Report

Suppression of Cotton Bollworm in Multiple Crops in China in Areas with Bt Toxin–Containing Cotton

See allHide authors and affiliations

Science  19 Sep 2008:
Vol. 321, Issue 5896, pp. 1676-1678
DOI: 10.1126/science.1160550

Abstract

Transgenic cotton that has been engineered to produce insecticidal toxins from Bacillus thuringiensis (Bt) and so to resist the pest cotton bollworm (Helicoverpa armigera) has been widely planted in Asia. Analysis of the population dynamics of H. armigera from 1992 to 2007 in China indicated that a marked decrease in regional outbreaks of this pest in multiple crops was associated with the planting of Bt cotton. The study area included six provinces in northern China with an annual total of 3 million hectares of cotton and 22 million hectares of other crops (corn, peanuts, soybeans, and vegetables) grown by more than 10 million resource-poor farmers. Our data suggest that Bt cotton not only controls H. armigera on transgenic cotton designed to resist this pest but also may reduce its presence on other host crops and may decrease the need for insecticide sprays in general.

Transgenic crops carrying insecticides have become an important tool for insect pest management worldwide and, in 2007, were grown on a total of 42.1 million ha, accounting for about 37% of all the transgenic crops (1). One of these, Bt cotton, produces insecticidal toxins from Bacillus thuringiensis (Bt) and occupied 14 million ha worldwide and 3.8 million ha in China in 2007 (1). Bt cotton can suppress populations of a target pest with a narrow host range, e.g., pink bollworm (Pectinophora gossypiella)(2), but its long-term and wider ecological consequences are unknown.

The cotton bollworm, Helicoverpa armigera, is one of the most serious insect pests of cotton, corn, vegetables, and other crops throughout Asia. There are four generations of H. armigera per year in northern China. In general, wheat is the main host crop of first-generation H. armigera larvae, and cotton, corn, peanuts, soybeans, and vegetables are the major hosts for subsequent generations (3). Because of its long-distance migrations between provinces and dispersal among different host crops, provincewide outbreaks of H. armigera on cotton and other crops were common in the early 1990s in China (3). Bt cotton was first approved for commercial use in 1997 in China and remains the only Bt crop registered. By 2001, Bt cotton had been extensively planted, especially in northern China, which resulted in increased yields and decreased use of insecticides (4).

We conducted long-term and large-scale field monitoring of H. armigera during 1992–2007 in multiple crops in six provinces (Hebei, Shandong, Jiangsu, Shanxi, Henan, and Anhui), covering 38 million ha of farmland in northern China (fig. S1), in which 3 million ha of cotton and 22 million ha of other host crops (corn, peanuts, soybeans, and vegetables) were cultivated annually by more than 10 million small farmers. Our results indicated that both the egg density of H. armigera on cotton and the larval density on other major host crops were negatively correlated with the number of years after the introduction of Bt cotton in the period of 1997–2006 (Figs. 1 and 2). Before Bt cotton commercialization, the H. armigera population was fairly high on cotton and other host crops over the period from 1992 to 1996. However, population density of H. armigera was drastically reduced with the introduction of Bt cotton, especially during the period from 2002 to 2006 (table S1). Using stepwise regression, we evaluated the contribution of temperature, rainfall, and deployment of Bt cotton on the population density of H. armigera in six provinces (Table 1). For all six provinces in northern China, Bt cotton acreage correlated best with the reduction in H. armigera populations (Table 1). For the second and third generations, the deployment of Bt cotton contributed more to the reduction of H. armigera density than temperature and rainfall during 1997–2006 and was the key factor for its long-term suppression in all the six provinces of northern China (R2 = 0.41 to 0.91, P < 0.05; Table 1). These results indicate that the regional occurrence of H. armigera on cotton and other major host crops in northern China was suppressed by the deployment of Bt cotton.

Fig. 1.

Egg densities of H. armigera from 1997 to 2006 on cotton in northern China. (A) Relation between egg density of the second generation (⚫) and planting year of Bt cotton. Linear model of egg density (black line), y = 157,076.05 – 78.21x, F = 32.16, df = 1,549, P < 0.0001, R2 = 0.06. (B) Relation between egg density of the third generation (⚫) and planting year of Bt cotton. Linear model of egg density (black line), y = 94,644.36 – 47.15x, F = 26.42, df = 1,558, P < 0.0001, R2 = 0.05. Data are means ± SEM. Values in parentheses are the numbers of sampling sites for each year.

Fig. 2.

Larval densities of H. armigera from 1997 to 2006 on corn, peanuts, soybeans, and vegetables in northern China. (A) Relation between larval density of the second generation (⚫) and planting year of Bt cotton. Linear model of larval density (black line), y = 480,293.95 – 239.28x, F = 16.50, df = 1,466, P = 0.0001, R2 = 0.03. (B) Relation between larval density of the third generation (⚫) and planting year of Bt cotton. Linear model of larval density (black line), y = 551,611.74 – 274.83x, F = 21.45, df = 1,462, P < 0.0001, R2 = 0.04. Data are means ± SEM. Values in parentheses are the numbers of sampling sites for each year.

Table 1.

