Diversity Of Arthropod Communities In Transgenic Bt Cotton And Nontransgenic Cotton Agroecosystems

PHYSIOLOGICAL ECOLOGY

Diversity of Arthropod Communities in Transgenic Bt Cotton and Nontransgenic Cotton Agroecosystems XINGYUAN MEN, FENG GE,1 XIANGHUI LIU,

AND

ERDAL N. YARDIM2

Key laboratory of integrated management of pest insects and rodents, Institute of Zoology, Chinese Academy of Sciences, Beijing 100080, P. R. China

Environ. Entomol. 32(2): 270Ð275 (2003)

ABSTRACT In a 3-yr Þeld experiment, possible effects of Bt transgenic cotton plants expressing cry1A(c) gene from Bacillus thrungiensis Berliner variety kurstaki on diversities of arthropod communities and pest and natural enemy sub-communities were assessed in insecticide treated and untreated cotton Þelds, as measured by the Shannon-Weaver diversity index. The treatments included: 1) nontransgenic cotton with no insecticide treatment (nontransgenic), 2) nontransgenic cotton with insecticide treatments (nontransgenic⫹insecticide), 3) transgenic Bt-cotton with no insecticide treatment (Bt-cotton), and 4) transgenic Bt-cotton with insecticide treatments (Btcotton⫹insecticide). The results indicated that Bt-cotton increased the diversity of arthropod communities and pest sub-communities; however, it decreased the diversities of natural enemy subcommunities. Insecticide treatments increased diversities of communities and sub-communities of arthropods in both transgenic Bt-cotton and nontransgenic cotton agroecosystems, but the increase may be an artifact of increased evenness through mortality of insecticide-targeted species. KEY WORDS transgenic, Bt-cotton, biodiversity, species richness, Shannon-Weaver diversity index

RESISTANCE OF COTTON BOLLWORM (Helicoverpa armigera Hu¨ bner) to insecticides is a major concern because of the extensive reliance on environmentally disruptive and costly insecticide applications for profitable cotton production (Hearn and Fitt 1992, Fitt 1994). Transgenic cotton, expressing the ␦-endotoxin gene from the bacterium Bacillus thrungiensis appears to be a promising new technology to manage cotton bollworm (Forrester et al. 1993). It also offers the potential to reduce the total use of broad-spectrum chemical insecticides to control lepidopterous pests dramatically (Gary and Fitt 1994), and might have fewer side effects on nontarget organisms (Meeusen and Warren 1989, Huang et al. 1999). A number of studies on the inßuence of transgenic Bt-cotton on target and nontarget arthropods have been carried out to assess its ecological consequences in agroecosystems (Frutos et al. 1999, Smith 1997, Huang et al. 1999, Van Tol et al. 1998, Watston 1995, Fitt et al. 1994, Daly 1994). Transgenic cotton can have a number of direct and indirect effects on arthropod communities in agroecosystems. The direct impact is the mortality of bollworms feeding on Bt-cotton (Fitt 1994), which can also provide effective or partial control of some other lepidopteran pests (Harris et al. 1996, Watston 1995, Bacheler et al. 1998, Bacheler et E-mail: [email protected]. Department of Plant Protection, Faculty of Agriculture, Yuzuncu Yil University, Turkey 1 2

al. 1996). Transgenic Bt cotton can affect natural enemies indirectly (Huang 1999) through the removal of eggs, larvae, and pupae of lepidopteran insects that serve as food sources for parasitic and predatory arthropods. Considerable reduction in the number of insecticide applications is another important factor that can affect arthropod communities in Bt-cotton Þelds. In Australia, ⬇80% (8 Ð9 sprays) of the total sprays applied to nontransgenic cotton agroecosystems can be eliminated in transgenic Bt cotton agroecosystems (Fitt 1994). Similarly, in Mississippi, an average of 3.3 insecticide applications needed to control H. zea (Boddie) in nontransgenic cotton has been reduced to 0.3 applications in transgenic Bt cotton (U.S. EPA 1998). Such major reductions in pesticide applications can result in increases in the abundance of beneÞcial insects and some minor pests, which are otherwise suppressed under heavy insecticide application regimes to control H. armigera. China, where transgenic cotton was Þrst grown in 1998 to control resistant bollworms, has a multipleÑ cropping pattern within the cotton zone instead of a mono and large area cropping pattern seen in the U.S.A and Australia. This study aimed to assess the inßuences of nontransgenic and transgenic Bt cotton varieties on the diversities of arthropod communities and pest and natural enemy sub-communities in insecticide treated and untreated Þelds in China.

