Materials and Methods

Field surveys were conducted in seven locations around Tift Co., GA, and the surrounding area from June to August 2020–2022, given the region’s prominent vegetable and cotton production. Selected sites were obtained with the help of the University of Georgia Extension agents and were scouted at least 24-h before sampling to ensure the presence of sufficient whitefly (B. tabaci MEAM1) numbers for bioassays. Sampling was conducted three times per site to collect dose–response data on B. tabaci adults through laboratory bioassays. Sampled sites varied in production size, but mostly represented vegetable crops (see Table 1 for site information).

Table 1. Surveyed Bemisia tabaci populations from Tift Co., GA, USA, and surrounding areas subjected to the modified maximum dose bioassay.

View Fullscreen

Untreated cotton seedlings (Gossypium hirsutum L., Stoneville® ST 4946-GLB2) were used as the standard host for all bioassays conducted, following other reports (Caballero et al. 2013b, Perier et al. 2023a, Schuster et al. 2010, Sparks et al. 2020). The cotton plants selected for this methodology were grown under 30 ± 2°C, 60 ± 5% relative humidity, and a photoperiod of 14:10 (L:D) h. The selection process involved washing and clipping the roots to a length of 5 cm. These prepared cotton plants were then inserted into the respective treatments as part of the experimental procedure following the steps outlined by Perier et al. (2023a).

A colony of B. tabaci cryptic species MEAM1 (LAB-1) maintained in a pesticide-free environment for at least 4 yr on rotations of squash (Cucurbita pepo L. subsp. pepo var. Golden Summer Crookneck) and cotton was used as the colony comparison for insecticide evaluations. The relative susceptibility of the colony (LAB-1) served as a reference for the expected susceptibility of the field populations to the tested insecticide (Caballero et al. 2013b).

Adult whiteflies were collected from all populations by using a funnel apparatus (Fig. 1) and stored in screened transparent plastic tubes (diameter: 2.86 cm; length: 20.3 cm, ClearTec® Packaging, Park Hill, MO) for the bioassay. Thirty-six tubes were collected per site per sample, with each containing at least 50 whiteflies. During transport to the laboratory, the samples were briefly insulated and cooled with ice packs to mitigate mortality due to transportation from the field; the travel time averaged 30 min from each collection site. As such, whiteflies, on average, were bioassayed within 1.5-h after field collection, including a 1-h acclimation period. Adult age and sex were not determined for this study, nor were they included as factors in any analysis.

View Figure

• Download as Powerpoint
• Download Figure

Bioassays were conducted at the Coastal Plain Experiment Station in Tifton, GA. Ambient conditions for the bioassays were 27 ± 2°C, 50% relative humidity, and a photoperiod of 24:0 (L:D) h for both the treatment period and the bioassay. Several populations of B. tabaci (Table 1) were subjected to a high dose of six insecticides: imidacloprid, dinotefuran, thiamethoxam, flupyradifurone (Insecticide Resistance Action Committee [IRAC] group 4), and cyantraniliprole and cyclaniliprole (IRAC group 28), in a dose–response bioassay. An additional “low” dose of each insecticide, one tenth of the high dose, was also tested (Table 2).

Table 2. Evaluated insecticides against Bemisia tabaci populations from Tift Co., GA, USA, and surrounding areas.

View Fullscreen

Stock solutions made up to 500 ml for each insecticide were used following previously published protocols (Perier et al. 2023a) and only modified for the single additional dose. This “low” dose was created by diluting 50 ml of the insecticide stock with 450 ml of water. The selected cotton plants were 3 weeks old (with at least one terminal true leaf with a 4-cm width) at the time of use. Similarly to the bioassay steps outlined by Perier et al. (2023a), only the terminal true leaf was used in this study after a 24-h root drench of the plant in the treatments and check (untreated control) at experimental conditions. Treated leaves were used to bioassay adult whiteflies and were placed inside the collection tubes toward a benchtop light. Treatments and check concentrations were replicated four times per bioassay, with mortality recorded immediately after leaf insertion into the tubes and again after 24-h.

Mortality data were corrected using Abbott’s formula (Abbott 1925) for consistency with the checks and treatments across all locations. An additional threshold of 40% was established for mortality in the check. This threshold was used to remove experimental sets with mortality in the check above the threshold. The final dataset subjected to analysis was percentage data and therefore had a binomial distribution. To ensure normality, the data were transformed (log transformed) and Gaussian error distribution was confirmed using residual and normality plots (Fernandez 1992) in SAS Enterprise Guide v. 8.3 (SAS Institute Inc., Cary, NC). Proc GLM was then used to determine the chemical, dosage, and population differences of the transformed data, whereas pairwise associations and differences were categorized using the LSMEANS statement and LINES option (P = 0.05). The separated responses were then ranked according to their effectiveness at managing B. tabaci. Overall comparisons involved analysis of either the pooled populations or pooled doses.

