Introduction

Corn (Zea mays L., Poaceae) is the most important cereal originating in the Americas and planted on large scale worldwide. Caterpillars and weeds are the main pests of the corn crop in Brazil (^{Zanuncio et al. 2013}; ^{Menezes et al. 2014}). The caterpillars are controlled using insecticides, however, these products can cause environmental contamination and alternative methods are necessary to manage these insect pests (^{Tavares 2010}). Natural enemies, especially egg parasitoids such as Trichogramma spp. (Hymenoptera: Trichogrammatidae), parasitize Spodoptera spp. (Lepidoptera: Noctuidae) and other caterpillars (^{Spínola-Filho et al. 2014}). These organisms can reduce damage by caterpillar pests in corn crops and their parasitism on eggs prevents larvae from hatching (^{Gardner et al. 2011}).

Weeds can reduce corn yield by up to 85 % (^{Fickett et al. 2013}; ^{Pascoaloto et al. 2017}) and herbicides are exclusively used for their control. Herbicides with atrazine (triazine class), which inhibits the photosystem II causing irreversible damage to plant cells (^{Chen et al. 2014}), are the most widely used to control dicotyledonous plants at pre- or post-emergence in this crop (^{Das et al. 2010}).

Herbicides can affect Trichogramma spp. parasitism because their active ingredient can penetrate the insects’ cuticle (^{Leite et al. 2015}). The herbicide’s effects on parasitoids may vary with the salt quantity and type and adjuvants or their mixture, which isusually done in the field (^{Stefanello Júnior et al. 2011}). Trichogramma spp. can be used as a model to determine the selectivity of agrochemicals to natural enemies (^{Menezes and Soares 2016}).

This study aimed to evaluate the compatibility of parasitoids with atrazine herbicide by toxicological and behavioral impacts of parasitoid females from 10 species of Trichogram-matidae (9 Trichogramma spp. + 1 Trichogrammatoidea sp.).

Materials and methods

This study was conducted at the Laboratory of Entomology and in the G.W.G. de Morais Insectarium of the Institute of Agricultural Sciences (ICA) of the “Universidade Federal de Minas Gerais (UFMG)” in Montes Claros, Minas Gerais state, Brazil, in 2014.

The experiment had a completely randomized design with 10 parasitoid species and one herbicide, besides the control, with 10 replications. Each parcel had two white paper cards (0.4 cm width × 2.0 cm length) with 45 Anagasta (Ephestia) kuehniella Zeller, 1879 (Lepidoptera: Pyralidae) eggs (200 total cards). Ten Trichogrammatidae species, commonly found in the crops in Brazil, were obtained from the Insectarium of the ICA/UFMG, with nine species of the genus Trichogramma - T. acacioi Brun, Moraes and Soares, 1984; T. atopovirilia Oatman and Platner, 1983; T. bennetti Nagaraja and Nagarkatti, 1973; T. brasiliensis Ashrnead, 1904; T. bruni Nagaraja, 1983; T. demoraesi Nagaraja, 1983; T. galloi Zucchi, 1988; T. pretiosum Riley, 1879, and T. soaresi Nagaraja, 1983, and one of the genus Trichogrammatoidea - T. annulata De Santis, 1972. The treatments were carried out with the herbicide atrazine - Gesaprim 500 Ciba-Geigy® (recommended commercial dose for corn) and distilled water (control) (^{Stefanello Júnior et al. 2008}; ²⁰¹¹).

