Selectivity of pesticides registered for soybean crop on Telenomus podisi and Trissolcus basalis 1

Brazil stands out among the largest pesticides consumers in the world, what brings concerns about the levels of pesticide residues above thresholds in foods, which may cause health problems to people (Vilca et al. 2017). In addition, pesticides may cause adverse effects to natural enemies, compromising their survival and the biological control of insect pests, requiring greater amounts of pesticides each year to correct the imbalance due to their intensive use (Silva & Bueno 2015). ABSTRACT RESUMO


INTRODUCTION
Brazil stands out among the largest pesticides consumers in the world, what brings concerns about the levels of pesticide residues above thresholds in foods, which may cause health problems to people (Vilca et al. 2017).In addition, pesticides may cause adverse effects to natural enemies, compromising their survival and the biological control of insect pests, requiring greater amounts of pesticides each year to correct the imbalance due to their intensive use (Silva & Bueno 2015).
The use of pesticides is the most effective tactic for pest control in the Brazilian agriculture; however, the choice for a pesticide should be based on its selectivity to non-target species.Thus, the integrated pest management enables the use of chemical and biological control for a more efficient and sustainable agriculture (Ramos et al. 2017).To date, few studies in Brazil have assessed the selectivity of pesticides registered for soybean crops in egg parasitoids of hemipterans (Stecca et al. 2017).Most selectivity tests focus on egg parasitoids of lepidopterans (Carmo et al. 2009, Vieira et al. 2012, Magano et al. 2013and 2015, Stecca et al. 2016).Thus, this study aimed to assess the selectivity of pesticides registered for soybean crop on the stink bug egg parasitoids T. podisi and T. basalis in adult stage.

MATERIAL AND METHODS
Selectivity tests with the parasitoids T. podisi and T. basalis were conducted at the Universidade Federal de Pelotas, in Pelotas, Rio Grande do Sul state, Brazil, in 2016.An adaptation of the standardized method established by the International Organization for Biological and Integrated Control of Noxious Animals and Plants (IOBC), for egg parasitoids of lepidopterans, proposed by Peres & Corrêa-Ferreira (2004), was used.
Three bioassays were carried out in a completely randomized design, with four replications.A total of 15 pesticides registered for soybean crop (Agrofit 2016) on adult parasitoids, using the maximum dosage for each product, were evaluated.The pesticides assessed were 9 insecticides, 4 fungicides and 2 herbicides, whose active ingredients and dosages are shown in Table 1.Each pesticide was compared with a control in which the application was made only with distilled water.
In the treatments, the commercial product was sprayed onto colorless glass plates (2 mm thick, 130 mm x 130 mm in dimension).On the glass plate, 1/ Product registered in Agrofit (2016) for soybean crop; 2/ concentration of active ingredient in the spray mix; 3/ dosage of commercial product (g or mL ha -1 ) registered for an application corresponding to 200 L ha -1 .an acrylic plate of equal size was placed, but with an inside cut to allow a 100 mm x 100 mm area to be hit by the spray.The product was sprayed with a 0.5 L manual sprayer, depositing on the plate 200 mg/100 cm 2 ± 10 % (totaling 2 mg of spray mix per cm 2 of the plate surface = 200 L ha -1 ), which was calibrated by weighing the plate on a precision scale (Holtz et al. 2014).After drying the spray mix, the plates were embedded in a steel structure (13 cm x 1.5 cm x 1.0 cm on each side) fenced on three sides with black voile fabric to prevent parasitoids from escaping.The other side of the cage had two holes, one for inserting the tube containing the parasitoids and the other for inserting cards containing host eggs of E. heros.The plates were covered with a brown cardboard that allowed light inside the cage and were fixed with two clips.
The cages were transferred to an airconditioned room under the same rearing conditions of the host and parasitoids, where emergence tubes were connected (glass tubes with diameter of 10 mm and 100 mm long), containing from 45 to 50 adult parasitoids of about 24 h old.The adults were drawn to the inside of the cages by positive phototropism and the food consisted of a fillet of pure honey offered on a piece of aluminum foil (1 cm x 5 cm) inside the cage.
After 24 h of the cages installation, the emergence tubes were disconnected and, in each cage, a blue card paper (1 cm x 5 cm) with 50 eggs of E. heros was fixed with Arabic gum.The same procedure was repeated at 48 h and 72 h.After 96 h of the bioassay setting, the cages were disassembled and each treatment card was packaged in colorless glass tubes of 8.5 cm x 2.5 cm and sealed on top with voile fabric and a rubber string.These tubes were transferred to an acclimatized room under the same rearing conditions of insects until assessment.