0046-225X/03/0270Ð0275$04.00/0 䉷 2003 Entomological Society of America

April 2003 Table 1.

FENG: DIVERSITY OF ARTHROPOD COMMUNITIES IN COTTON AGROECOSYSTEMS

271

Insecticide applications in the insecticide-treated nontransgenic and Bt-cotton plots during the three years of study

Pesticides Dicofol Phosalone Monocrotophos Cypermethrin Omethoate Carbaryl

Target pests Cotton mites Cotton bollworms Cotton mirids Cotton aphids Cotton aphids Cotton leafhoppers

Action threshold 70 adults/plant 100 eggs/100 plants 10 adults/100 plants 2000 individuals/100 plants 2000 individuals/100 plants 200 individuals/100 plants

Materials and Methods Field Experiments. The study was carried out between 1999 and 2001 in Fugou County, Henna Province, China (34N, 115E), where wheat and cotton are commonly intercropped. Wheat was planted in the experimental Þelds in October and harvested in June of the following year. Cotton was planted in the wheat Þeld in May and harvested in October. The experimental Þeld was a medium-textured silty loam soil and was fertilized twice with urea and calcium superphosphate at the rates of 30 kg/ha and 120 kg/ha on 20 June, and 50 kg/ha and 180 kg/ha on 25 July in all years. The experiment was set up based on a randomized complete block design involving four treatments each with three replicates. Each plot was ⬇0.4 ha. A 5-m gap was left between plots to avoid inßuence of treatments on arthropods in neighboring plots. All the vegetation between plots was removed when necessary to avoid effects of surrounding environment. The treatments included: 1) nontransgenic cotton with no insecticide treatment (nontransgenic), 2) nontransgenic cotton with insecticide treatments (nontransgenic⫹ insecticide), 3) transgenic Bt-cotton with no insecticide treatment (Bt-cotton), and iv) transgenic Btcotton with insecticide treatments (Btcotton⫹insecticide). We used the transgenic Btcotton variety “Deltapine NuCOTN 33B,” which contains the Bollgard gene expressing Cry1 A(c) (Monsanto, ST. Louis, MO), and the nontransgenic cotton variety “Chun Aizao.” Table 1 summarizes the insecticides used, target pests, action thresholds, application dates and dosage in the insecticide treated plots for the three years of study. Insecticides were applied based on an integrated pest management (IPM) program for cotton pests in China (Chen et al. 1990). Pest and Predator Sampling. Five 1-m2 sampling sites, each consisting of six cotton plants, were selected randomly in each plot. Numbers of sedentary arthropods (except cotton aphids) were counted visually on the plants at each site every Þve days from 15 May to 10 September in all years. Flying insects were sampled with Þve sweeps of a sweepnet near each sampling site. Arthropods collected by sweepnet were taken to the laboratory, identiÞed to species and counted. Five plants were randomly selected in each plot for sampling aphids. The numbers of aphids were counted on three leaves taken from three different positions within the plant canopy as described by