Results and Discussion

The differences in B. tabaci mortality were a result of treatment (F_5,1200 = 31.79; P < 0.0001), dose (F_2,1200 = 71.14; P < 0.0001), and population (F_7,1200 = 102.44, P < 0.0001). Significant interactions were only seen in parings including dose, mainly dose and chemical (F_5,1200 = 4.61; P < 0.001) and dose and population (F_14,1200 = 3.76; P < 0.0001). When all three interacted, there was no significant impact on mortality, nor was there an interaction between chemicals and population. When mortality was pooled for each population, all treatments were significantly separated from the check, untreated control (Fig. 2). Similarly, treatments were separated regardless of dose (high and low). They were then ranked according to their efficacy (Fig. 2). At this level, dinotefuran, cyantraniliprole, and flupyradifurone caused the highest mortality on B. tabaci, followed by thiamethoxam, with imidacloprid and cyclaniliprole offering the least contribution in the Tift Co. and surrounding area (Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

Regarding insecticide dose, only two insecticides, dinotefuran and thiamethoxam, showed dose-related different responses (high or low dose). For the other insecticides, the doses were comparable at P < 0.05 (Fig. 2). However, clearer separation could be seen for each chemical when the population was considered an effect at P < 0.05 (Fig. 3). Separation was observed for several insecticides at different locations for a specific dose (Fig. 3), while at the same time there was a lack of separation among the doses for the other chemicals when population was not considered (Fig. 2), illustrating the need for site specific data. Certain populations exhibited higher mortality at the lower dose than at the high doses (e.g., TF-1, Figs. 3E). Interestingly, there were cases where mortality between the doses was unchanged and had no separation (e.g., CH-1, Figs. 3C); as such, more evaluations would be required to determine the insecticide response of these locations to these insecticides (LC₅₀ bioassays). Or, perhaps, the lack of dose separation in these populations aids in characterizing the overall efficacy of these insecticides. Nevertheless, the trends identified in the lower dose, that is, good or poor efficacy, were also evident in the high dose, as seen with the Ponder population and many of the insecticides (Fig. 3).

View Figure

• Download as Powerpoint
• Download Figure

The 24-h maximum dose bioassay used in this study revealed higher mortality rates in several populations of B. tabaci when exposed to the high dose of the tested insecticides. Overall, cyantraniliprole was highly effective in the Georgia B. tabaci populations, resulting in an average of 88.7% mortality across all field populations at the high dose. Notably, at the low dose, the average mortality was 86% for the same insecticide, indicating substantial control. Therefore, using the proposed low-dose estimates the same level of adequate control of B. tabaci for cyantraniliprole. In the LT, Gibbs, and Lang-Rigdon populations, a higher mortality rate was obtained at the low dose of cyantraniliprole, whereas in Chula and Ponder, susceptibility to the insecticide was already evident at the low dose, highlighting the potential benefit of dose characterization in individual populations. The general efficacy of dinotefuran and flupyradifurone, compared with the other two neonicotinoid insecticides tested at the high dose, was consistent with previous reports from Florida, USA (De Marchi et al. 2021, Smith et al. 2016). In these Georgia B. tabaci populations, dinotefuran exhibited higher efficacy, resulting in high mortality at the high dose. The efficacy of imidacloprid and thiamethoxam in these same populations was poor, warranting further resistance management. Similarly, less effective control was obtained with cyclaniliprole, as the results revealed low B. tabaci mortality following exposure. In imidacloprid, poor efficacy was evident at the low dose except for the Tifton and Ponder populations. This was expected, given the potential for fast upregulation of metabolic mechanisms in response to imidacloprid exposure (Karunker et al. 2008, Perier 2023). However, the shifts in mortality observed in these populations due to exposure to the low dose of imidacloprid hint toward the existence of sublethal effects. Some biological processes, such as fecundity and other behavioral changes, could be disrupted (He et al. 2011, Nauen et al. 1998, Sohrabi et al. 2011, Wang et al. 2016). However, sublethal effects were beyond the scope of this study, but could provide insight for improved resistance management as seen with other neonicotinoids (Wang et al. 2016).

