To the editor:

A news item by Jayaraman et al. in the February issue (Nat. Biotechnol. 23, 158, 2005) reports concerns in India about the potential rapid evolution of insect resistance to transgenic cotton that produces Bacillus thuringiensis (Bt) toxin. I would like to correct a misleading statement of mine, which was quoted in this article.

Contrary to my statement, the Indian government does require planting of non-Bt cotton refuges to delay insect resistance to Bt cotton. The rule is that every Bt cotton field will be surrounded by a belt of non-Bt cotton that covers at least 20% of the total area planted with cotton1. As in the United States, in India the refuge strategy aims to promote survival of susceptible insects and thereby slow evolution of resistance2.

Thus far, field-evolved resistance to Bt crops has not been documented for insect populations2, even though Bt cotton and Bt corn have grown on more than 90 million hectares worldwide since 1996 (ref. 3). The refuge strategy has probably contributed to this success.

In India, compliance with the refuge strategy is uncertain1. Thus, the issues raised in the article about the risk of insect resistance are relevant, particularly if several crops other than cotton are also genetically modified to produce Cry1Ac, the toxin in Bt cotton. Non-Bt varieties of some of these other crops may thwart resistance by acting as refuges for the cotton bollworm4 (Helicoverpa armigera), the major pest in India targeted by Bt cotton. Similar to the modeling predictions for the cotton bollworm cited in the article, genetic models predicted that a closely related cotton pest, Helicoverpa zea, could evolve resistance to Bt cotton in the US in three to four years, assuming only a 4% refuge5. Larger refuges of non-Bt cotton and non-Bt varieties of other crops have likely helped to delay this pest's resistance for more than nine years2,6.