Rate of application (a.i.) (per ha) 113g 505g 150g 46g 180g 650g

Date Nontransgenic cotton 1999

2000

2001

22-Jul

20-Jul

22-Jul 18-Jul 9-Aug

20-Jul 8-Aug 8-Aug

Bt-cotton 1999

2000

2001

25-Jun 10-Jul 22-Jul 20-Aug 20-Aug

10-Jul 20-Aug

8-Aug 8-Aug

Hardee et al. (1993). Position one was the fourth fully expanded leaf from the terminal, position two was the Þrst main stem green leaf about one-third the distance from the terminal, and position three was the Þrst main stem green leaf above the Þrst fruiting branch. Data Analysis. The Shannon-Weaver diversity index H⬘ (Shannon and Weaver 1949), H⬘ ⫽ ⫺¥ pi logep i where pi is the proportion of the ith species in the total sample, was calculated to measure the arthropod community diversity (all species of insects and mites), as well as pest and natural enemy sub-communities in the treatment plots. Data on abundances, species richness and diversities of communities and sub-communities were analyzed using Two-way analysis of variance (ANOVA). Means were separated using the DuncanÕs multiple range test. Results Abundance. A total of 36 species were observed in the experimental plots during the three years of study (Table 2). Abundances of total arthropods differed signiÞcantly among the treatment plots in 1999 (F ⫽ 18.32; df ⫽ 3; P ⬍ 0.001) and 2001 (F ⫽ 4.23; df ⫽ 3; P ⫽ 0.022). The nontransgenic cotton plots had the highest abundance of arthropods, followed by the Bt-cotton, the nontransgenic⫹insecticide, and the Btcotton⫹insecticide plots (Table 3). Pest abundance differed signiÞcantly among treatments in 1999 (F ⫽ 17.69; df ⫽ 3; P ⬍ 0.01) and 2001 (F ⫽ 7.18; df ⫽ 3; P ⫽ 0.003), showing a trend similar to that of total arthropods. Treatments had a signiÞcant effect on natural enemy abundance in 1999 (F ⫽ 7.45; df ⫽ 3; P ⫽ 0.002) and 2001 (F ⫽ 10.22; df ⫽ 3; P ⫽ 0.001). Insecticide applications signiÞcantly (P ⬍ 0.05) reduced natural enemy abundance in both the nontransgenic⫹ insecticide and the Bt-cotton⫹insecticide plots in 1999 and 2001. Bt-cotton alone did not signiÞcantly affect the abundance of natural enemies. Species Richness. Species richness of arthropod communities differed signiÞcantly between treatments in 1999 (F ⫽ 27.746; df ⫽ 3; P ⫽ 0.0001) and 2001 (F ⫽ 6.41; df ⫽ 3; P ⫽ 0.016) (Table 4). The species richness of arthropod communities was signiÞcantly lower (P ⬍ 0.05) in the Bt-cotton plots than in the nontransgenic cotton plots in both 1999 and 2001. Insecticide applications led to signiÞcant (P ⬍ 0.05)

272 Table 2.

ENVIRONMENTAL ENTOMOLOGY

Vol. 32, no. 2

Number of individuals of arthropod species sampled from treatment plots during the three years of study Species

Family

Pestiferous Species Aphis gossypii Glover Empoasca biguttula Shiraki Empoasca flavescens Fabricius Trialeurodes vaporariorum Westwood Adelphocoris faciaticollis Reuter Lygus lucorum Meyer-Du¨ r Adelphocoris suturalis Jakovlev Halyomorpha picus Fabricius Helicoverpa armigera Hu¨ bner Anomis falva Fabricius Laphygma exigua Hubner Argyrogramma agnata Staudinger Spilarcitia subcarnea Walker Sylepta derogata Fabricius Tetranychus cinnabarinus and T. truncatus Thrips tabaci Lindeman Phytoscaphus gossypii Chao Pleonomus canaliculatus Faldemann Maladera orientalis Motschulsky Natural Enemies Propylaea japonica Thunberg Coccinella septempunctata Linnaeus Harmonia axyridis Pallas Scymnus hoffmanni Weise Chrysopa sinica Tjeder Chrysopa septempunctata Wesmael Misumjenops tricuspidatus Fabricius Eringonidium graminicola Sundevall Theridion octomaculatum Boes Tetragnatha maxillosa Thorell Geocoris pallidipennis Costa Orius minuius Linnaeus Aphidius gifuensis Ashmead Charops bicolor Szepligeti Epistrph balteata De Geer Melanostoma scalare Fabricius

