Direct and indirect effects of selective insecticides on 2 generalist predators of Bemisia tabaci (Hemiptera: Aleyrodidae)

Abstract

Generalist arthropod predators have historically contributed to the suppression of arthropod pests in many agroecosystems. The successful implementation of integrated pest management (IPM) programs hinges on the incorporation of insecticides that are compatible with the biological attributes of natural enemies of pests. A potentially promising pathway is improving biological control by natural enemies through the timely application of selective insecticides. In our study, adult predators were exposed to commercially available insecticides (cyantraniliprole and pyriproxyfen) using a combined laboratory and field approach to assess their effects on survivorship and predation. We isolated 2 predators, Hippodamia convergens Guérin-Méneville and Geocoris punctipes (Say), in the laboratory to estimate the survivorship and consumption of whitefly nymphs, Bemisia tabaci (Gennadius). In the field, we deployed whitefly nymph-infested potted cotton plants in replicated cotton plots with both insecticide treatments. We enumerated whitefly nymph populations on enclosed (predator-exclusion) and open (predator-accessible) potted plants. While pyriproxyfen had a negligible effect on the predators, cyantraniliprole exposure directly affected H. convergence by reducing survivorship duration and indirectly influenced both predators by reducing prey consumption and altering the consumption of alternative prey. In field conditions, regardless of pesticide exposure, whitefly-infested potted plants that excluded predators had more whiteflies than predator-accessible potted plants. Overall, pyriproxyfen demonstrated minimal impact on the predators in the laboratory or field, while cyantraniliprole adversely influenced mortality and indirect foraging under controlled laboratory conditions but did not have a significant impact in the field.

Introduction

Integrated pest management (IPM) involves the regulation of pest species through multi-layered regulation techniques (Bueno et al. 2017). Two common pillars of IPM applied to insects involve employing biological control agents and selective insecticide applications in tandem (De Clercq et al. 1995, Gray et al. 2009). However, growers often rely solely on chemical options to manage pests and commonly do so with nonselective insecticides (Machado et al. 2019). Consequently, the excessive use of broad-spectrum insecticides results in mortality and reduced fecundity of natural enemies, leading to a lower abundance of biological control agents and associated pest control services (Fleeger et al. 2003, Soundararajan 2012). In addition, the overuse of broad-spectrum insecticides indirectly reduces biological control by lowering feeding activity and altering the foraging behavior of natural enemies (Desneux et al. 2007, Soundararajan 2012, Ndakidemi et al. 2016) and can result in insecticide resistance in pest populations that spark pest outbreaks (Bueno et al. 2017). Given the multiple negative effects tied to the overuse of broad-spectrum insecticides, finding ways to pair selective insecticides with minimal non-target effects on natural enemies (Fernandes et al. 2010, Perdikisa et al. 2011) has the potential to reduce risks to natural enemies and synergize the activity effects of biological control agents on pests (Duso et al. 2020).

Selective insecticides target specific insect pests while minimizing harm to biological control agents and the environment (Atanassov et al. 2003, Roubos et al. 2014). These compounds have demonstrated their efficacy in managing pests while preserving populations of natural enemies like parasitoids and predators (Singh et al. 2004, Naranjo and Ellsworth 2009a, Ohnesorg et al. 2009, Torres and Bueno 2018, Vandervoet et al. 2018, Machado et al. 2019, Bordini et al. 2021). For example, Naranjo et al. (2004) reported that the application of insect growth regulators (IGRs) buprofezin and pyriproxyfen to manage whiteflies in cotton fields led to increased predator survival and enhanced predation on secondary cotton pests compared to a conventional chemical regime. Therefore, the use of selective insecticides can help promote biological control by preserving natural enemy communities (Kraiss and Cullen 2008, Roubos et al. 2014). However, natural enemy responses to insecticides and cropping systems may be species-specific. For instance, Grafton-Cardwell et al. (2006) reported that pyriproxyfen and buprofezin interfere with the pupation and molting of the predatory beetle, Rodolia cardinalis (Mulsant) (Coleoptera: Coccinellidae) in citrus orchards. Other laboratory and field studies support the detrimental impacts of pyriproxyfen, chlorantraniliprole, and spirotetramat on the survival and development of some natural enemies, such as Podisus maculiventris, Tenuisvalvae notata, Eretmocerus mundus, and Ceraeochrysa cincta. (De Clercq et al. 1995, Francesena et al. 2012, Rugno et al. 2016, Barbosa et al. 2018, Schmidt-Jeffris 2023). Therefore, while some “selective” insecticides can be compatible with biological control efforts, careful consideration, and continued research are needed to optimize IPM practices.

