Diversity , Composition and Population Dynamics of Arthropods in the Genetically Modified Soybeans Roundup Ready ® RR 1 ( GT 40-3-2 ) and Intacta RR 2 PRO ® ( MON 87701 x MON 89788 )

Knowledge of insect diversity is essential for ecological studies and pest management. The aim of this study was to study the occurrence, abundance of target and non-target pests in genetically modified insect resistant (Bt) and glyphosate-tolerant soybeans (RR1 and RR2), with and without the application of insecticides. Experiments were carried out in the agricultural year of 2011/2012 in four municipalities in the State of Mato Grosso do Sul. The treatments were: 1 Roundup Ready RR1 soybeans without insecticide application; 2 Roundup Ready RR1 soybeans with application whenever the control level was reached; 3 Intacta RR2 PRO soybeans without insecticide application; and 4 Intacta RR2 PRO soybeans with application whenever the control level was reached. The evaluations were initiated soon after emergence of the plants at weekly intervals. In order to obtain representative gradients of species composition, we used the method of ordering by non-metric multidimensional scaling (NMDS) and the Bray-Curtis dissimilarity index. Insects of the order Lepidoptera presented 7547 specimens, composing more than 70% of the insect community. The orders Coleoptera and Hemiptera also stood out, consisting of 2066 and 331 insects, respectively. Most of the samples were recorded in stages V8 to R2 for defoliating caterpillars and between R5.2 and R6 for the phytophagous stink bug complex. The Bt technology significantly reduced the target insect pests and favored populations of natural enemies. The treatments with insecticide application resulted in reduction of arthropods collected and changes in population outbreaks when compared to areas without spraying.


Introduction
The soybean [Glycine max (L.)Merrill (Fabaceae: Phaseoleae)] stands out as the main agricultural commodity Brazil, where in the crop year 2011/2012 about 25 million hectares were planted, producing 66,383,000 tons.In 2012/2013, the acreage increased by about 8.8% to 27.241 million acres and the expected harvest is about 82.627 million tons, a 24.5% increase in national production of this crop (Conab, 2012).
Soybean productivity is threatened by abiotic factors including climatic and biotic factors such as attack of insect pests, which can be controlled by natural enemies such as predators, parasitoids and pathogens.However, their populations often reach levels of economic damage to the crop.High population levels of these pests can be a result of favorable environmental conditions and/or inadequate management, including overuse of pesticides (Embrapa, 2011).
There has been a long search for specific technologies with low persistence in the environment, including Integrated Pest Management "IPM" based on plant resistance to insects.With the advent of genetic transformation based on recombinant DNA technology, it has become possible to insert foreign genes into the genomes of plants, thus conferring insect resistance (Bennet, 1994).Genes from bacteria, such as Bacillus thuringiensis (Bt) and Bacillus sphaericus have been the main organisms used to confer plant resistance to insects on the commercial scale (Sharma et al., 2000).
Recent advances in agricultural biotechnology have resulted in transgenic plants which are efficient alternatives and have less environmental impact for controlling lepidopteran pests in areas where these are of agronomic concern (Williams et al., 1998).
Knowledge of diversity and arthropods associated with agriculture is essential for ecological and IPM studies, considering that in agroecosystems the arthropod communities change according to the management exercised (Milane et al., 2012).Some arthropod species can cause damage to plants, can sometimes lead to death, however some beneficial species can control these pest species through predation or parasitism (Garcia et al., 2004).
Population dynamic studies provide useful information for developing models involving pest management (Gilbert et al., 1976), considering the possibility of obtaining population distribution data over a given period of time (Odum, 1988).Therefore, such studies can be successfully employed in IPM programs (Silveira Neto et al., 1976).
The objective of this work was to evaluate the abundance, diversity, population dynamics and influence of phenological growth stages on the composition of insect pests and beneficial arthropods in genetically modified soybeans resistance to insects Intacta RR2 PRO ® (Bt) and glyphosate-tolerant RR1 (non-Bt).

