A colony of T. pretiosum was established in the Laboratory of Ecotoxicology at the Department of Entomology (Universidade Federal de Lavras). The colony was kept at 25 ± 2 C, 70 ± 10 % RH and a 12:12 h L:D photoperiod, and reared on UV-sterilized eggs of Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) (Insecta Produtos Biológicos, Lavras, Minas Gerais, Brazil). Host eggs (≤ 24 hours old), glued with arabic gum (50%) onto blue paper cards (8 × 1 cm²), were exposed to a 24-hours T. pretiosum kept in 1-L plastic containers sealed with plastic film. Then, the cards were transferred to new plastic containers and kept until adult T. pretiosum emergence, when the adults were supplied new host eggs. Adult wasps were fed honey droplets carefully placed on the container wall.
Adhering to IOBC guidelines, the insecticides were tested at their highest recommended dose rates for the control of cotton pests based on a spray solution of 200 L/ha (AGROFIT 2020). The active ingredients (a.i.), trade names, chemical groups, and concentrations (g or mL of a.i./L of spray solution) were as follows: teflubenzuron 150 g/L (Nomolt, Benzoylphenylurea, 0.125 mL), thiodicarb 800 g/kg (Larvin, Oxime methylcarbamate, 2.5 g), chlorfenapyr 240 g/L (Pirate, Pyrazol analogue, 3.0 mL), flupyradifurone 170.9 g/L (Sivanto, Butenolide, 1.875 mL), and the positive control methomyl 215 g/L (Lannate, Oxime methylcarbamate, 1.0 mL). The negative control consisted in the application of distilled water, the solvent of all insecticides.
Twenty-five T. pretiosum females were individualized in glass vials (8.5 × 2.5 cm) sealed with plastic films and offered, for a 24-hours period, UV-sterilized E. kuehniella eggs (n = 125) glued with arabic gum (50%) onto blue paper cards (8 × 1 cm²). Then, females were removed and the vials were kept at rearing conditions for three pre-determined periods (1, 4 and 8 days) corresponding to the egg-larval, prepupal and pupal parasitoid stages, respectively. Thus, cards with host eggs containing T. pretiosum at the varying life stages were sprayed under a Potter tower (Burkard, Uxbridge, UK) calibrated at 15 lb/pol2 pressure to deposit 1.5 ± 0.5 µL/cm2, as recommended by the IOBC (Sterk et al. 1999). The sprayed cards were air-dried at room temperature and subsequently individualized in glass vials kept at rearing conditions until wasp emergence. Evaluations were performed with the aid of a stereoscopic microscope (40× magnification) to determine egg parasitism (indicated by the blackening of the vitelline membrane of the host egg) and parasitoid emergence (presence of a wasp exit hole in the host egg) and, following adult emergence, the wasp sex (male or female) and deformation status (occurrence of antennaless and/or wingless adults). The bioassay was held in a completely randomized design (CRD) and a factorial layout (3 parasitoid stages × 6 insecticidal treatments), yielding 18 treatments with 5 replicates (a vial with 5 cards with parasitized host eggs). The assessed endpoints included the F0 emergence rate (emerged adults × 100 ÷ parasitized host eggs), F0 deformation rate (deformed wasp adults × 100 ÷ total wasp adults) and F0 sex ratio [proportion female = Σ♀ ÷ Σ(♀+♂)].
Transgenerational effects of the insecticides were also tested on the surviving F0 adults. Twenty females surviving from the insecticidal treatments were randomly collected and individualized in glass vials, as described previously. Each female was offered 125 UV-sterilized E. kuehniella eggs, free from insecticidal treatment, glued onto paper cards (5 × 0.5 cm²) for 24 hours. F0 female T. pretiosum were held in the vials, honey-fed, and assessed daily for survival until death, whereas the board cars were transferred to new vials for further assessments. Evaluations were also done to determine the F0 parasitism (parasitized host eggs × 100 ÷ total host eggs) and, following the F1 cycle completion, the F1 emergence rate, F1 adult deformation rate and F1 sex ratio. The bioassay was held in a CRD and a factorial layout across a range of parasitoid stages (egg-larva, prepupa and pupa) by insecticidal treatment (negative control and five a.i.) combinations. Due to the low survival of F0 adults emerging from methomyl-treated T. pretiosum, transgenerational studies were not performed for this insecticide.
2.4. Data analysis
All analyses were performed using R software (V. 3.5.3) and RStudio (V. 1.2.5001) (R Core Team, 2019). Data on adult emergence and deformation (F0 and F1), parasitism (F0) and sex ratio (F0 and F1) were checked for normality (Shapiro-Wilk test) and homoscedasticity of residuals (Bartlett test). A two-way ANOVA (ExpDes package; Ferreira et al., 2018) was performed to verify the effect of parasitoid stage, insecticidal treatment and their interaction on the assessed biological traits. In case of significance of the main factors, the Scott–Knott cluster analysis was used to separate their means. Significant interactions were probed with additional ANOVA and Scott–Knott cluster analyses.
Data on F0 female survival were analyzed using Kaplan–Meier estimators to obtain survival curves and estimates of median lethal times (LT50). The overall similarity of the survival curves was tested through the Log-Rank test, and pairwise comparisons among the curves were performed with the Holm–Sidak's test. These analyses were implemented with functions from survival and survminer packages (Therneau 2020; Kassambara et al. 2021).
The insecticides were also categorized into toxicity classes based on their reduction in beneficial capacity (adult emergence of F0 and F1 and F0 parasitism rates) (Sterk et al. 1999). The reduction (R) was calculated thus: R = 100 - (insecticide treatment value × 100 ÷ control value). Toxicity categories were as follows: class 1 = harmless (R < 30%), class 2 = slightly harmful (30% ≤ R < 80%), class 3 = moderately harmful (80% ≤ R ≤ 99 %), and class 4 = harmful (R > 99%).
Afterwards, a hierarchical clustering analysis was performed to assist on interpretation of the bioassays’ outcomes by applying functions (dist and hclust) from stats package. Dissimilarity (Euclidean distance) between all insecticidal treatments was quantified using mean values of the assessed endpoints (emergence, deformation and sex ratio – for F0 and F1, and LT50 and parasitism – for F0). The attributes (pooled mean values for each combination of insecticidal treatment with parasitoid stage) were standardized (mean = 0 and variance = 1) prior to the analysis (Wickham 2018). The dendrogram was constructed using the Ward's minimum variance method (“ward.D”), and the cut-off point was determined as per Mojena (Mojena 1977).