Mango (Mangifera indica L.; Anacardiaceae) is one of the most widely grown and exported tropical fruits in Costa Rica, with about 53,288 metric tons produced annually ( http://www.fao.org/faostat/en/#data/QC). One of the major constraints affecting mango production in Central America is the constant presence of thrips on mango blossoms and young leaves. The dominant species present on blossoms is Frankliniella invasor Sakimura whereas young leaves are mostly affected by Scirtothrips citri (Moulton) (both Thysanoptera: Thripidae) (Mound & Hoddle 2016). It has been reported that a single mango inflorescence may contain more than 600 adults plus the immature stages of F. invasor (Rocha et al. 2012), and may lead to significant yield reduction, as observed by Gehrke (2008). Frankliniella invasor is native to the Caribbean–Central American region and was reported for the first time feeding on mango in Hawaii (Sakimura 1972).

Regular use of insecticides, especially the organophosphate malathion to control pestiferous thrips populations on flower inflorescences has not yielded much success for farmers. In fact, malathion has been proven to be an insecticide with low efficacy for some thrips species in Central America (Walter et al. 2018), even though it is known as an efficient compound against the western flower thrips, Frankliniella occidentalis Pergande (Thysanoptera: Thripidae) (Helyer & Brobyn 1992). We report here on laboratory toxicity studies of chlorfenapyr, spinetoram, chlorpyrifos, and thiamethoxam against F. invasor populations on mango in Costa Rica. This assay was conducted as a first step towards the development of an integrated pest management program for thrips management in mango in Costa Rica and Central America.

Samples of mango inflorescences infested with thrips were collected from the field during late Jan and early Feb of 2017 (10.6376190°N, 85.5213268°W; 25 masl). Samples were placed in a labeled plastic bag and transported to the laboratory. Ten adult F. invasor were aspirated into glass tubes (6 mm diam) from the samples and emptied into individual 35 mL plastic diet cups (Fill-Rite Corporation, Newark, New Jersey, USA) with uncoated paper tops.

Insecticides were purchased from local agrochemical stores for bioassays (Table 1). Eight concentrations of each insecticide were prepared using distilled water at 0, 128, 0, 64, 32, 16, 80, 400, 2,000, and 10,000 ppm. Fresh snap bean pods (Phaseolus vulgaris L.; Fabaceae) were rinsed with 1% bleach then flushed thoroughly with tap water and allowed to air dry on paper towels. Beans were cut into 20-mm long sections and the excised pods were sealed at either end with a thin layer of paraffin (Fisherfinest™ Histoplast Paraffin Wax, Fisher Scientific, Waltham, Massachusetts, USA). After bean pods were sealed, they were immersed in the different insecticide concentrations for 5 min in 50 mL glass containers then air-dried on paper towels for about 15 min. For the control, bean pods were immersed only in distilled water. Pods previously treated with an insecticide were placed individually inside 35 mL transparent plastic cups (3 cm diam) with uncoated paper tops. Ten adults of F. invasor were placed inside each cup. Each treatment consisted of 10 replicates plus a control. Individual cups were placed into a sealed plastic rearing container (5.7 L) with the bottom lined with a paper towel to reduce condensation. Containers were kept in a controlled-environment chamber maintained at 26 ± 2 °C, 60% RH, and a photoperiod of 16:8 h (L:D). Thrips mortality was evaluated after 24 h and were considered dead when they were unable to move after being probed for a couple of s.

Table 1.

Insecticides used in laboratory bioassays to determine the LC₅₀ and LC₉₅ estimates of Franklinellia invasor toxicity of treated bean pods (Phaseolus vulgaris L.) after 24 h exposure previously collected from mango.

Fig. 1.

Logistic mortality curves for the thrips Franklinellia invasor for 8 selected insecticides with the dip method. (A) Chlorpyrifos, (B) malathion, (C) α-cypermethrin, (D) imidacloprid, (E) thiamethoxam, (F) spinetoram, (G) spinosad, (H) chlorfenapyr.

Table 2.

The LC₅₀ and LC₉₅ estimates for 8 selected insecticides after 24 h exposure to control Franklinellia invasor previously collected from mango. CI = confidence interval.

Following this first trial, a second toxicology assay was conducted where concentration ranges were chosen to have at least 5 data points between 20 to 80% mortality in order to create a reliable concentration response curve (Yu 2015). Concentration ranges of the second assay varied between 0.05 and 0.64 µg per mL for spinetoram and between 0.18 to 400 µg per mL for malathion.

The LC₅₀ and LC₉₅ estimates with associated 95% confidence intervals were determined for data from the second toxicological assay using the probit model (Finney 1971). Concentration-mortality data were then modeled using the LOGISTIC option in PROC PROBIT (SAS Institute 2011) to determine the intercept and slope. The P-values for the goodness-of-fit tests (Pearson chi-square) were used to indicate an adequate fit of the log concentration-response regression line. Treatment mortality was automatically corrected for that in controls by the PROBIT software. Logistic mortality curves for each insecticide were generated with SigmaPlot version 14 (SYSTAT, San Jose, California, USA).

Our results indicated that spinosad, spinetoram, chlorpyrifos, α-cypermethrin, and chlorfenapyr were highly effective, while malathion and imidaclopid were weakly active to control F. invasor (Fig. 1; Table 2). Infante et al. (2014) reported similar results for spinosad, malathion, imidacloprid, and α-cypermethrin to control F. invasor Mexican populations. However, in our study the neonicotinoid thiomethoxam showed a higher level of toxicity compared with imidacloprid. Unfortunately, there are a number of limited rotational choices with respect to the modes of action available for insecticidal control of F. invasor. Because of this situation, there is a risk that insecticide resistance will increase in this species (Bielza 2008). We believe that additional investigations to identify potential insecticides to control F. invasor should be continued for managing this pest as well as exploration of biological control agents.

We thank Joaquín Calderón of the Natural Science Laboratory of Earth University for helping with the data collection in the field. Funding for the project was provided by Mangarica Costa Rica.