Pesticides were classified based on parasitism reduction, in comparison with the control, and calculated by the equation: PR where PR is the percentage of parasitism reduction; Vt the average parasitism for the treatment; and Vc the average parasitism of the control.Therefore, pesticides were classified according to the IOBC standards, as it follows: class 1: innocuous (PR < 30 %); class 2: slightly harmful (30 % ≤ PR < 80 %); class 3: moderately harmful (80 % ≤ PR < 99 %); class 4: harmful (PR ≥ 99 %).
The data of average number of eggs parasitized per female were submitted to normality analysis by the Shapiro-Wilk test and variance homogeneity by the Bartlett test.When assumptions were not met, the non-parametric Kruskal-Wallis analysis was used, as well as the ranking of means by the Dunn test with Bonferroni correction.Otherwise, the variance analysis (Anova) was carried out.All tests were applied at 5 % of significance.

RESULTS AND DISCUSSION
Insecticides with the active ingredients imidacloprid + beta-cyfluthrin, thiamethoxam + lambda-cyhalothrin, acephate, deltamethrin and fenitrothione differed significantly from the control, regarding the number of eggs/female parasitized.These treatments caused 100 % of parasitism reduction and, therefore, were classified as harmful (class 4) for both parasitoids (Table 2).
The mixture of insecticides of the neonicotinoid group with the pyrethroid group, as in the formulations imidacloprid + beta-cyfluthrin and thiamethoxam +  (Moura et al. 2004).These insecticides act immediately by "shock effect" caused by pyrethroids, as well as by the effect provided by neonicotinoids acting systemically in the plant tissue.Pazini et al. (2016) concluded that this combination of chemical groups of insecticides have a deleterious effect on species of egg parasitoids such as T. podisi and T. pretiosum.Turchen et al. (2015) also observed a rapid mortality caused by lambdacyhalothrin + thiamethoxam to T. podisi, in direct exposure tests in laboratory.Both the researches assessed the same doses of the insecticide evaluated in the present study.
Insecticides of the organophosphorus chemical group, which act by inhibiting an important neurotransmitter, acetylcholinesterase (AChE), have a low molecular weight, which possibly increases its toxicity by favoring the penetration of the molecule insecticide into the insect cuticle (Fukuto 1990).This feature confers these insecticides a broad spectrum of activity, compromising the survival of parasitoids, as observed by Bastos et al. (2006) with monocrotophos and metamidophos insecticide, which reduced the parasitism of T. pretiosum by more than 95 %.Suh et al. (2000) also proved the harmful activity of profenophos to Trichogramma exiguum (Pinto & Platner 1978) (Hymenoptera: Trichogrammatidae), in laboratory tests.Vieira et al. (2001) also reported the harmful effect of organophosphorus trichlorfon to adult Trichogramma cordubensis (Vargas & Cabello 1985) (Hymenoptera: Trichogrammatidae).In their study, the exposure to the pesticide caused a mortality rate of more than 94 % of the parasitoids in the first 24 h of contact with residual insecticide.Although the active ingredients of organophosphorus insecticides in this study differ from those in the studies presented, the fact that they belong to the same chemical group confers them a low selectivity.
The active ingredients diflubenzuron, flubendiamide and lufenuron and the biological insecticide Bacillus thuringiensis did not differ significantly from the control to both parasitoids.Hence, they were classified as innocuous (class 1), with less than 10 % of parasitism reduction (Table 3).Although the average number of parasitized eggs per T. basalis female was low in this bioassay, if compared to the others, the results were validated, because it was a characteristic of the parasitoid strain and did not interfere in the selectivity test.
Insecticides that act as insect growth regulators, such as diflubenzuron and lufenuron (Table 3), act after ingestion and do not normally affect adults of egg parasitoids (Bastos et al. 2006).The insecticides actions are attributed to the inhibition of chitin synthesis in the insect integument, leading to its death by a malformation of the exoskeleton (Tunaz & Uygun 2004).Thus, insecticides belonging to this group tend to be more specific, because they act exclusively on immature stages of insects, with reduced impact on adult hymenopterans that are egg parasitoids (Dhadialla et al. 1998).
The action of the diamides chemical group insecticides activates ryanodine receptors in insects.These receptors play a fundamental role in the insect musculature, and the uncontrolled calcium release in the sarcoplasmic reticulum provoked by insecticides of this group causes a permanent paralysis that tends to lead to the death of the contaminated insect (Cordova et al. 2006).The results of the present study show that the active ingredient flubendiamide is innocuous to T. podisi and T. basalis (Table 3).Pazini et al. (2016) also proved the harmlessness of this insecticide to the egg parasitoids T. podisi and T. pretiosum, which did not have the parasitism affected by contact with the pesticide dry residue.This active ingredient was also innocuous to T. chilonis, characterized as little persistent and with a low potential to change the parasitism rate, despite causing a mortality rate of about 31 % to the parasitoids (Sattar et al. 2011).