Materials and Methods
Local installation of the experiment -The study was conducted during the agricultural year of 2011/2012 under field conditions using four treatments and four replications in randomized blocks.Treatments: 1 -Roundup Ready ® RR1 soybeans without application of insecticides, 2 -Roundup Ready ® RR1 soybeans with application of insecticides, both treatments used to cultivar Brasmax Potência RR1 (event GT 40-3-2), 3 -Intacta RR2 PRO ® soybeans without insecticide application and 4 -Intacta RR2 PRO ® soybean with insecticide application; for treatments 3 and 4 the experimental strain used was Monsoy L6910 (events MON 87701 and MON 89788).The cultivars tested showed similar cycle, growth habit and phenology.
Sowing was performed in areas free of weeds, using a John Deere model 1109 vacuum mechanized planter, with 9 rows and 45cm row spacing; seed density was between 13 and 14 seeds per linear meter.
For insect management in treatments RR1 and Intacta RR2 PRO ® with insecticide application, the control levels adopted in the study were maintained: defoliating caterpillars in the vegetative stage ≥ 30% defoliation or 20 large larvae (≥ 1.5 cm), and in the reproductive stage ≤ 15% defoliation or 20 large larvae (≥ 1.5 cm) per linear meter of soybeans.Sucking insects in the reproductive stage ≥ 1 and stink bug ≥ 0.5 cm per meter (level adopted for seed production) (Embrapa, 2011).
Using control levels recommended by Embrapa (2011) were performed 3.75 insecticide sprayings, on average, in the four sites studied considering the treatment RR1 (with insecticide spraying) for management of defoliating caterpillars, had no spraying in the treatment Intacta RR2 PRO ® (with spraying).For management of phytophagous stink bugs were sprayed on average 2.5 and 2.75 times, in the four locations, in soybean Intacta RR2 PRO ® and RR1 (both treatments with insecticide), respectively.
Sampling -Evaluations were performed immediately after plant emergence and periodically using intervals between 7 to 10 days, identifying the phenological stages proposed by Fehr and Caviness (1977) and Farias et al. (2007), from plants emergence to physiological maturity of the crop.This proposed design for classifying the stages of soybean development considers the vegetative and reproductive stages.Vegetative stages are described by the letter V and reproductive stages by the letter R. Excluding VE (emergence) and VC (cotyledon) stage, stages V and R are followed by numerical indices that indicate the actual stage of plant development (e.g., V1, V2, V3, V5, V6… Vn), where the n is the number of nodes with fully developed leaves above the cotyledons, and stage R, from 1 to 8, which describe the flowering period maturation, comprising the plant flowering (R1 and R2), the pod development (R3 and R4), developing seeds (R5 and R6), and physiological maturity (R7 and R8).
The total area used for each plot measured 18 meters wide by 25 meters long (450 m 2 ), where 10 collection points were used per repetition in the respective culture phenological stages.
For sampling of the complex of insect pests and beneficial arthropods a ground cloth was used, consisting of two wood sticks connected by a white cloth, with a length of 1m and a width of 1.4 m (large enough to cover soybean line adjacent sampled).For sampling, one end of cloth was placed between the soybean rows, adjusted to the base of one row and other extended over the plants of the adjacent row.The plants in one row (area=0.45m 2 ) were shaken vigorously so as to cause the insect pests to fall on the cloth (Stürmer et al., 2012).
Statistical analysis -The faunistic analysis was performed according to Silveira Neto et al. (1976), calculating the indices of frequency, constancy, abundance and dominance.Frequency was defined as the percentage of individuals of a species in relation to the total.The constant is defined as the percentage of samples in which a determined species was present.Having obtained this constant, the species were rated as: constant (w) when present in over 50% of weekly samples; accessory (y) when present in 25 to 50% of the samples; and accidental (z) when present in less than 25% of the samples.
Abundance, which is the number of individuals of a particular species found in a given area (unit area or volume), may vary across space and time (Southwood, 1995).To estimate the abundance classes, limits established by the confidence intervals (CI) of 5% and 1% probability were used to determine the following classes: rare (r), the number of individuals of the species below the lower threshold of the CI at 1% probability; dispersed (d), number of individuals among the lower limits of the CI at 1% and 5% probability; common (c), number of individuals within the CI at 5%; abundant (a), number of individuals between the upper limits of the CI at 1% and 5% probability, and very abundant (va), number of individuals greater than the upper limit of the CI at 1% probability.
To define dominance, which is the action exerted by the dominant organisms in a community, we used the method of Laroca and Mielke (1975).The species is considered dominant when presenting relative frequency greater than 1/S x 100, where S is the total number of species found in the sampling period.
In order to measure the ecological dominance, that is the degree to which a taxon is more numerous than its competitors in an ecological community, we used the method of Laroca and Mielke (1975).In this method, species is considered dominant when presenting relative frequency greater than 1/S x 100, where S is the total number of species found in the sampling period.
To measure the diversity of the insect community, the following indices were calculated: Margalef richness index (α), that is the ratio between the number of species and number of individuals of a community; Shannon-Wiener Insects of the order Lepidoptera were the most abundant and presented 7,547 specimens, or slightly more than 70% of the total collected; most of these are considered insect pests and are of great economic importance (Table 1).Next are the orders Hemiptera, Coleoptera, Neuroptera and Dermaptera, which respectively accounted for 2066, 331, 88 and 6 individuals.These results were expected, since these orders include a large number of pests and natural enemies of soybean (Chiaradia et al., 2011;Hoffmann-Campo et al., 2000).
Arthropods of the order Araneae totaled 400 specimens, thus forming an important group of predators, since they feed on insects and generally show high species richness in natural and agricultural areas, sometimes being very abundant (Triplehorn & Johnson, 2005).
In RR1 soybeans with insecticide application 2870 specimens were collected, divided in 19 orders, where two species were classified as super-dominant, 11 as dominant, three super abundant, four plentiful, six rare, three super frequent, four very frequent and eight species common.In the RR1 treatment without insecticide application 4833 specimens were collected, also distributed in 19 species, of which three were super dominant, 12 dominant, two super abundant, two very abundant, four rare, three super frequent, two very frequent and eight species were classified as constant (Table 2).The orders Lepidoptera and Hemiptera stand out by presenting greater frequency of insects sampled and are also the major pest species of this crop.Two of the three super-dominant species are contained in RR1 soybeans, A. gemmatalis and C. includens, key pests of the culture and targets of Bt technology.
In the Intacta RR2 PRO ® soybean with insecticide application, 934 specimens were collected, distributed in 22 species, three species being super dominant, 11 dominant, three super abundant, six very abundant, eight rare, three super frequent, six very frequent and seven species constants, respectively.In the treatment utilizing Intacta RR2 PRO ® without insecticide application 1401 specimens were collected, belonging to 23 species, one super dominant, 15 dominant, one super abundant, three very abundant, seven rare, one super frequent, three very frequent and nine constants, respectively (Table 3).