1999

Order

a

b

2000 c

d

a

b

2001

c

d

a

b

c

d

Aphididae Cicadellidae Cicadellidae Aleyrodoidea Miridae Miridae Miridae Pentatomidae Noctuidae Noctuidae Noctuidae Noctuidae Arctiidae Pyralidae Tetranychidae

Homoptera Homoptera Homoptera Homoptera Hemiptera Hemiptera Hemiptera Hemiptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Acarina

1380 828 1002 298 340 248 235 247 811 348 512 207 8 2 138 16 1 4 12 24 37 3 52 20 15 0 8 2 8 28 19 24 15 13 11 7 23 32 12 10 0 0 0 0 84 74 55 37 23 4 12 13 5 11 9 7 8 5 9 5 1 0 0 1 0 0 0 0 0 1 3 0 5 1 0 2 0 2 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 3 1 1 0 2 2 2 0 2 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 17 9 7 9 1 0 0 1 5 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 2 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2520 980 1747 449 11 41 24 111 7 2 18 9

Thripidae Curculionidae Elateridae Melolonthidae

Thysanoptera Coleoptera Coleoptera Coleoptera

42 21 4 0

10 29 3 2

20 14 5 0

3 7 2 0

5 0 0 0

0 3 0 0

0 5 1 0

0 0 0 0

15 21 0 0

5 21 0 0

10 25 1 0

7 11 0 1

Coccinellidae Coccinellidae Coccinellidae Coccinellidae Chrysopidae Chrysopidae Thomisdae Erigoninae Theridiidae Tetragnathidae Lygaeoidae Anthocoridae Aphidiidae Ichneumonidae Syrphidae Syrphidae

Coleoptera Coleoptera Coleoptera Coleoptera Neuroptera Neuroptera Araneida Araneida Araneida Araneida Hemipterodea Hemipterodea Hymenoptera Hymenoptera Diptera Diptera

13 2 1 13 0 1 14 47 18 2 0 2 2 0 1 0

19 4 6 12 1 0 11 15 3 0 0 0 3 0 0 0

17 1 2 21 2 0 26 62 21 9 3 0 0 1 0 2

10 0 0 9 0 0 3 17 13 0 0 5 0 0 0 0

3 1 0 0 0 2 1 3 0 0 0 7 4 0 0 0

16 0 0 3 0 0 4 8 2 0 0 0 3 0 1 0

21 3 0 1 0 0 9 11 0 0 0 0 13 0 3 0

15 1 0 0 3 0 15 7 2 0 0 0 17 0 0 0

9 0 3 0 1 0 13 25 0 0 2 7 10 0 0 2

7 0 0 0 0 0 0 2 4 0 3 3 7 1 0 0

5 2 6 0 0 0 16 30 7 0 0 2 5 0 0 0

4 0 3 0 3 0 3 5 0 0 2 2 3 0 0 0

a: nontransgenic; b: nontransgenic ⫹ insecticide; c: Bt-cotton, d: Bt-cotton ⫹ insecticide.

reductions in species richness of arthropod communities in the nontransenic⫹insecticide and the Btcotton⫹insecticide plots compared with the nontransgenic cotton and Bt-cotton plots, respectively. There were signiÞcant treatment effects on the species richness of pest subcommunities in the experimental plots Table 3.

in 1999 (F ⫽ 12.852; df ⫽ 3; P ⫽ 0.002) and 2001 (F ⫽ 9.75; df ⫽ 3; P ⫽ 0.005) (Table 4). The species richness of pest subcommunities was signiÞcantly lower (P ⬍ 0.05) in the Bt-cotton plots than nontransgenic cotton plots in 1999. Insecticide applications signiÞcantly (P ⬍ 0.05) reduced species richness of pest sub-com-

Average numbers of arthropods (mean ⴞ SEM) sampled in treatment plots during the three years of study Nontransgenic

