Sublethal effects of imidacloprid on the whitefly parasitoid Encarsia formosa Gahan

SUMMARY Acute toxicity of an imidacloprid-based product (Confidor 200 SL) to pupae of the whitefly parasitoid Encarsia formosa Gahan (Hymenoptera: Aphelinidae), and its effects on life history traits and population growth in F 1 generation of the surviving parasitoid females of a commercial strain (“Dutch” strain, D) and two local populations from Serbia (Bujanovac, B; Negotin, N) were examined in laboratory bioassays. All trials were carried out at 27±1°C temperature and 60±10% relative humidity, and under 16/8 h daylight/darkness photoperiod in four replications. In acute toxicity bioassays, tobacco leaves carrying parasitoid pupae were treated with a series of symmetrical concentrations (800, 400, 200, 100, 50 and 25 mg a.i./l) covering a 10-90% mortality range. The product based on imidacloprid, applied directly onto parasitoid pupae at mean lethal concentrations (LC 50 ) determined in the acute toxicity assays (30 mg/l, 20 mg/l and 25 mg/l, for populations B, N and D, respectively), significantly affected the survival of females developed from the treated pupae, extended the duration of juvenile development (by 1.81, 1.59 and 1.73 days for populations B, N and D, respectively), significantly reduced total parasitism of females D (25.92 %), total female adult emergence in populations B (27.48 %) and D (17.92 %), and significantly reduced the instantaneous rate of increased only of females N (4.23 %). Considering the high acute toxicity of the imidacloprid product to the pupal stage of E. formosa , and significant reductions in life table and population parameters, imidacloprid is not considered compatible for simultaneous use with the parasitoid E. formosa . A more precise assessment of risks involved in the use of that insecticide requires a more detailed testing in the field. The implications of these results for the concept of integrated control of the greenhouse whitefly are discussed.


INTRODUCTION
The greenhouse whitefly Trialeurodes vaporariorum Westwood (Hemiptera: Aleyrodidae) is a cosmopolitan and highly polyphagous pest species of greenhouse crops, especially tomato (Gerling, 1990). The incidence of established populations of T. vaporariorum has been documented in many countries of the world's temperate regions (Manzano & van Lenteren, 2009). It has been widespread in Serbia since the 1970s and frequently found in greenhouses as one of the most serious pests of vegetables and ornamentals (Perić et al., 2009;Prijović et al., 2014).
As the widespread use of chemical insecticides has caused an evolution of whitefly resistance to compounds with various modes of action, including newer chemical classes such as neonicotinoids (Whalon at al., 2008(Whalon at al., , 2017, durable and sustainable strategies for control of this pest species should be based on integration of chemical, biological, cultural and other measures (Gerling, 1990;Gentz et al., 2010). The parasitic wasp Encarsia formosa Gahan (Hymenoptera: Aphelinidae; Coccophaginae) has been used for many years now for biological control of T. vaporariorum as one of the most successful biological agents for greenhouse and ornamental crops protection around the world (van Lenteren & Martin, 1999;Enkegaard & Brǿdsgaard, 2006).
Pesticides are necessary when natural enemies fail to keep a pest population below its economic threshold, as a result of adverse environmental conditions, inadequate host species or cultivar, excessive density of pest population, etc. (Hoddle et al., 1998;Albajes et al., 1999). Despite many restrictive factors, natural enemies can be effectively integrated with chemical treatments based on the knowledge of pesticides and their effects on natural enemy populations (Croft, 1990).
Imidacloprid is the first compound in the neonicotinoid class that has a systemic activity, so that its formulations for soil and foliar application have been developed for use in a variaty of agricultural crops. Imidacloprid has demonstrated good efficacy in controlling a considerable number of pests in the order Hemiptera, including the greenhouse whitefly, which is usually treated with foliar applications in different crops (Cahill et al., 1996;Nauen et al., 1996Nauen et al., , 1998Horowitz et al., 2004;Grafton-Cardwell et al., 2008).
Most studies that have quantified the risk involved in imidacloprid applications in combination with E . formosa have been based on the use of its recommended application rates and they were conducted as acute toxicity and/or persistence assays as proposed by the IOBC testing scheme (Yankova et al., 2011;Sterk et al., 2002;Richter et al., 2003). It has also been noted that different methods of imidacloprid application may also cause variable effects on Е. formosa (van de Veire & Tirry, 2003) and other parasitoids (Stapel et al., 2000). Many laboratory tests have also assessed the safety of imidacloprid for different development stages of other parasitoids (Stapel et al., 2000;Morales, et al. 2006;Sabahi et al., 2011;Carvalho et al., 2012) and predators (Sclar et al., 1998;Torres et al., 2002). On the other hand, stimulation of reproduction of pests or natural enemies by imidacloprid, and its safety to beneficial insects have been documented in a number of studies (Elbert et al., 1991;Elzen, 2001;James & Vogele, 2001;James & Price, 2002;Sohrabi et al., 2012). However, only a few studies have focused on sublethal effects of imidacloprid by estimating population-level response of parasitoids (Saber et al., 2009;Saber, 2011;Kheradmand et al., 2012;Sohrabi et al., 2012Sohrabi et al., , 2013 or predators (James, 1997). To the best of our knowledge, there are no data in literature focusing on the effects of imidacloprid on population growth of Е. formosa.
Our previous examination of biological parameters of six local populations and a commercial "Dutch" strain (D) of Е. formosa had shown that the local population Bujanovac (B) had the greatest potential for controlling greenhouse whitefly, while population Negotin (N) demonstrated the least ability (Drobnjaković et al., 2016). The present study focused on the effects of an imidacloprid-based insecticide on life history traits and population growth of F 1 generation, following treatment of pupae as the least susceptible development stage of the parasitoid. Data obtained in this study are discussed in terms of potentials for improving T. vaporariorum integrated management strategy.