Bemisia tabaci mortality was surprisingly different in response to the two diamides tested in this study. Although cyantraniliprole exposure resulted in higher B. tabaci mortality, exposure to cyclaniliprole (another relatively new anthranilic diamide with a novel mode of action; Tsukamoto et al. 2021) yielded an insufficient B. tabaci control in this study. In other locations, cyclaniliprole continues to offer control with low levels of resistance (Gill and Chong 2021, Guo et al. 2020). As such, further evaluation is needed to confirm whether cyclaniliprole has poor uptake in plants from root drench applications, as translocation throughout the plant would be limited. This is notable because differences in response for nymphs were also seen when sprays and drenches were previously compared (Gill and Chong 2021). This could explain the poor control observed in this study from cyclaniliprole, given the root drench insecticide treatment method. Nevertheless, cyclaniliprole separated from the check and offered some levels of B. tabaci control.

Chula and Ponder, on average, were the most susceptible populations tested against the insecticides. By contrast, the laboratory colony LAB-GA was the most resistant to these insecticides, possibly because the original colony was created from resistant populations at a time when heavy B. tabaci infestation plagued Tift Co. and its neighboring areas. With the lack of interfering field environmental factors, such as genetic diversity from migrating populations, in a controlled ecosystem, such as colony rearing conditions, an established case of resistance could be sustained by haplodiploidy (Denholm et al. 1998). Therefore, certain resistance traits may persist without contention in a colony, allowing more susceptible populations to exist in field populations. However, the identification of these populations would require more than just a “maximum dose” testing and warrant future efforts.

In this study, we propose a rapid bioassay that produces insecticide response data within a 48-h time frame (the approach involves 24-h systemic treatment of the plants followed by a 24-h experimental period). Although insecticide activity can persist past this period, other factors may interfere with the recorded results. In this study, leaf desiccation and whitefly feeding behavioral changes were considered when standardizing to the 24-h experiment period once the leaf was detached. Mainly, these factors could contribute to mortality with a longer experiment duration. Not to mention the potential degradation of the applied insecticide. Because B. tabaci is a systemic feeder, a systemic approach was the goal of this study. However, this methodology targets insecticides that offer quick responses on adults as well as nymphs, not necessarily growth regulators, which require longer experimental times. In addition, the tested insecticide should have some systemic nature that allows for translocation from the roots. There is a possibility of differences between this systemic approach and other insecticide evaluation methodologies that tend to offer complete contact coverage of the treatment plant or direct application to the insect (Yu 2014). However, earlier comparisons found no such differences, with mortality only being due to the treatments, regardless of the methodology (Sparks et al. 2020). Plant uptake and retention of the chemicals being tested can also be estimated for greater accuracy, thereby correlating response to the treatment concentrations applied (Perier et al. 2023a). Therefore, if the molecular chemistry of the insecticide allows, it can be tested with this methodology and more accurately. The low dose is proposed as a companion to the high dose and is meant to provide more information on complex populations that may benefit from more long-term control options using sublethal approaches. The application of the low dose identified many areas of concern during this study. Initially, there was a need for field characterization, as insecticide response was impacted by the geographical location of the population. Moreover, the application rate of these insecticides at the high dose control was sufficient for most. However, the low dose identified populations that could be actively selected for resistant populations due to reduced mortality at the higher dose. Also, the low dose provides a clearer starting point for future LC₅₀ evaluations that would be based on the high dose and require multiple adjustments of serial doses.

Including a low dose alongside a high dose (=maximum labeled rate) bioassay is a crucial enhancement to the original maximum dose bioassay proposed by De Marchi et al. (2021). Incorporating the low dose creates a more comprehensive understanding of the dose–response relationship, because the high dose alone provides only a single data point on the hypothetical dose–response curve of a given population to a specific compound. This addition improves the bioassay and aligns with the need for faster assay development, because it requires only two different dosages instead of a larger set of concentrations (Perier et al. 2023b, Yu 2014). This reduction in doses saves time, reduces the sample size, and minimizes the material required for the laboratory bioassay test. Therefore, the modified high–low dose bioassay proposed in this study provides a quick, easy to setup tool for integration into insecticide resistance management strategies. Further studies regarding model validity for precision and accuracy comparing the laboratory bioassays with real-world in-field responses are needed (Cremonez et al. 2023a, 2023b).

This study evaluated commonly used insecticides for B. tabaci management. Greater control was found with the novel neonicotinoids dinotefuran and flupyradifurone than with imidacloprid and thiamethoxam. As such, these two novel neonicotinoids should be considered as alternatives. With the diamides, only cyantraniliprole offered substantial control. The use of the low dose in this study identified field trends that were different from those displayed at the high dose. As such, the impact of population should always be considered during these evaluations. The addition of the low dose provided more insight into the insecticide response of each population and could be used for evaluations prioritizing insecticides for use or rotation.