The sweetpotato whitefly, Bemisia tabaci (Gennadius, 1889) (Hemiptera: Aleyrodidae), is a recurrent pest in agricultural landscapes producing cotton and vegetables (Simmons et al. 2008, Li et al. 2021, Perier et al. 2022). High infestations of B. tabaci create significant economic losses (Riley and Palumbo 1995, Naranjo et al. 1998, Li et al. 2021) in the form of direct (i.e., reduced quality of crops) or indirect (i.e., virus transmission) damages (Xue et al. 2010, Velasco-Hernández et al. 2013). The use of insecticides is the most common management tactic for B. tabaci (Horowitz et al. 2020). While it is preferable to use insecticides with low impact on natural enemies to minimize the likelihood of secondary pest outbreaks in cotton areas, sometimes farmers turn to broad-spectrum insecticides (e.g., organophosphate, pyrethroids, and neonicotinoids) to manage pests infesting cotton fields (Barros et al. 2018). Consequently, repetitive use of insecticides has led to the development of insecticide resistance and the loss of natural enemy populations (Palumbo et al. 2001, Horowitz et al. 2020, Mahmood et al. 2023). Currently, pyriproxyfen and cyantraniliprole are among the commonly used insecticides for managing B. tabaci in the cotton agroecosystem (Horowitz et al. 2011, Perring et al. 2018, Abubakar et al. 2022, LaTora et al. 2022). However, few studies have attempted to integrate the use of these insecticides with the activity of natural enemies. (e.g., Naranjo and Ellsworth 2009a, Prabhaker et al. 2017, Kheirodin et al. 2020, Bordini et al. 2021).

In this study, we aimed to determine the effects of 2 selective insecticides (pyriproxyfen and cyantraniliprole) on 2 known predators of B. tabaci found in many agricultural systems (Elzen 2001, Hagler 2002, Naranjo 2002, Naranjo et al. 2002, Hagler and Naranjo 2005, Perdikisa et al. 2011, Hagler and Blackmer 2013, Kheirodin et al. 2022). The generalist predators Geocoris punctipes (Say) (Hemiptera: Geocoridae) and Hippodamia convergens Guérin-Méneville (Coleoptera: Coccinellidae) are among the most abundant predator species found in pest-infested crops and contribute to pest control services (Whitcomb and Bell 1964, Hagler 2002, Naranjo et al. 2002, Prabhaker et al. 2017). Through a controlled set of microcosm experiments and a field experiment, we investigated the compatibility of 2 commonly applied selective insecticides (pyriproxyfen and cyantraniliprole) with the 2 aforementioned common whitefly predators. In the laboratory, we measured the impact of insecticides on predator feeding rate (voracity), survivorship, and 2 behaviors: foraging location (i.e., the movement from plants exposed to insecticides) and foraging on alternative prey. In the field, we quantified the responses to insecticide exposure on whitefly biocontrol predation services provided by Hippodamia convergens and Geocoris punctipes. For laboratory experiments, we hypothesized that exposure to cyantraniliprole and pyriproxyfen-treated prey would not lower influence predator survivorship and or alter predator foraging behavior, which, if truly “selective,” then may result in synergy or compatibility with whitefly control. When provided a choice, we expected predators would detect the presence of an insecticide and avoid feeding on treated plants. For the field experiment, we hypothesized that predators avoid insecticide-treated plants; therefore, we predict lower predation services within insecticide-treated plots.

Materials and Methods

Study System

Host Plant, Prey Colonies, and Insecticides

Cotton, Gossypium hirsutum L., was the host plant used for all predator experiments. Likewise, cotton plants served as host plant material for maintaining prey colonies. Libertylink Cotton (var. ‘ST-4550GLTP’, Stoneville, BASF Corp., Research Triangle Park, NC, USA). was seeded in PRO-MIX (BX Mycorrhizae) medium and was fertilized weekly with Miracle-Gro All Purpose Plant Food (NPK=20-20-20; Scotts Miracle-Gro Company, Marysville, OH, USA). All plants were grown in growth chambers set at 30 ± 2°C, a 14L/10D photoperiod, and relative humidity of 60% ± 5%. Specifically, we used 4- to 6-wk-old cotton plants to establish the desired number of whitefly nymphs for subsequent experiments.

Prey colonies were established and maintained under controlled conditions at 25 ± 3 °C, a 14L/10D photoperiod, and relative humidity of 50% in rearing rooms on site. These conditions were kept constant for all colonies. The B. tabaci (whiteflies), specifically the MEAM1 cryptic species, colony was established using specimens collected from a preexisting laboratory colony at the Coastal Plains Research Station, University of Georgia, Tifton, GA, USA, that had been maintained for at least 4 yr. Cucumber plants, Cucumis sativus L., were often supplemented as a host to bolster the population density of the whitefly colony. In addition, a colony of aphids, Aphis gossypii (Glover, 1877) (Hemiptera: Aphididae) was established on cotton plants under conditions like the whitefly colony. The original aphid specimens used to create the colony were field collected at the Lang-Rigdon Farm in Tifton, GA, USA (31°30ʹ58″N, 83°32ʹ51″W) during the summer months. Both aphids and whitefly colonies served as prey and were used to infest cotton plants as required for each specific experiment.