Test organism (origin and rearing)
Local populations of E. formosa have been established from insects collected at the pupal stage (parasitized whitefly larvae) in extensive crops of vegetables and ornamentals in locations known not to have a history of biological control of greenhouse whiteflies by a commercial E. formosa strain so as to avoid possible mixing of populations. Specimens of population B were collected in Bujanovac, Southern Serbia (GPS: 42°30'27'' N, 21°48'30'' E) on Solanum nigrum Linnaeus, while population N was collected in Negotin, Eastern Serbia (GPS: 44°13'00'' N, 22°31'00'' E) on Hibiscus spp. The male and female parasitoid wasps issued from eclosed pupae in both populations were identified as E. formosa according to a determination key provided by Polaszek et al. (1992). All wasp populations demonstrated the asexual pattern noted also by other researchers (Zchori-Fein et al., 1992), i.e. males were very rare under regular conditions. A commercial strain of E. formosa (Koppert Biological Systems, Netherlands), referred thereafter as "Dutch" strain (D) of normal susceptibility to pesticides, was used as the reference strain.
The commercial strain and two selected representatives of local E. formosa populations were reared on tobacco leaves infested with the host pest T. vaporariorum at 27±1°C temperature, 60±10% relative humidity, and under 16/8 h light/darkness photoperiod. A T. vaporariorum population was established in 2008 from insects collected from weed plants in the territory of Belgrade, and it has been reared ever since on tobacco plants, cv. Samsun, in ventilated muslin cages according to EPPO (2004) methodology.

Bioassays
All assays were performed at 27±1°C temperature and 60±10% relative humidity, under 16/8 h photoperiod, and in four replications. All assays were conducted in Petri dishes (12 cm diameter) with four openings (1 cm) on the lids, covered with muslin for ventilation and to prevent condensation inside the dish, on tobacco leaves set upon 1% agar layer. Imidacloprid was applied as a dilution in distilled water by spraying the entire surface of both the lid and the dish containing a tobacco leaf. A Potter Precision Spray Tower (Burkard Scientific, UK) was used for spraying 2 ml liquid under 100 kPa air pressure to make a 2.7 ± 0.2 mg/cm 2 water deposit in each dish.

Acute toxicity bioassay
In an acute toxicity bioassay, tobacco leaves with parasitized whitefly nymphs (parasitoid pupae) were fixed onto tin aluminium foil with Traganth-kit (natural, non-toxic adhesive). After drying, the leaves were cut into pieces, each containing 20 parasitoid pupae (four days old, i.e. 12 days after egg laying) per replicate, and set onto filter paper inside plastic Petri dishes (filter paper was moistened to ensure that tobacco leaves stay in place during exposure). The Petri dishes (lids too) containing tobacco leaves with pupae were treated with serially diluted solutions (at least 6 concentrations: 800, 400, 200, 100, 50 and 25 mg a.i./l) to cover a 10-90% mortality range (determined in preliminary testing). Control pupae were treated with distilled water only. Two hours after treatement, the leaves were transferred into new Petri dishes to await emergence of parasitoid adults from pupae. Mortality was assessed from the number of emerged adults against the number of treated pupae, nine days after treatment (EPPO, 2004;Simmonds et al., 2002;Sohrabi et al., 2012).
Concentration-mortality data were processed by probit analysis (Finney, 1971) in Polo Plus software (LeOra Software, Berkeley, USA) calculating lethal concentrations (LC 50 , LC 90 ) and the slope of regression line. A lethal dose ratio test was used for comparison of the calculated LC/EC values: when 95% confidence interval between two LC values included 1, those values were considered not significantly different (Robertson et al., 2007).