Effects of temperature, rainfall, and deployment of Bt cotton on the population density of H. armigera in northern China. Stepwise regression analysis was used for analyzing the association between population density (egg density on cotton or larval density on other host crops) of H. armigera and temperature (Temp.), rainfall, and deployment of Bt cotton. F, generation; R2, coefficient of determination. Only variables from which the regression coefficient met the criteria of P < 0.05 are shown. NS, without significant effects (P > 0.05) on population density. + and - represent positive and negative associations between the population density and the factors, respectively. *P < 0.05; **P < 0.01.

View this table:

We also sampled H. armigera in cotton fields from 1998 to 2007 at Langfang Experiment Station in Hebei Province (5). The densities of eggs on Bt and non-Bt cotton and larvae on non-Bt cotton were negatively associated with the number of years after Bt cotton commercialization (R2 = 0.52 to 0.63, P < 0.05). The population density of H. armigera can be described by the linear regression model (Fig. 3). The data also showed that the densities of H. armigera eggs were not significantly different between Bt and non-Bt cotton over the period of 1998–2007 (P > 0.05) (Fig. 3A). However, larval densities on non-Bt cotton were significantly higher than those on Bt cotton from 1998 to 2006 (P < 0.05) (Fig. 3B), with an exception in 2007 when the population density was low and larval density was not significantly different between the two treatments (P > 0.05). Using Bt cotton also reduced the duration of H. armigera's oviposition period on cotton, because of decrease of moth density. Three peaks of egg density, representing the second, third, and fourth generations, respectively, were detected each year from 1998 to 2000, and in recent years, there was only one oviposition peak evident in the second generation, and no evident peak in generations 3 or 4 (fig. S2). The abundance of each generation and the peak duration of the third and fourth generations decreased linearly as Bt cotton commercialization proceeded through 1998 to 2007 (fig. S3). Thus, all data indicate that the commercial use of Bt cotton in northern China was associated with long-term areawide suppression of H. armigera after 10 years.

Fig. 3.

Egg and larval densities of H. armigera on cotton at Langfang site, Hebei Province, China, from 1998 to 2007. (A) Relation between egg density on Bt cotton(red circles) and non-Bt cotton (black circles) and planting year of Bt cotton. Linear model on Bt cotton (black line), y = 185,476.90 – 92.42x, F = 69.05, df = 1,58, P < 0.0001, R2 = 0.54. Linear model on non-Bt cotton (red line), y = 171,365.94 – 85.37x, F = 62.59, df = 1,58, P < 0.0001, R2 = 0.52. (B) Relation between larval density on Bt cotton (red circles) and non-Bt cotton (black circles) and survey years. Linear model on non-Bt cotton (black line), y = 87,107.86 – 43.41x, F = 97.56, df = 1,58, P < 0.0001, R2 = 0.63. Data are means ± SEM. There are six samples for each point in the graphs.

Regional control of H. armigera in multiple crops in China has been attained in recent years through the use of Bt cotton. Our results suggest that Bt cotton led to reduced populations of H. armigera not only on cotton but also on other host crops. This may be because cotton usually is the main host for the moths of the first generation to lay eggs and acts as the source of the subsequent generations on other host crops (6). Bt cotton kills most of the larvae of the second generation and, accordingly, works as a dead-end trap crop for H. armigera population. Interest in trap cropping, a promising agroecological approach for insect pest control, has increased considerably for modern agriculture (7, 8), but few trap crops were used on such a large scale as that of Bt cotton in northern China, which shows that Bt crop can have a great advantage to expand the traditional view of a trap crop. This dependence on Bt cotton might also contribute to a reduction in both occurrence of H. armigera and the need for insecticide sprays in non-Bt host crops such as corn, soybeans, peanuts, and vegetables.

However, a major challenge for planting Bt cotton for pest control is the potential for insects to evolve resistance to Bt. Continuous monoculture of varieties that express the same Bt toxin could select for resistance, particularly when the amount of Bt toxin decreases as the plants age (9, 10). A promising resistance management strategy entails the use of plants with a high dose of toxin in combination with the maintenance of “refuge” crops that encourage proliferation of Bt-susceptible insects within the pest population (1113). To this end, the U.S. Environmental Protection Agency requires that each cotton farm set aside some land for cotton that does not produce Bt if farmers plant transgenic Bt cotton producing Cry1Ac toxic protein (1416). Although successful in the United States (17), this strategy is difficult to implement in China because of the challenges associated with educating and monitoring millions of small farmers. In China, a multiple cropping system consisting of soybeans, peanuts, corn, and vegetables is common. These crops also serve as hosts for H. armigera, and, because they do not express Bt toxin, they serve as refuges for nonresistant insects (10). Because cotton is not the only host crop, Bt cotton comprises about 10% of the major host crops in any province or throughout northern China. This accidental approach to refuge management appears to have, so far, warded off the evolution of resistance (10). Nevertheless, as a result of decreased spraying of broad-spectrum pesticides for controlling cotton bollworm in Bt cotton fields, mirids have recently become key pests of cotton in China (18, 19). Therefore, despite its value, Bt cotton should be considered only one component in the overall management of insect pests in the diversified cropping systems common throughout China.

Supporting Online Material

www.sciencemag.org/cgi/content/full/321/5896/1676/DC1

Materials and Methods

Figs. S1 to S3

Table S1

References

Data Files S1 to S7

References and Notes

View Abstract

Navigate This Article