For laboratory experiments, newly emerged adult whiteflies were placed in cages with insect-free cotton plants for 72 h for oviposition. After oviposition, adults were removed, and the infested plants were placed under colony-rearing conditions consistent with the abovementioned rearing conditions. Whitefly eggs were allowed to develop into third-instar nymphs over 16 days with periodic observation every 3 days. Fourth-instar aphid nymphs were collected from the aphid colony and transferred to insect-free cotton plants as needed for the experiments. Aphids were used as alternative prey to evaluate predator–prey preference and consumption of alternative prey when exposed to insecticides.

Two insecticides commonly used for whitefly and aphid management were selected as insecticide treatments. The selected insecticides, cyantraniliprole (Exirel, 0.83SC, FMC Corp., Philadelphia, PA, USA) and pyriproxyfen (Knack, 0.86EC, Valent LLC, San Ramon, CA, USA), were obtained from suppliers. For the specific experiments, treatments were diluted according to the recommended label rates, calculated to a volume of 1 L (Table 1).

Predators

Due to their high abundance in Georgia, USA cotton fields and predation on both prey species (Tillman and Mulrooney 2000, Bowers et al. 2021, Kheirodin et al. 2022, 2023, Schmidt et al. 2024), we selected 2 generalist predators, Geocoris punctipes and H. convergens. Predators were selected to represent 2 major functional groups of predators (i.e., piercing-sucking and chewing predators, respectively). Geocoris punctipes adults were collected from a field at the University of Georgia Ponder Research Farm, GA, while H. convergens adults were purchased from an online supplier (Nature’s Control, Phoenix, OR, USA). Both predators were maintained in growth chambers under controlled conditions (i.e., 25 ± 3°C, a 14L/10D photoperiod, RH = 50%) and provided diets of A. gossypii and 5% sugar solution ad libitum for sustenance. Before all trials, predators were food-deprived for 48 h to standardize hunger levels.

Laboratory Bioassay Conditions

All laboratory experiments were conducted at the Coastal Plains Research Station, University of Georgia, Tifton, GA. These experiments included evaluations of predator survivorship, predator foraging location in response to insecticide application, prey preference/ consumption of alternative prey by predators, and predator voracity on whitefly nymphs (i.e., the proportion of eaten prey from the total number of initially given prey). The laboratory conditions were consistent with the colony-rearing conditions as previously described.

Cotton leaves infested with second- to third-instar whitefly nymphs were retrieved from previously infested plants. Second- to third-instar nymphs were chosen because they remain feeding and do not move around on plants. The petioles of these leaves were then placed inside individual 1.5 ml microcentrifuge tubes (Eppendorf) containing Miracle-Gro fertilizer solution. Afterward, the leaves were sprayed with treatments, including a control (water), pyriproxyfen, or cyantraniliprole (Table 1). All treatment solutions included a methylated seed oil nonionic surfactant to ensure the treatments adhered to the leaves after application (Tazer 40 µl, Table 1). Individual 1 L spray bottles were used for applications with 3 pumps (~0.3 ml) representing the application rate. Immediately after, the leaves were left to air dry for 1 h before being placed in bioassay chambers (473.18 ml plastic containers). The lids of the bioassay chambers were perforated to facilitate airflow, and cotton balls soaked in water were added for hydration and humidity. The starved predators were then released into the bioassay chambers, and the lids were sealed (Fig. 1A). Each predator (H. convergens and G. punctipes) and treatment (control, pyriproxyfen, and cyantraniliprole) was replicated 10 times for each experiment. The bioassay chambers were then placed inside growth chambers under controlled conditions at 25 ± 3°C, a 14L/10D photoperiod, and relative humidity of 50% for the experiment duration. Generally, after each experiment, we examined the cotton leaves under a stereomicroscope (Labomed Luxeo 6Z LED) and recorded the number of consumed preys; missing prey was assumed to be consumed.

Natural Enemy Survival and Predator Voracity Responses to Insecticide-exposed Prey

To evaluate the direct effects of insecticide application on H. convergens and G. punctipes, both species were supplied with fresh prey for 7 days on cotton leaves treated with insecticide in their respective bioassay chambers. Specifically, we replaced the leaves with newly sprayed cotton leaves infested with 3 second- to third-instar whitefly nymphs daily for 7 days. In addition, fresh water-soaked cotton balls were provided as a source of water and humidity. Predator survivorship was evaluated every 24 h throughout the 7-day period. To confirm that the predators fed on the whitefly nymphs, leaves were examined under a stereomicroscope (Labomed Luxeo 6Z LED) periodically.

Assessments were carried out on predator voracity to evaluate the indirect effects of these insecticides on both predators. Predator voracity in this study refers to the ability of an individual predator to consume a specific number of preys within a designated timeframe. In this study, we estimated predator voracity by subtracting the number of whitefly nymphs consumed in 24 h from 10 whitefly nymphs, the initial number of whitefly nymphs offered. The voracity was then compared between insecticide-treated and non-treated control to examine the potential effect of insecticides on predator voracity.

Consumption of Alternative Prey

Aphids were selected as alternative prey to whiteflies because of their similar feeding behavior. Aphids are common in production systems containing whiteflies (Kheirodin et al. 2023). Furthermore, recent studies show that the presence of aphids is correlated with changes in whitefly predation, and field collected predators frequently test positive for the presence of aphid DNA (Bowers et al. 2021, Kheirodin et al. 2023, Schmidt et al. 2024). Therefore, aphids were used in this evaluation to assess potential changes in whitefly predation following insecticide exposure when alternative prey is available.