Bioassay testing effects on life history parameters and population growth in F 1 generation
The effects of the imidacloprid-based insecticide on life history parameters and population growth of E. formosa were assessed by focusing on the vitality and reproduction potential of females surviving treatment at the pupal stage. The bioassay had four replications. Concentrations of the imidacloprid insecticide (30, 20 and 25 mg/l for parasitoid populations B, N and D, respectively) were chosen based on data noted in the previous toxicity bioassay at the pupal stage of the parasitoid (concentrations were within 95 % confidence limits for the LC 50 calculated in acute toxicity bioassays). All surviving female adults that emerged from 40 treated pupae were transferred to Petri dishes containing third and fourth instar larvae/nymphs of the pest whiteflies that were offered for parasitizing at two day intervals until the death of the last female. Female adult survival, longevity and total parasitism, total emergence of adults and instantaneous rate of increase in F 1 generation of the parasitoid from all three populations were noted (Gholamzadeh et al., 2012).
To determine the development time of surviving juveniles, when parasitoid adults were just before emergence from pupae, the number of eclosed adults was noted at 12 h intervals (Enkegaard, 1993). Development time was calculated as the total number of days from egg laying to adult emergence from the parasitoid pupae. To determine the longevity of survived adults, the number of surviving parasitoid females was checked every other day and the longevity of females was calculated as the total number of days a female was alive assuming that their ultimate death occurred at the midpoint of 48 h. Before analysis, data on juvenile development time and female longevity were √x transformed. Parasitism was determined based on the number of parasitized host nymphs (black pupae) in each inspection interval (parasitism/48 h period), as well as the total number of parasitized host nymphs (total, lifetime parasitism) (Stouthamer & Mak, 2002). After determining the number of parasitoid pupae, leaves carrying them were transferred into new Petri dishes to monitor the emergence of parasitoid adults from the pupae. The data were then used to calculate the number of parasitized host nymphs (parasitism/48 h period and total parasitism), and total adult emergence from parasitized host nymphs (total fertility). Parasitism was calculated as the number of parasitized host nymphs per female alive at the midpoint of 48 h (parasitism/48 h period) and during the entire life span of females (total, lifetime parasitism). Adult emergence was calculated as the number of wasps that emerged from parasitized nymphs and reached adult stage. Parasitism and adult emergence data were transformed by √(x+0.1). Parasitism and adult emergence data were used to calculate at the end of each interval the instantaneous rate of increase (r i ), using the formula: Nf where N 0 is the initial number of wasps (e.g. 40 parasitoid pupae per replicate), N f -final number of wasps at the end of t-day (number of black parasitized pest larvae -parasitoid pupae and number of eclosed adults), ∆t -number of days that elapsed from the beginning of bioassay (in our assay ∆t was determined as the shortest oviposition period in the examined population in each bioassay to make the most reliable and uniform comparison of population status). Positive r i indicated a population growth, negative r i was indicative of a population decline, while r i = 0 showed a stable population (Stark et al., 1997;Stark & Banks, 2003). Arcsin √x transformation was applied to the instantaneous rate of increase data. Kaplan-Meier analysis (SPSS for Windows, Version 17) was used to calculate the average female longevity and survival curves were constructed (Enkegaard, 1993;Qiu et al., 2004), which were analyzed by Long-rank test. Development time, longevity, parasitism/48 h, total parasitism and adult emergence, and instantaneous rate of increase data were analyzed by two-way ANOVA (azadirachtin treatment and population were the factors) with means separated by Fisher´s LSD test (p<0.05). Means of all parameters for treatment and control, for each population individually, were separated by Student's t-test (p<0.05), using the software Statsoft Statistica 7.0.

Acute toxicity to parasitoid pupal stage
Acute toxicity parameters of imidacloprid after direct treatment of E. formosa pupae originating from two local populations and a commercial strain are shown in Table 1. The resulting lethal concentrations were considerably below the insecticide maximum recommended concentration (0.1 % = 200 mg a.i./l). A lethal dose ratio test revealed no significant difference between the calculated LC 50 and LC 90 values for the parasitoid pupae.