Nontransgenic ⫹ insecticide

Bt-cotton

Bt cotton ⫹ insecticide

Arthropods Pests Natural enemies

231.07 ⫾ 26.80 a 225.59 ⫾ 26.58 a 6.71 ⫾ 1.30 a

1999 111.91 ⫾ 8.69 c 109.23 ⫾ 8.72 c 4.06 ⫾ 0.43 b

171.34 ⫾ 21.12 b 163.89 ⫾ 21.09 b 9.45 ⫾ 3.6 a

50.23 ⫾ 8.67 d 47.71 ⫾ 8.82 d 3.11 ⫾ 0.20 b

Arthropods Pests Natural enemies

21.98 ⫾ 2.54 a 20.22 ⫾ 3.65 a 1.31 ⫾ 0.23 a

2000 22.02 ⫾ 3.72 a 18.33 ⫾ 2.03 a 2.07 ⫾ 0.45 a

20.86 ⫾ 1.52 a 17.01 ⫾ 1.73 a 3.38 ⫾ 0.94 a

23.48 ⫾ 1.66 a 22.76 ⫾ 1.95 a 3.09 ⫾ 0.50 a

Arthropods Pests Natural enemies

59.24 ⫾ 5.54 a 55.71 ⫾ 4.95 a 3.78 ⫾ 0.52 a

2001 29.19 ⫾ 2.27 bc 24.88 ⫾ 1.94 bc 1.24 ⫾ 0.30 b

40.81 ⫾ 3.41 b 36.59 ⫾ 1.94 b 3.87 ⫾ 1.46 a

19.26 ⫾ 1.86 c 16.75 ⫾ 2.00 c 1.36 ⫾ 0.29 b

Within a row, values indicated by different letters are signiÞcantly different.

April 2003 Table 4.

FENG: DIVERSITY OF ARTHROPOD COMMUNITIES IN COTTON AGROECOSYSTEMS

273

Average numbers of arthropod species (mean ⴞ SEM) sampled in treatment plots during the three years of study Nontransgenic

Nontransgenic ⫹ insecticide

Bt-cotton

Bt cotton ⫹ insecticide

Arthropod Pest Natural enemy

25.67 ⫾ 1.20 a 14.00 ⫾ 0.58 a 11.68 ⫾ 0.68 a

1999 20.67 ⫾ 0.88 b 12.33 ⫾ 0.33 b 8.33 ⫾ 0.33 b

22.33 ⫾ 0.33 b 11.67 ⫾ 0.67 b 10.67 ⫾ 0.33 a

16.00 ⫾ 0.00 c 9.67 ⫾ 0.67 c 5.33 ⫾ 0.88 c

Arthropod Pest Natural enemy

13.00 ⫾ 0.58 a 7.33 ⫾ 0.33 a 5.67 ⫾ 0.33 a

2000 13.00 ⫾ 0.58 a 7.00 ⫾ 0.58 a 6.00 ⫾ 0.00 a

11.67 ⫾ 1.33 a 6.33 ⫾ 0.67 a 5.33 ⫾ 0.67 a

11.67 ⫾ 0.67 a 5.67 ⫾ 0.67 a 5.67 ⫾ 0.33 a

Arthropod Pest Natural enemy

19.33 ⫾ 0.67 a 11.67 ⫾ 0.33 a 8.00 ⫾ 0.57 a

2001 14.67 ⫾ 0.67 bc 8.00 ⫾ 0.58 bc 6.33 ⫾ 0.33 b

16.00 ⫾ 1.53 b 10.00 ⫾ 1.00 ab 6.67 ⫾ 0.67 b

12.33 ⫾ 1.45 c 7.00 ⫾ 0.58 c 5.33 ⫾ 0.88 c

Within a row, values indicated by different letters are signiÞcantly different.