In this experiment, we used cotton leaves with 5 whitefly nymphs from previously infested plants. Five fourth-instar aphid nymphs (alternative prey) were introduced onto the leaves. Whiteflies and aphids were presented as prey options to predators, with the aim of evaluating the consumption of whiteflies and alternative prey. The treatments (control, pyriproxyfen, and cyantraniliprole) were consistent with earlier methods evaluating natural enemy survival and predator voracity responses to insecticide-exposed prey and followed the same application method (Table 1). After 24 h, the predators were removed, and leaves were observed under a stereomicroscope (Labomed Luxeo 6Z LED) to record the number of whiteflies and aphid nymphs remaining; missing prey was assumed to be consumed. The proportion of whitefly nymphs consumed by predators was then compared between treated and non-treated exposure conditions to determine the potential effects of insecticides on predator preference.

Predator Foraging Location in Response to Insecticide Presence

Sixty polyethylene tube-caged chambers were created for the experiment. Caged chambers consisted of 2 polyethylene tube arms (15 cm diameter, 34.29 cm length), sealed at the top with nylon chiffon, connected by a passageway (5 cm diameter, 21.59 cm length tube) (Fig. 1B). Choice chambers were created to investigate the foraging behavior of individual predators exposed to both prey species (aphid nymphs and whitefly nymphs) with a choice of plants treated with insecticides or untreated. One chamber contained a plant sprayed with insecticide, and the connected choice chamber contained a plant that was insecticide-free. Because insecticides are often applied with surfactants, the untreated received a surfactant application. For this bioassay, we recorded 2 parameters: (i) predator choice, which is defined as their location after a 24-h exposure period in one of the 2 areas (insecticide-treated or insecticide-free) and (ii) feeding on prey available (i.e., aphid nymphs and whitefly nymphs).

A selection of 5-wk-old cotton plants infested with whitefly nymphs was used for this experiment. Plants with 5 second- to third-instar whitefly nymphs were selected to standardize the developmental stage of nymphs. Whitefly nymphs were removed from leaves that contained more than 5 nymphs using a camel hair brush. Five fourth-instar aphids were transferred to each plant. Whitefly nymphs and aphid nymphs were used as foraging prey options for predators. Plants were treated with either control (water), pyriproxyfen, or cyantraniliprole (Fig. 1A). Cages were placed over the cotton plants and secured into the soil to prevent predator escape. A pair of cotton plants were enclosed within separate arenas with a connection tunnel. Specifically, each insecticide-treated plant was paired with a control (water). Both control types were also paired; water + surfactant (treated control) and water only (non-treated control), to test the impact of the surfactant on the parameters mentioned above. Before being released in the center of the passageway, the predators underwent 48 h of starvation (Fig. 1B). Each treatment was replicated 10 times for each predator (H. convergens and G. punctipes) and treatment pairing (insecticide + non-treated control and treated control + non-treated control). After 24 h, the location of the predator was recorded (insecticide-treated or insecticide-free chamber), and their survivorship was recorded (dead or alive). Thereafter, cotton leaves were observed under a stereomicroscope (Labomed Luxeo 6Z LED) to record the number of remaining prey individuals.

Field Evaluation of Natural Enemy Predation

Two University of Georgia research stations were used for the field experiment: Lang-Rigdon Farm in Tifton, GA (31°30ʹ58″N, 83°32ʹ51″ W), and Stripling Irrigation Research Park in Camilla, GA (31°16ʹ49″N, 84°17ʹ39″W). Land preparations at both sites began in June 2022. They included an herbicide program of pre-plant incorporated trifluralin, post-planting overtop clethodim, and a single application of the pre-emergent herbicide S-metolachlor (see Table 1 for application rates). Fertilizer (NPK = 10-10-10) (~92 kg ha⁻¹) was also incorporated into the soil via broadcast application. The insecticide treatments were consistent with the laboratory trials: control (water), pyriproxyfen (Knack), and cyantraniliprole (Exirel) along with methylated seed oil nonionic surfactant (Tazer) diluted according to the recommended application rates and applied to trial plots (see Table 1 for application rates). Insecticide treatments were applied using a CO₂ tank-coupled hand sprayer at a constant working pressure of 206.8 kPa. Insecticide-treated plants were left undisturbed for 6 days at the site locations. The experiments at both locations (Tifton and Camilla) were established in a spilt-plot RBD (split being half was smooth (glabrous) leaf variety Deltapine “DP 2055 B3XF” (Bayer CropScience LP, St. Louis, MO, USA), and hairy (pubescent) leaf variety Stoneville “ST 4550GLTP” (BASF Corp., Research Triangle Park, NC, USA), with 3 field blocks and 4 replications per treatment, a total of 18 trial plots were established at both sites (Tifton and Camilla) (Fig. 2). Specifically, each plot covered four rows with a 0.91 m width and 12 m length at 0.30 m plant spacing. An additional 1.83 m buffer zone between plots was established to avoid insecticide drifting among plots (Fig. 2).