The number of taxa collected and analyzed was slightly higher for the Intacta RR2 PRO ® soybean, presenting 22 and 23 species compared with the RR1 soybean with only 19 species.The lower number of specimens collected, mainly from target pests, may have promoted the development and establishment of rare species populations such as M. latipes, S. subsignatus, M. armatus and Chinavia spp., non-target insects of Bt technology.The calculated indices of richness, diversity and evenness (Table 4) show that despite the number of specimens and families sampled, in the area studied there occurred many rare and few abundant species (Tables 2 and 3).Thus, we can infer that the community studied shows similarity in structure compared with other agroecosystems and natural ecosystems.For example, Nunes et al. (2009) found values of up to 1.09 for the Shannon index, 10.75 for species richness and 0.32 for equitability in studies of soil fauna with different management.
The Margalef richness (α), Shannon-Wiener diversity (H") and equitability (E) indices presented similar values for the soybeans RR1 and Intacta RR2 PRO ® (with and without insecticide application) (Table 4).These results represent equality in species richness within technologies RR1 and Intacta RR2 PRO ® , indicating no difference in species composition when applying insecticides in the fields, where only the technologies differ for with respect to this component.
Values of the Equitability index (Table 4) can be considered low for the four treatments.This demonstrates that some species show a much larger number of individuals than others.This is further confirmed by the difference between the number of abundant and rare species, indicating that the distribution of species in the community is not uniform, since the more similar the representativeness of each species in the community, the higher this index (Peet, 1974).
Population fluctuation of the insect pest complex over time, considering the culture stages and the occurrence of the species, indicated that the total population of insect pests, regardless of treatment, increased from stage V6 to the full vegetative stage.The insect pest population grew significantly in treatments with RR1 soybeans (non-Bt) when compared with treatments with Intacta RR2 PRO ® (Bt).However, only in cultivation of the RR1 soybean did population growth persist throughout the development of the culture (Figure 1).
During the soybean vegetative and reproductive stages, the lepidopterous targets of the Bt technology were significantly controlled and remained at levels below those of the RR1 treatments, even with application of insecticides.This is due to the toxic action of the Cry 1AC protein expressed in the genetically modified Bt cultivar, causing very high mortality of neonate caterpillars of the species A. germmatalis and C. includens (Macrae et al., 2005;Mcpherson & Macrae, 2009).
Greater fluctuations in arthropod populations were observed in the treatment using RR1 soybeans with insecticide application, caused by the reduction of insects due to use of chemical control among the reproductive stages R2 and R6 (Figure 1).
For non-target insect pests of Bt technology, especially the stink bug complex, in treatments without insecticide application an increase in population was observed in the reproductive stages R6 to R7.3, corroborating with.Salvadori et al., (2007) and Correa-Ferreira and Panizzi (1999).This is due to overlapping of generations which, when infesting the culture in the initial reproductive states R1 to R3, culminates in high populations in the later stages R6 and R7.3.
The population of insects in RR1 soybeans with insecticide application showed two outbreaks, both in the reproductive phase of the crop, the first between stages R1 and R3 and the second between R5.4 and R7.1.Decline in the insect population between outbreaks in R3 to R5.4 was observed occurred due to the application of insecticide and control of defoliating caterpillars (Figure 1A).For the RR1 soybean without insecticide application high population levels were reached and maintained during the reproductive stage of the culture (stages R1 and R6) (Figure 1B).
Intacta RR2 PRO ® (Bt) with and without spraying presented population fluctuation of pests and similar fluctuation during vegetative and reproductive development, with total populations much lower than those observed in the two treatments using the RR1 technology (Figure 1C and D).
For Intacta RR2 PRO ® soybeans without insecticide application an increase in the number of insect pests was observed in the final stages of the culture R6 and R7.1, probably due to non-use of insecticides for managing of the stinkbug complex, pests not targeted by the Bt technology (Figure 1D).
Populations of beneficial arthropods (natural enemies), represented by the orders Hemiptera, Dermaptera, Coleoptera, Neuroptera and Araneae occurred in all treatments, with the largest populations obtained in the later stages of culture (Figure 2), i.e., the presence of predators progressed during the vegetative and reproductive stages.
In treatments using Intacta RR2 PRO ® with and without insecticide application, the population of natural enemies was relatively greater when compared with the treatments using RR1, especially with the application of insecticides (Table 1, Figure 2C and D).No use of insecticides for managing caterpillars may have favored the establishment and abundance of insect predators.The use of selective insecticides therefore allows for obtaining a greater number of beneficial insects in agricultural systems (Godoy et al., 2010;Carmo et al., 2009).
However, changes in arthropod populations may be functions of plant density, cultures involved, adaptability of the predator to the culture, density of prey specific to the predator, the availability of other food sources such as pollen, soil moisture and crop microclimate (French et al., 1998;Clark et al., 1997;Cárcamo & Spence, 1994).Thus, the need for studies that consider the complexity of the plant-insect interaction in genetically modified plants is evident to permit greater understanding of population dynamics.Several factors may explain the greater abundance of predators in the treatment using Intacta RR2 PRO ® without insecticide spraying, differentiated from the RR1 treatment by: the absence of defoliation, greater shading and hence a different microclimate, milder humidity and temperatures, conditions that favor an increased number of natural enemies (Letourneau, 1987).Composition of the insect pests studied showed positive interaction for the factors of stage F=37.86,df=22 and 350,p <0.001),F=37.86;dof=22 and 350;p<0,001),F=8.16;dof=2 and 174;p<0.001) and treatment versus application (Pillai-Trace=0.16;F=16.85; dof=2 and 174; p<0.001) (Figure 3), there was no significant interaction of these factors when considering beneficial arthropods, due to the low number of individuals collected.
The statistical difference observed for phenological interaction among stages, population distribution and arrangement of insects is explained by overlapping generations and species of arthropods occurring in the culture throughout plant development.The RR1 and Intacta RR2 PRO ® soybeans technologies significantly affected pest composition when taking into account the target pests of Bt technology.However use of the insecticide applied whenever the action level was exceeded caused a significant reduction in the number of specimens when compared with the treatments without insecticides.
Some species contributed significantly to the difference between treatments evaluated.Arrows indicate the importance of each species within the arrangement (Figure 3), where the species E. heros and Spodoptera spp.contributed most to the ordering of the arrangement obtained because these were common in Intacta RR2 PRO ® soybeans with and without insecticide application.This is justified because these species are not targeted insect pests of Bt technology.As for the species A. gemmatalis and C. includens, susceptible to the Cry 1Ac protein (Bernardi et al., 2012), the samples located at the left and top (Figure 3) are similar for the RR1 soybean with or without application, independent of the culture stage.