munities in both 1999 and 2001. SigniÞcant differences occurred in the species richness of natural enemy subcommunities in 1999 (F ⫽ 21.949; df ⫽ 3; P ⬍ 0.001) and 2001 (F ⫽ 12.852; df ⫽ 3; P ⫽ 0.002) (Table 4). The species richness of natural enemies was lower in the nontansgenic⫹insecticide and Bt-cotton⫹insecticide plots than in the nontransgenic cotton and Bt-cotton plots in both years. The Bt-cotton plots had signiÞcantly lower (P ⬍ 0.05) species richness of natural enemy subcommunities than the nontransgenic cotton plots in 2001. Diversity. Shannon-Weaver diversity indexes of arthropod communities differed signiÞcantly (P ⬍ 0.01) among the treatments in all years (F ⫽ 5.740; df ⫽ 3; P ⫽ 0.007 in 1999; F ⫽ 5.567; df ⫽ 3; P ⫽ 0.008 in 2000; F ⫽ 18.621; df ⫽ 3; P ⫽ 0.008 in 2001) (Table 5). The Bt-cotton⫹insecticide plots had signiÞcantly higher (P ⬍ 0.05) diversity indexes of arthropod communities than the other treatment plots in all years. No significant differences occurred among the nontransgenic cotton, the Bt-cotton and the nontransgenic⫹ insecticide treatments with respect to species richness of arthropod communities. Diversity indexes of pest sub-communities differed signiÞcantly among treatment plots in 1999 (F ⫽ 3.173; df ⫽ 3; P ⫽ 0.043) and 2001 (F ⫽ 42.73; df ⫽ 3; P ⫽

0.002) (Table 5). The index in the Bt-cotton plots was signiÞcantly higher (P ⬍ 0.05) than in the nontransgenic cotton plots in 2001. Like arthropod communities, pest sub-communities appeared to be more diverse in the insecticide treated plots. The nontransgenic⫹insecticide and the Bt-cotton⫹ insecticide plots had signiÞcantly higher (P ⬍ 0.05) diversity indexes of pest sub-communities than the nontransgenic and the Bt-cotton plots, respectively in 2001. There were signiÞcant treatment effects on the diversities of natural enemy sub-communities in the treatment plots in 1999 (F ⫽ 7.502; df ⫽ 3; P ⫽ 0.002) and 2001 (F ⫽ 58.73; df ⫽ 3; P ⫽ 0.001). The indexes did not differ signiÞcantly with respect to variety in 1999. However, unlike the pest sub-communities, the diversity index of the natural enemy sub-communities was signiÞcantly higher (P ⬍ 0.05) in the nontransgenic plots than that in the Bt-cotton plots in 2001. Insecticide applications led to increases in the diversity of natural enemy subcommunities. The diversity index in the nontransgenic⫹insecticide plots was signiÞcantly higher (P ⬍ 0.05) than that in the nontransgenic plots in 1999. Similar results were observed in the Bt-cotton⫹insecticide plots where the diversity index of natural enemy subcommunities was signiÞ-

Table 5. Diversity indexes (mean ⴞ SEM) of arthropod communities and pest and natural enemy sub-communities in treatment plots during the three years of study Nontransgenic

Nontransgenic ⫹ insecticide

Bt-cotton

Bt cotton ⫹insecticide

1.04 ⫾ 0.08 a 0.89 ⫾ 0.06 ab 0.45 ⫾ 0.04 a

1.36 ⫾ 0.14 b 1.01 ⫾ 0.08 b 0.55 ⫾ 0.04 a

Arthropod community Pest sub-community Natural enemy sub-community

0.89 ⫾ 0.03 a 0.78 ⫾ 0.02 a 0.44 ⫾ 0.04 a

1999 1.04 ⫾ 0.02 a 0.93 ⫾ 0.02 ab 0.77 ⫾ 0.02 b

Arthropod community Pest sub-community Natural enemy sub-community

1.14 ⫾ 0.13 a 0.81 ⫾ 0.12 a 1.00 ⫾ 0.05 a

2000 1.29 ⫾ 0.09 a 0.94 ⫾ 0.11 a 0.98 ⫾ 0.15 a

1.26 ⫾ 0.08 a 0.80 ⫾ 0.11 a 0.99 ⫾ 0.05 a

1.36 ⫾ 0.05 b 0.95 ⫾ 0.09 a 1.13 ⫾ 0.06 a

Arthropod community Pest sub-community Natural enemy sub-community

0.93 ⫾ 0.08 a 0.86 ⫾ 0.06 a 1.12 ⫾ 0.03 a

2001 1.45 ⫾ 0.03 a 1.20 ⫾ 0.05 b 1.15 ⫾ 0.06 a

1.35 ⫾ 0.03 a 1.16 ⫾ 0.03 b 0.83 ⫾ 0.04 b

1.51 ⫾ 0.09 b 1.35 ⫾ 0.02 c 1.44 ⫾ 0.05 c

Within a row, values indicated by different letters are signiÞcantly different.