We estimated whitefly nymph predation (whitefly nymph removal) within cotton plots using pairs of two 6-wk-old potted cotton plants that were placed alongside 6-wk-old cotton plants planted in the field, with one potted plant open to predation (i.e., without cage) and the other one closed. The closed cage was constructed using a modified tomato cage (i.e., 83.82 cm galvanized steel wire round with top ring removed) and a fine mesh (nylon chiffon, 0.01 cm openings). Closed cages enclosed the potted plant to protect the whitefly nymphs from natural predation. Cotton plants were pre-infested with 20 second- to third-instar nymphs. Cages were placed in the center of each cotton plot spaced 1 m from each other for 24 h the day before insecticide applications to treated plots (before spray commenced). We repeated the deployment of newly prepared infested plants that were in open and closed cages 7 days after insecticide application (7DAA) and removed them after 24 h. Plants were collected, and the remaining nymphs were counted. Using the methodology of Naranjo and Ellsworth (2017), consumption was defined as partially eaten nymphs, dried/desiccated nymphs (sucking insect feeding), or missing nymphs, which became the response variable, whitefly nymphs remaining (nymphs remaining on the leaves of the plants from the initial 20 that were there; missing prey was assumed to be consumed).

Statistical Analysis

All analyses and graphical summaries were conducted in R (RCoreTeam 2023). We estimated Kaplan–Meir survivorship curves for G. punctipes and H. convergens using the package {survival} on right censored data because not all predators were dead at the end of the experiment (Therneau and Granbsch 2000, Therneau 2023). Within the {survivial} package, the Mantel–Haenszel test (function survdiff), a log-rank test, was used to compare survival curves for each predator in relation to insecticides (Harring and Fleming 1982). From the same data set where the predators were provided prey, using all predators that survived, we analyzed predation responses (whitefly nymphs consumed in 24 h) in relation to insecticides, predator taxa, and the interaction using ANOVA on square-root transformed (sqrt x +1) whitefly nymph predation. To assess changes in predator foraging on whitefly nymphs when alternative prey was available in the same proportions, we used the proportion of whitefly nymphs consumed out of total prey available as the response in relation to predator species, insecticide treatment, and interaction with a generalized linear model (glm) using the glm function and specifying the distribution family as a binomial (link= “logit”). To estimate the effects of insecticide-treated areas on foraging (chamber condition: insecticide-treated or nontreated) on prey removal by predators, we used linear mixed-effects models (LMEs) for each predator on total prey, whitefly nymph, or aphid nymph removal (predation). Here, log-transformed prey removal (natural log x+1) as the response variable for predation in relation to the application (one chamber treated and the other chamber untreated), insecticide treatment, and interaction because both chambers were connected. We used LMEs given the experimental arena with both chambers connected was an experimental unit, so the chamber was coded as a random effect, and models were fit using {nlme} function “lme” (Pinheiro and Bates 2023) with a Gaussian distribution in R language family= “gaussian.”

Finally, we again fit an LME to assess the effects of field insecticide treatments on prey removal from caged (closed) and uncaged (open) sentinel plants. Because natural enemies would have access to the whitefly nymphs in the open plants, we were interested in the difference between opened and closed potted plants in relation to insecticides and pre- and post-application of insecticides. Because the experiment was carried out at 2 locations, we coded the LME with location as a random effect and used the model section to reduce the complexity of models (i.e., ∆AIC, Burham and Anderson 2002). Second, the split-plot of traits was removed because there were no differences between traits for whitefly nymph removal from sentinel plants (F_1,140 = 0.5917, P = 0.4431). Model selection was then used for this experiment because there were 3 possible interactions: cage condition (open or closed cages) × pesticide treatment, the timing of sampling (pre- and post-application) × treatment and condition × application, which would indicate possible interdependencies in determining predation-based on design elements. However, interactions also create overly complex models, so we used model selection to simplify the analysis and report the best-fitting model explaining variation in whitefly nymph removed from experimental plots. For all analyses, we inspected the data for the distribution and variance structure to make decisions on the analysis approach. Where needed, appropriate transformations were used to improve the fit of models to meet model assumptions (i.e., there was no pattern in residuals of fitted models or evidence of heterogeneity of variance of LMEs). Any significant main effects or interactions were followed with linear contrasts using the package {emmeans} (Lenth 2023), and graphical displays were created using the package {ggplot2} (Wickham 2016).

Results

Direct Effects: Natural Enemy Survival Post-exposure to Insecticide-treated Prey

The laboratory evaluations showed that the tested predators responded differently to prey exposed to insecticide treatments. Insecticide exposure did not significantly affect the survivorship of G. punctipes over the 7 days (χ² = 0.10, df = 2, P = 1.000; Fig. 3A), while the survivorship of H. convergens was significantly reduced (χ² = 13.30, df = 2, P = 0.001; Fig. 3B). Geocoris punctipes survival was not significantly impacted after being exposed to pyriproxyfen or cyantraniliprole-treated prey, as compared to the control group. Conversely, H. convergens exhibited a significant decrease in survivorship when exposed to cyantraniliprole, with less than 50% of H. convergens surviving until day 4 and by day 5, 100% mortality (Fig. 3B). In addition, pyriproxyfen exposure caused a significant reduction in survivorship of H. convergens compared to the control group, with less than 40% surviving by day 7 (Fig. 3B).

