any impacts are likely to depend on the persistence of the toxin in the soil. Laboratory studies suggest that in neutral soil the toxicity of Bt toxin from Bt cotton and corn rapidly declines, so that by 120 days its effect on larval growth is 17-23% of its initial activity (Sims & Holden 1996; Sims & Ream 1997; quoted in Wolfenbarger & Phifer 2000). However, active toxin readily binds to soil particles, inhibiting breakdown by microbes (Stotzky 2000). High clay content and low pH increase the toxin's persistence (Stotzky 2000). Bt toxin may enter the soil via pollen or when Bt residues are ploughed in, but a laboratory study suggests it may also enter the soil by means of exudates from the roots of Bt corn (Saxena et al. 1999), in which case it may be present throughout the cropping season.
Was this article helpful?