274

ENVIRONMENTAL ENTOMOLOGY

cantly higher (P ⬍ 0.05) than in the Bt-cotton plots in 2001. Discussion The results of the study revealed signiÞcant differences in abundance and species richness between nontransgenic and Bt-cotton and insecticide treated and untreated plots. Insecticide applications reduced the abundance of total arthropods, pests and natural enemies in both the nontransgenic⫹insecticide and the Bt-cotton⫹insecticide plots. Total arthropods and pests were more abundant in the nontransgenic plots than in the Bt-cotton plots. Bt cotton can effectively reduce numbers of lepidopteran pests such as Heliothis virescens (F.), H. zea (Boddie), and Pectinophora gossypiella (Saunders), and can partially suppress several nontarget lepidopteran pests, such as Spodoptera exigua (Hu¨ bner), Trichoplusia ni (Hu¨ bner), Buccalatrix thruneriella Busck, Ostrinia nubilalis (Hu¨ bner), Spodoptera frugiperda (J. E. Smith), Spodoptera eridania (Stoll), Pseudoplusia includens (Walker) (Bacheler and Mott 1996, Bacheler et al. 1998, Harris et al., 1996, Watston 1995). Decreases in pest abundance in Bt-cotton are expected to arise primarily from decreases in lepidopterans. However, reductions of some other nontarget insects, such as aphids and thrips, apparently contributed substantially to the signiÞcant decreases of pest subcommunities in the Bt-cotton plots. Insecticide applications and Bt-cotton decreased species richness of communities and subcommunities. SigniÞcant decreases in species richness of natural enemy subcommunities in the Bt-cotton plots may have resulted from migration or mortality of some species, perhaps in response to decreased prey populations. Most of the signiÞcant differences in abundance, species richness and diversity indexes occurred in 1999 and 2001. In 2000, heavy and long lasting rainfalls in July and August may have suppressed arthropod populations and masked any possible signiÞcant treatment effects. The diversity indexes of arthropod communities and pest subcommunities were generally signiÞcantly higher in the insecticide treated nontransgenic and Bt-cotton plots. This might be explained in that the Shannon-Weaver index of diversity operates as a function of both species richness and evenness (Shannon and Weaver 1949). Hence, reductions in abundance of some pests caused by insecticide applications in the insecticide treated plots and by Bt toxin in the Btcotton plots could contribute to evenness among the species, thus resulting in higher diversity indexes of arthropod communities and pest subcommunities. Similarly, signiÞcantly higher diversity indexes of natural enemies in the insecticide treated plots than in the untreated plots may be attributed to the effects of the insecticides on evenness by reducing natural enemy abundances. However, signiÞcantly lower diversity indexes in the Bt-cotton plots in 2001 can be attributed to decreases in the species richness of natural enemy

Vol. 32, no. 2

subcommunities because Bt-toxin did not affect natural enemy abundance. In conclusion, our Þndings suggest that the use of transgenic Bt-cotton and application of insecticides decreased species richness of communities and subcommunities. As a result, species composition was altered by eliminating some species and increasing diversity of pest subcommunities, as measured by the Shannon-Weaver diversity index. However, the diversity of natural enemies was decreased in Bt-cotton. Acknowledgments This project was supported by the Program of Chinese Academy of Sciences (Grant No. KSCX2-1-02 and KSCX2SW-103) the State Key Basic Research “973” Program (Grant No. G2000016209) and the Chinese National Science Fund (Grant No. 39970137).