Indirect Effects: Natural Enemy Foraging Responses to Insecticide-treated Prey

The results indicate predator-specific differences of post-insecticide exposure on predation (Fig. 4A). Predators differed in their predation on whitefly nymphs (F_1,38 = 5.19, P = 0.028), with no significant pesticide treatment effect (F_2,38 = 0.82, P = 0.447), and a significant interaction between predator species and insecticide exposure treatment (F_2,38 = 9.05, P < 0.001; Fig. 4A). The significant interaction indicates that the effects of insecticides on whitefly nymphs foraging were predator-dependent and explained by a reduction in predation by H. convergens when exposed to prey treated with cyantraniliprole (Fig. 4A). In a second experiment, we provided a prey choice of whitefly nymphs and aphids nymphs that were treated in the same way (Fig. 4B). Focusing on whitefly nymphs, we analyzed the proportion of whitefly nymphs consumed when both whitefly nymphs and an alternative prey, aphid nymphs, were provided and exposed or not exposed to insecticides. In this case, there were no significant differences in the proportion of whitefly nymphs consumed between the 2 predators (χ² = 0.03, df = 1, P = 0.855). Insecticide exposure had significant effects (χ² = 16.78, df = 2, P < 0.001) and a non-significant interaction (χ² = 0.03, df = 2, P = 0.867), indicating that the effects are independent of predator or exposure treatment (Fig. 4B). Therefore, both predators consumed relatively equal proportions of whitefly nymphs in the pyriproxyfen and control treatment, with a significantly lower proportion consumed when prey were treated to cyantraniliprole (Fig. 4B).

Indirect Effects: Predator Foraging Location

To simulate a situation where predators could actively choose between insecticide-treated plants and untreated, 2 chambers containing plants with prey (whitefly nymphs and aphid nymphs) were connected by a central corridor for predators to move between plants with treated prey (insecticides or water). We tested for predation differences when predators had the choice to move and forage on treated and untreated prey. Geocoris punctipes consumed significantly more total prey (fewer prey remaining) in treated plant chambers than untreated (F_1,26 = 5.45, P = 0.028; Fig. 5A), and no significant insecticide treatment effect (F_2,26 = 0.02, P = 0.979; Fig. 5A). While G. punctipes generally consumed fewer whitefly nymphs in the untreated plant chamber (F_1,26 = 5.59, P = 0.026; Fig. 5B), and higher predation on whitefly nymphs was observed in the pyriproxyfen treatment (F_2,26 = 3.20, P = 0.057), aphid predation did not vary significantly by chamber insecticide exposure or chamber condition (F_2,26 = 1.84, P = 0.179, F_1,26 = 1.77, P = 0.195, respectively; Fig. 5C). For H. convergens, the results show no significant effects of insecticides or chamber exposure for total prey consumed (F_2,26 = 2.00, P = 0.155, F_1,26 = 0.004, P = 0.947; Fig. 5). However, for whitefly nymph consumption by H. convergens, the fewest whitefly nymphs remained when exposed to plant chambers treated on one side with pyriproxyfen as compared to plant chambers of water only (control) or cyantraniliprole (F_2,26 = 5.00, P = 0.015; Fig. 5B). Aphid predation by H. convergens was not influenced by chamber condition or insecticides (F_1,26 = 0.84, P = 0.369, F_2,26 = 0.18, P = 0.833, respectively; Fig. 5C).

Field Study: Effects of Insecticide Applications on Whitefly Nymph Removal

In the field, we deployed sentinel plants with a standardized number of 20 whitefly nymphs to replicated field plots exposed to water, pyriproxyfen, or cyantraniliprole. From sentinel plants, we estimated the effect of the insecticides on whitefly nymph removal and compared between open and closed cages. A simple model containing only the main effects (cage, application timing, and insecticide treatment) without any interactions was the best-fitting model (∆AIC > 2.00 for any model containing an interaction, and the difference in AIC between simple main effects and full interaction model was > 10 AIC). Therefore, we retained and interpreted the main effects, which suggest that either before insecticide application or 7 days after insecticide application, significantly fewer whitefly nymphs remained on open plants (no cage) as compared to the closed cages (F_1,139 = 157.09, P < 0.001; Fig. 6). In addition, more whitefly nymphs were present on open cages that were deployed following insecticide application as compared to before (F_1,139 = 6.88, P = 0.009; Fig. 6), which potentially indicates a slight reduction in biological control of the whiteflies following application of selective insecticides into the experimental plot area. The different insecticide treatments had no significant effect on whitefly nymphs remaining on sentinel plants (F_2,139 = 1.18, P = 0.308).