Conclusions
The Intacta RR2 PRO ® soybean provides greater diversity of species compared to RR1 soybeans.
The arthropod population suffered a gradual increase during the development of soybean, where the greatest abundance occurs between stages V8 and R6.
The beneficial arthropods in Intacta RR2 PRO ® soybeans are more abundant compared to RR1 soybean.
The main target insect pests and non-targets of Bt technology significantly contribute to the differentiation between pest populations in RR1 and Intacta RR2 PRO ® soybeans, regardless of the use of chemical control.

Figure 1 .
Figure 1.Population fluctuation of insect pests in the vegetative and reproductive stages of soybeans in the treatments, (A) RR1 with insecticide application, (B) RR1 without insecticide application, (C) Intacta RR2 PRO ® with insecticide application, (D) Intacta RR2 PRO ® without insecticide application at four locations in the state of Mato Grosso do Sul, Brazil, 2013

Figure 2 .
Figure 2. Population fluctuation of beneficial arthropods in the vegetative and reproductive stages of soybean treatments, (A) RR1 with insecticide application, (B) RR1 without insecticide application, (C) Intacta RR2 PRO ® with insecticide application, (D) Intacta RR2 PRO ® without insecticide application at four locations in the state of Mato Grosso do Sul, Brazil, 2013

Table 4 .
Diversity indices of Shannon-Wiener, Margalef and Equitability calculated for genetically modified soybeans, RR1 and Intacta RR2 PRO ® at four locations in the state of Mato Grosso doSul, Brazil, 2013