References Cited Bacheler, J. S., and D. W. Mott. 1996. Potential utility and susceptibility of transgenic Bt cotton against bollworms European corn borer and stink bugs in NC. Proceedings Beltwide Cotton Conferences, pp. 927Ð931. Bacheler, J. S., D. W. Mott, and D. E. Morrison. 1998. Large scale evaluation of bollgard resistance to multiple pests in North Carolina under grower conditions. Proceedings Beltwide Cotton Conferences, pp. 961Ð984. Chen, Q. Y., D. M. Li, and C. Y. Cao. 1990. Recent advances in research on integrated management of cotton diseases and insect pests in China. China Agricultural ScientiÞc Press, Beijing. Daly, J. C. 1994. Ecology and resistance management for Bacillus thuringiensis transgenic plants. Biocontrol Sci. Technol. 4: 563Ð571. Fitt, G. P. 1989. The ecology of Heliothis species in relation to agro-ecosysem. Annual Review of Entomology. 34: 17Ð52. Fitt, G. P. 1994. Cotton pest management: part 3. An Australian perspective. Annu. Rev. Entomol. 39: 543Ð562. Forrester, N. W., M. Cahill, L. J. Bird, and J. K. Layland. 1993. Management of pyrethroid and endofulfan resistance in Helicoverpa armigera (Huber) in Australia. Bull. Entomol. Res. 1: 132 pp. Frutos, R., C. Rang, and M. Royer. 1999. Managing insect resistance to plants producing Bacillus thuringiensis toxins. Crit. Rev. Biotechnol. 19(3): 227Ð276. Gary, P., and G. P. Fitt. 1994. Field evaluation and potential ecological impact of transgenic cottons (Gossypium hirsutum) in Australia. Biocontrol Sci. Technol. 4: 535Ð548. Hardee, D. D., M. T. Smith, and A. A. Weathersbee. 1993. Sampling of the cotton aphid, Aphis gossypii Glover, in cotton, pp. 738 Ð740, In Proceedings, Beltwide Cotton Production and Research Conferences. National Cotton Council of America, Memphis, TN. Harris, F. A., J. R. Furr, and D. S. Calhoun. 1996. Cotton insect management in transgenic Bt cotton in the Mississippi Delta, 1992Ð1995. Proceedings Beltwide Cotton Conferences, pp. 854 Ð 858. Hearn, A. B., and G. P. Fitt. 1992. Cotton cropping system, pp. 85Ð142. In C. Pearson [ed.], Field Crop Ecosystem of the World. Elsevier Press, Amsterdam. Huang, F. N., R. A. Higgins, and L. L. Buschman. 1999. Transgenic Bt-plants: successes, challenges and strategies. Pestology 2: 2Ð29.

April 2003

FENG: DIVERSITY OF ARTHROPOD COMMUNITIES IN COTTON AGROECOSYSTEMS

James, C. 1997. Global review of the Þeld testing and commercialization of transgenic plants. ISAAA Briefs No. 5. ISAAA Ithaca, NY. Meeusen, R. L., and G. Warren. 1989. Insect control with genetically engineered crops. Annu. Rev. Entomol. 34: 373Ð382. Shannon, C., and W. Weaver. 1949. Thee mathematical theory of communication. University of Illinois Press, Urbana, IL. 117p. Smith, R. H. 1997. An extension entomologistÕs 1996 observation of Bollgard (Bt) technology. Proceedings of the Beltwide Cotton Conference, 1977. National Cotton council, Memphis, TN, pp. 856 Ð 861.

275

[EPA] US Environmental Protection Agency. 1998. The Environmental Protection Agency White Paper on Bt Plant-pesticide Resistance Management, US EPA, p. 84. Van Tol, N. B., and G. L. Lentz. 1998. Inßuence of Bt cotton on beneÞcial arthropod populations. Proceeding Beltwide Cotton Conferences, pp. 1052Ð1054. Watston, T. F. 1995. Impact of transgenic cotton on pink bollworm and other lepidopteran insects. Proceedings Beltwide Cotton Conferences, pp. 759 Ð760. Received for publication 13 March 2002; accepted 12 November 2002.