Discussion

Developing a successful IPM system requires studies that reveal the trade-offs and potential incompatibilities among its programs and individual parts. Therefore, it is essential to study the compatibility of chemical insecticides used in IPM with biocontrol efforts to guide pest management tactics for target pests (Torres and Bueno 2018). Such compatibility requirements go beyond the direct impacts of insecticides on predator survivorship and involve indirect effects on predator activities (Schmidt-Jeffris 2023). Our laboratory bioassays provide evidence of the direct and indirect impacts of selective insecticides on 2 generalist predators. Like previous studies (Naranjo et al. 2004, Barros et al. 2018, Machado et al. 2019), our evaluation showed minimal impact of pyriproxyfen on predators. Specifically, pyriproxyfen, as compared to cyantraniliprole, had a low to moderate effect on survivorship and foraging of G. punctipes and H. convergens. Conversely, Cyantraniliprole appears to have both direct and indirect effects on G. punctipes and H. convergens. Our study revealed that H. convergens consuming cyantraniliprole-exposed prey reduced survivorship compared to water- or pyriproxyfen-exposed prey. In addition, an indirect effect was observed where H. convergens consumed lower numbers of cyantraniliprole-exposed whitefly nymphs, and both predators consumed a lower proportion of whitefly nymphs when offered an alternative prey choice in cyantraniliprole-exposed prey. However, in the field, sentinel plants deployed in insecticide-treated plots suggest that predators are still active when exposed to either pyriproxyfen or cyantraniliprole treatments and provide whitefly biocontrol by reducing the number of whitefly nymphs.

Direct and indirect exposure to cyantraniliprole, through contact with sprayed leaves and feeding on treated prey, significantly impacted predators by reducing survival for H. convergens and reducing H. convergens and G. punctipes consumption of prey. Cyantraniliprole acts on ryanodine receptors within muscle cells to deplete stored calcium. It disrupts muscle contraction, ultimately leading to paralysis and the eventual demise of targeted insects. Widely recommended for controlling various pests such as lepidopterans, beetles, whiteflies, leaf miners, and thrips (Selby et al. 2013, Barry et al. 2015), cyantraniliprole shares target receptors with both pest insects and their natural enemies. However, despite this overlap, several studies indicate a low risk of adverse effects on natural enemies when exposed to cyantraniliprole (Mandel 2012, Misra 2012, Patel et al. 2012). Bojan (2021) reported that a high dose of cyantraniliprole had considerably fewer toxic effects on the natural enemies, especially on predatory coccinellids. The negative effect of cyantraniliprole on a species of predatory Coccinellidae, revealed in our study, corroborates Jiang et al. (2020), showing that cyantraniliprole exposure caused reduced adult longevity and a severe reduction in the predation potential of Coccinella septempunctata L. (Coleoptera: Coccinellidae) (Amarasekare et al. 2016, Beers et al. 2016, Mills et al. 2016). For example, our results show the lowest predation on whitefly nymphs when H. convergens was exposed to cyantraniliprole treated prey, suggesting an indirect effect on the foraging rate. Furthermore, we observed altered preference for prey by both G. punctipes and H. convergens, as the proportion of whitefly nymphs consumed was lower when prey was exposed to cyantraniliprole. The exact cause for the altered prey preference of whitefly nymphs treated with cyantraniliprole is unknown, it could be due to the predator’s sensory detection capacity. Both predators may be able to detect the nutritional quality and avoid contaminated or treated prey (Silva et al. 2020, 2021). Here, our results show the direct and indirect effects of cyantraniliprole on the biological control agents of whiteflies.

Conversely, given the mode of action of pyriproxyfen, which targets and disrupts insect growth during the immature stages, there were negligible effects of pyriproxyfen on G. punctipes and H. convergens adults. In previous studies, negative impacts following exposure to this insecticide have been identified in the egg and immature stages for other coccinellids and hemipteran species (Liu and Stansly 2004, Li et al. 2015). In this study, pyriproxyfen exposure did not significantly impact predator survivorship or foraging behavior, highlighting its compatibility with adult predators. However, our evaluations were confined to the recommended label rate of this insecticide. It is possible that sublethal doses of pyriproxyfen could further alter the behaviors seen here (Iftikhar et al. 2020). Iftikhar et al. (2020) reported longer pre-adult developmental time and reduced adult H. convergens fecundity and longevity, highlighting the indirect effect of pyriproxyfen sublethal doses on this lady beetle population. Similarly, Noelia et al. (2023) reported high mortality of eggs and larvae of the coccinellid Eriopis connexa (Germar) in response to pyriproxyfen application, indicating direct adverse effects. In contrast, previous studies indicated the compatibility of pyriproxyfen with immature stages of Chrysoperla carnea (Stephens) (Neuroptera: Chrysopidae) (Mansoor and Shad 2020). Therefore, to draw a full picture of the compatibility of pyriproxyfen with whitefly biocontrol efforts, future work is needed to determine their potential adverse effect on the immature stages of these generalist predators. Currently, our study suggests that pyriproxyfen is compatible with adult H. convergens and G. punctipes.