The biological insecticide B. thuringiensis has a fast action against pests of the Lepidoptera order and is a control agent that should be used in integrated pest management (Polanczyk & Alves 2003).The selectivity of this commercial product is due to the high specificity of the insecticide proteins to Lepidoptera, Coleoptera and Diptera, not affecting Hymenoptera egg parasitoids (Praça et al. 2004).Adults of T. pretiosum also did not show a parasitism reduction in contact with B. thuringiensis in laboratory, demonstrating the harmlessness of the product to non-target species (Silva & Bueno 2015).
Fungicides containing pyraclostrobin + metconazole, trifloxystrobin + cyproconazole, azoxystrobin + flutriafol and trifloxystrobin + protioconazol did not differ statistically from the control, and were innocuous (class 1) to T. podisi.However, they were slightly harmful (class 2) to T. basalis, differing statistically from the control (p < 0.05), except for trifloxystrobin + protioconazol, which, despite no statistical difference, reduced the parasitism by 33.62 %, being classified as slightly harmful (class 2) (Table 4).The desiccant herbicides glufosinate ammonium salt and glyphosate isopropylamine salt did not differ statistically from the control (p > 0.05), being classified as harmless to both parasitoid species.Pazini et al. (2016) classified fungicides containing the active ingredients trifloxystrobin + protioconazole as innocuous to T. podisi, in agreement with the results in this study (Table 4).However, Magano et al. (2015) classified trifloxystrobin + protioconazole as slightly harmful to T. pretiosum in laboratory tests, showing that this fungicide is slightly harmful to some parasitoid species.The authors also classified pyraclostrobin + epoxiconazole and trifloxystrobin + cyproconazole as moderately harmful, once again proving that mixtures of triazole and strobilurins have bad effects on the parasitoid species.
The chemical group of triazole in mixture with the group of strobilurins is used in the formulation of many fungicides, because they combine the 1/ Mean ± standard error of parasitized eggs per female in each treatment, after 96 h of exposure of parasitoids to pesticide residues.ns Not significant by the variance analysis (F = 0.417 curative power of triazole added to the residual effect provided by strobilurins (Pimenta et al. 2011).Although fungicides focus their actions on fungi, toxicity comes from their formulation with ingredients sometimes toxic and with differentiated selectivity to each parasitoid species (Carvalho et al. 2012).
The major share of herbicide glyphosate on the world market is attributed to its broad spectrum of action and the advent of soybean cultivars resistant to this herbicide molecule (Shaner 2000).Stecca et al. (2016) classified the glyphosate isopropylamine salt as selective to T. remus, when adult parasitoids were exposed to the pesticide applied on glass plates, highlighting its differentiated selectivity to parasitoid species.
Desiccant herbicides are usually selective to egg parasitoids, such as Palmistichus elaeisis (Delvare & La Salle 1993) (Hymenoptera: Eulophidae), which was not affected by the ammonium salt-based desiccant.However, the impact of desiccant herbicides must be studied in other developmental phases of parasitoids for a better understanding of their impact on non-target species (Hassan & Abdelgader 2001).
The results of this research show that the parasitoid T. basalis is more susceptible to some pesticides in its adult stage, such as fungicides (Table 4), if compared to T. podisi.This may influence the predominance between one or another species in the field.This study was conducted with parasitoid lineages coming from the Brazilian midwest, and assays were carried out with species from other regions of the country.Thus, tests with lineages derived from southern Brazil are also necessary for a more accurate conclusion about the selectivity to T. podisi and T. basalis in this region.
Neurotoxic insecticides should be avoided, since they were not selective to parasitoids.The group of diamides, insect growth regulators and biological insecticides should be preferred whenever possible, as they are selective to the two parasitoid species tested here.Triazole-based fungicides combined with strobilurins were not selective to T. basalis, although semi-field and field tests are needed for a better understanding of their impact under these conditions.Other pest control methods, such as transgenic plants, do not have an effect on Hemipterans yet; thus, the use of pesticides selective to parasitoids of stink bug eggs is of great importance to the integrated pest management application in the soybean crop.CONCLUSIONS 1.The neurotoxic insecticides imidacloprid + betacyfluthrin, thiamethoxam + lambda-cyhalothrin, acephate, deltamethrin and fenitrothione are not selective to T. podisi and T. basalis, but the insecticides diflubenzuron, flubendiamide, Bacillus thuringiensis and lufenuron are selective to both parasitoids; 2. Fungicides formulated with pyraclostrobin + metconazole, trifloxystrobin + cyproconazole, azoxystrobin + flutriafol and trifloxystrobin + protioconazol are selective to T. podisi, but not to T. basalis; 3. Herbicides containing glufosinate ammonium salt and isopropylamine salt are selective to the egg parasitoids T. podisi and T. basalis.