A total of 45 A. kuehniella eggs were glued per white paper card with 10 % Arabic gum, exposed to ultraviolet radiation (UV) for 60 min, placed in glass vials (7.5 cm diameter × 13.0 cm height), sealed with plastic polyvinyl chloride (PVC) and an elastic film, and stored in a refrigerator at 5 °C and 80 % R.H. for 24 h. After this period, each card was sprayed with atrazine using a Guarany® hand sprayer (Itú, São Paulo, Brazil) until runoff began (Gesaprim 500 Ciba-Geigy®) at 8.07 L.ha-1 (20.16 × 10-5 mL of commercial product/card, 10.08 × 10-2 mg of active ingredient/card). The control eggs were sprayed with distilled water. The hand sprayer used was measured and had its conformity evaluated by the “Instituto Nacional de Metrologia, Qualidade e Tecnologia (INMETRO)” of the “Ministério do Desenvolvimento, Indústria e Comércio Exterior (MDIC)” of Brazil. The hand sprayer was tested three times (replicates) with a 30 min calibration process before its use. The cards were subsequently kept in the shade outdoors for 2 h to evaporate water excess and placed in sealed transparent glass test tubes (1.0 cm diameter × 9.0 cm height) with a newly emerged female parasitoid (< 24 h old, without food) for 48 h at 12:12 h (light:dark) photoperiod and 24.39 ± 0.01 °C (^{Soares et al. 2012}; ²⁰¹⁴). Each test tube received two cards with treatment and control. The parasitism was allowed until Trichogrammatidae female death (after about six days). The experimental plot was repeated if females died before 48 h of parasitism.

The parasitized eggs, male and female emergence and the female-biased sex ratio (number of females ÷ number of males + number of females) were evaluated. Parasitized eggs were those with dark colour after 20 days of subjection to parasitism, and the non-parasitized of yellow colour (^{Pratissoli et al. 2004}; ^{Prezotti et al. 2004}). Male and female Trichogrammatidae were identified according to the antenna dimorphism (males have feathery antennae and females nailed ones) (^{Zuim et al. 2017}), using a binocular microscope with 40 × magnification.

Herbicide toxicity was classified based on the parasitism and emergence rate reduction as follows: I = harmless (< 30 % reduction), II = slightly harmful (30 - 79 % reduction), III = moderately harmful (80 - 99 % reduction), and IV= harmful (> 99 % reduction) (^{Sterk et al. 1999}). The reduction in the emergence rate of the parasitoid species was calculated as follows: % reduction = 100 - mean [(% mean of the treatment ÷ % mean of the control) × 100] (^{Carvalho et al. 2010}). The female parasitism and emergence reduction were calculated with the formula: % reduction = 100 - mean [(% overall of the treatment with insecticide ÷ % overall of the control treatment) × 100]. The data were transformed to arcsine, also submitted to analysis of variance (one-way ANOVA) and the means examined using the Tukey’s HSD (honest significant difference) test at 1 % or 5 % probability. The “sistema para análises estatísticas (SAEG)”, version 9.1 (Supplier: UFV) was the program used.

Results

Atrazine changes the behavior of parasitoid females by reducing the parasitism rate (≈ 71 %) of the Trichogrammatidae species, except for T. Galloi and T. bennetti. This herbicide was harmless (< 30 % reduction) to these two natural enemies but it was moderately harmful (80 - 99 % reduction) to the other parasitoid species (Table 1).

The emergence rate (≈ 74 %) of Trichogrammatidae females, except those of T. galloi and T. bennetti, was lower with atrazine. No parasitoid female emerged from eggs parasitized by T. acacioi, T. annulata, T. atopovirilia, T. bruni, T. brasiliensis, T. demoraesi, and T. soaresi that had been previously sprayed with atrazine. This herbicide was harmless to T. bennetti and T. galloi, but moderately harmful to T. pretiosum and harmful (> 99 % reduction) to the other Trichogrammatidae species (Table 1).

The sex ratio (≈ 74 %) of ten Trichogrammatidae species, except for T. galloi and T. bennetti, was lower with atrazine. This herbicide was harmless to these two Trichograma species but moderately harmful to T. pretiosum, and harmful to the other Trichogrammatidae species (Table 1).