Behavioral avoidance of insecticides is desirable in natural enemies because it reduces exposure and can increase the likelihood of survival in field conditions (Haynes 1988, Desneux et al. 2007, Cordeiro et al. 2010, Campos et al. 2011). Studies have emphasized the adaptive nature of beneficial insects in detecting and responding to insecticides within their surroundings, showcasing altered search behaviors to evade treated areas (Perera 1982, Bos and Masson 1983, Hoy and Dahlsten 1984, Desneux et al. 2005, Thornham et al. 2007). However, in our connected choice chamber experiments with treated and untreated plant chambers, Geocoris punctipes appeared to actually forage at higher levels in treated plant chambers (Fig. 5), and H. convergens did not appear to change foraging overall on prey. Contrary to our predictions, both predators consumed more whitefly nymphs in plant chambers exposed to pyriproxyfen. This increase in consumption may be due to prey immobility since capturing them should be easier than capturing the mobile prey (aphid nymphs). However, our field experiments revealed no significant difference between the number of whitefly nymphs consumed in control and plots exposed to either pyriproxyfen or cyantraniliprole treatments. In contrast, our laboratory experiments indicate a significant impact of cyantraniliprole on the survival of H. convergens and prey foraging by both natural enemies. One explanation for this disparity is the confinement of predators within small-scale laboratory bioassay experiments. In such settings, where predators are confined, they may experience heightened exposure, whereas predators in the field can hide from insecticides or avoid contact with insecticides. Alternatively, in field conditions, the spillover of natural enemies into sprayed plots from adjacent habitats could be another potential explanation, which masked the effect we saw in the laboratory trial. What we did see in the field was a general reduction in whitefly nymph removal from sentinel plants following insecticide application to field plots, which suggests that all treatments were affected by the use of insecticides within the experimental area. In this particular field experiment, we deployed sentinel plants with known whitefly nymphs and measured whitefly nymph removal. Our lack of differences could be because there were few predators in the area or whitefly nymphs were removed at random. However, the discrepancy between laboratory and field results emphasizes the importance of pairing studies to provide detailed and controlled data from laboratory bioassays and field evaluations.

Our study shows the susceptibility of natural enemies to direct and indirect exposure to insecticide-treated prey and insecticide residue on plant leaves. Chewing predators such as Coccinellidae, which consume the entire prey, face potential exposure to any insecticide persisting in or on the prey, including the gut. On the other hand, hemipteran piercing-sucking predators drink their prey and, therefore, are likely indirectly exposed at comparatively lower levels (Gentz et al. 2010). The natural enemies tested here displayed variable survival rates when exposed at the adult stage to insecticide-treated nymphs and dried residues of different insecticides recommended for whitefly control. Notably, while cyantraniliprole did not cause significant mortality in G. punctipes, it led to a reduction in prey consumption by adults. Results from the laboratory may not capture the effects occurring in more complex field conditions. Bioassays create closed systems where animals are unable to emigrate, and in field conditions, predators may be exposed to multiple pesticides. In addition, prior rearing conditions of predators might play a significant role in their reaction to insecticide (Naranjo et al. 2004, Naranjo and Ellsworth 2009b). Nevertheless, results from our field study suggest the different insecticide treatments had no significant differences in the number of whitefly nymphs consumed between caged and open-potted plants, challenging the laboratory results and emphasizing the complexity of translating controlled experiments to real-world field conditions. Consequently, further research is required to understand the interactive effects of insecticides on biological control communities. Our current data, with combined field and laboratory results, indicate that pyriproxyfen does not appear to have negative direct or indirect effects on G. punctipes or H. convergens.

Acknowledgments

We thank Melissa Thompson for her technical support, and Aspen Foster and Lily Sheffield for assisting with predator scouting and sample collection. We thank Calvin D. Perry and Bobby J. Washington for their assistance in the field experimental setup in Camilla, GA, and Charles L. Gruver and associate personnel for assistance in the field experiments in Tifton, GA. We also thank The University of Georgia for providing laboratory and office spaces. This research was funded by USDA-ARS Non-Assistance Cooperative Agreement #58-6080-9-006. This study reports the results of the research only. The mention of a proprietary product does not constitute a recommendation by the USDA or UGA.

Author contributions

Albertha Parkins (Conceptualization [equal], Data curation [lead], Formal analysis [supporting], Investigation [lead], Methodology [equal], Writing—original draft [lead], Writing—review & editing [equal]), Arash Kheirodin (Conceptualization [equal], Investigation [equal], Supervision [lead], Writing—review & editing [equal]), Jermaine Perier (Writing—review & editing [equal]), Paulo Cremonez (Writing—review & editing [supporting]), David Riley (Funding acquisition [equal], Writing—review & editing [equal]), Alvin Simmons (Funding acquisition [equal], Writing—review & editing [supporting]), and Jason Schmidt (Data curation [equal], Formal analysis [equal], Funding acquisition [equal], Project administration [equal], Resources [equal], Writing—review & editing [equal])

References

Author notes