Discussion

Trichogrammatidae species showed a reduction in the parasitism behavior and the rates of emergence of females in eggs sprayed with atrazine, probably due to the effect of repellency to oviposition and also to the mortality of immatures inside the host eggs (^{Leite et al. 2015}). The effects of atrazine-based herbicides on Trichogrammatidae species depend on the doses used and formulations of this chemical (^{Stefanello Júnior et al. 2008}; ^{Menezes and Soares 2016}). The lower parasitism rate by the Trichogrammatidae species (repellence effect of atrazine), except in T. galloi and T. bennetti, can be explained by the ability of these species to detect nocive substances, rejecting hosts and, consequently, reducing parasitism. Trichogramma galloi may be more resistant to pesticides as chlorantraniliprole (insecticide, anthranilic diamides chemical group), sulfometruron methyl (plant growth regulator, sulfonylurea chemical group), and triflumuron (insecticide, benzoylurea chemical group) that were demonstrated to not repel or reduced its parasitism rate in a previous Brazilian study (^{Antigo et al. 2013}). Tolfenpyrad (insecticide, pyrazole chemical group) reduced the foraging behavior of T. pretiosum but this was not found with 11 other pesticides, indicating that parasitoids could successfully forage on eggs treated with many pesticides in a Pakistani study (^{Khan et al. 2015}).

The lack of impact of atrazine on the emergence of T. galloi and T. bennetti females may be due to its detoxification capacity (^{Stefanello Júnior et al. 2011}; ^{Oliveira et al. 2014}). This indicates greater resistance of these parasitoid species to atrazine. Primoleo® (herbicide, triazine chemical group) and Siptran 500 SC® (herbicide, triazine chemical group) were classified as class I (harmless, < 30 % reduction) and the Gesaprim GrDA® (herbicide, triazine chemical group) as class II (slightly harmful, 30 - 79 % reduction) for the mortality of T. pretiosum adults in a previous Brazilian study (^{Stefanello Júnior et al. 2008}). The survival of the predator Podisus nigrispinus Dallas, 1851 (Hemiptera: Pentatomidae), the parasitoids T. atopovirilia, T. bennetti and T. brunni, and soil arthropods such as Collembola, mites (Acari), and ants (Hymenoptera: Formicidae) was lower than 50 % when submitted to herbicides with atrazine in three different Brazilian studies (^{Pereira et al. 2005}; ^{Lins et al. 2007}; ^{Menezes et al. 2012}). The higher emergence rate of T. pretiosum, T. demoraesi, T. galloi, and T. soaresi with atrazine in a no-choice test in a Brazilian study may be related to the “hormesis” phenomenon (^{Leite et al. 2015}) where in sublethal quantities of a stressor benefit an organism (^{Pratissoli et al. 2010}). This hypothesis is possible, especially considering that the herbicide quantity reaching the parasitoid inside the host egg is very low (^{Leite et al. 2015}). However, the “hormesis” phenomenon with these four Trichogramma species was not detected in this work (free-choice test), possibly due to the reduction of the parasitism and, consequently, female emergence rate.

No impact of atrazine on T. galloi and T. bennetti sex ratio may show, as explained before, host egg protection and detoxification capacity (^{Stefanello Júnior et al. 201}1; ^{Oliveira et al. 2014}), but it needs to be better studied to understand the responses of both parasitoids to the herbicide. The emergence of Aedes (= Stegomyia) aegypti L., 1762 (Diptera: Culicidae) and Aedes (= Stegomyia) albopictus Skuse, 1894 (Diptera: Culicidae) were higher with atrazine than with glyphosate (herbicide, glycine chemical class) in a North American study. The sex ratio was distorted with male bias observed in the control and glyphosate treatments, but not with atrazine (^{Bara et al. 2014}). The emergence period for both sexes of these two mosquito species was longer with atrazine than with glyphosate and in the control (^{Bara et al. 2014}). The reduced sex ratio of T. bruni, T. atopovirilia and T. bennetti and higher values for T. demoraesi and T. soaresi in no-choice tests in a Brazilian study (^{Leite et al. 2015}) shows that the impact of atrazine, a widely used herbicide, on life-history traits as reported for mosquitoes (^{Bara et al. 2014}) depends on the species of insect.

Conclusions

The parasitoids T. bennetti and T. galloi showed adequate values of parasitism and emergence in eggs previously sprayed with atrazine and can be used simultaneously with the application of this herbicide.

Trichogrammaacacioi, T.annulata, T.atopovirilia, T.bruni, T.brasiliensis, T. demoraesi, and T. soaresi species should be released, after the residual toxicity period of the herbicide, to avoid reducing its efficiency in biological control.