Effects on life history parameters and population growth of F 1 generation
The juvenile development time was significantly affected by treatment (F 1,18 =97.1, p<0.001), while the examined population (F 2,18 =2.0, p=0.165), and interaction between treatment and population (F 2,18 =2.0, p=0.932) had no significant effect. Juvenile development of the parasitoid wasps in treated pupae of all three populations of E. formosa was 1.59-1.81 days longer than the development in untreated pupae (Table 2). Regarding treated juveniles, population B needed 1.81 days more, population N 1.55 days more and juveniles D 1.73 days more for development than control juveniles. Female wasps that emerged from parasitoid pupae after treatment with the insecticide LC 50 in the acute toxicity bioassay lived either shorter or longer than control females, but the difference in longevity was not significant (F 1,18 =44.31, p=0.083) (Figure 1). Females B lived 0.35 days shorter, females N 0.47 days longer, while females D lived as long as control females. Both treated and control females B had the leading longevity (Bt:Bc = 9.61 days: 9.96 days). The survival curves for female wasps of the examined populations are shown in Figure 2. Females that emerged from pupae treated with imidacloprid had lower survival rates than females which emerged from pupae treated only with distilled water (Bt vs. Bc: ww=38.00,p<0.001;Dt vs. Dc: ww=46.94,p<0.001;Nt vs. Nc: ww = 42.522,p<0.001). Considering only the treated females, there were no significant differences in survival between D and N females p=0.355), nor between B and D females (Bt vs. Dt: ww=194.06, p= 0.055), while B females had better survival data than N females p<0.05). Considering control females, population B had a better survival than populations D p<0.05),p<0.01). Control females D had a higher survival rate than control females N, but the difference is not statistically significant p=0 .527).
Repetitive ANOVA analysis showed that parasitization rates of the females that emerged from pupae treated with imidacloprid in all test populations were significantly affected by the observation period (F 7,126 =195.41, p<0.001). All main effects and associated interactions were found to be statistically significant between and within observation period, at p=0.05 significance level (Table 3).  Table 3. Repetitive ANOVA parameters for main effects, and associated interactions, on parasitization rates of E. formosa females that emerged from imidacloprid-treated pupae, local populations Bujanovac (B) and Negotin (N), and commercial "Dutch" strain (D); Control = distilled water; Treatment = imidacloprid 30 mg/l (B), 20 mg/l (N) and 25 mg/l (D) Parasitization rates (parasitism/48 h period) of the tested E. formosa females that emerged from treated or untreated parasitoid pupae are shown in Figure 3. Considering the examined populations, imidacloprid treatment affected only females D by shortening their oviposition period by two days (14 days) in comparison with control females. Females of the local populations B and N that emerged from treated pupae laid eggs as long as control females (B females 16 days, N females 14 days). Parasitization rate was not reduced significantly by imidacloprid in any of the three populations, except during the oviposition interval of 0-2 days of females B, and the interval of 12-14 days of females D. Females B achieved the highest parasitism within 48 h period; maximum parasitization was noted on the fifth and sixth days of oviposition (25.2 pupae/ female/48 h in treatment, and 25.93 pupae/female/48 h in control).
The results of the two-way ANOVA analysis revealed that reduction in total parasitism was significantly affected by treatment (F 1,18 =5.02, p<0.05), population (F 2,18 =69.37, p<0.001) and their interaction (F 2,18 =6.25, p<0.01). Total parasitism (Figure 3) of wasp females from treated pupae B and N showed no significant difference from control females, in contrast to females D, whose total parasitism was significantly lower by 25.92 %. In both cases, i.e. treatment and control, the highest average number of pest nymphs was parasitized by females B (Bt:Bc = 143.79 pupae/female:145.15 pupae/female). Reductions in adult emergence in populations B (27.48 %) and D (17.52 %) were significantly affected by treatment (F 1,18 =43.09, p<0.001), population (F 2,18 =51.98, p<0.001) and their interaction (F 1,18 =14.01, p<0.001). The emergence of F 1 female adults from pupae treated with imidacloprid in populations B and D differed from the emergence of control adults, while no such difference was noted for population N (Table  4). Control population B had the highest value of this parameter (131.61 adults/female), while population D had the highest value considering treatment data (105.52 adults/female).  The reductions in instantaneous rates of increase (r i ) of females that survived treatment with imidacloprid at the pupal stage were caused by treatment (F 1,18 =19.3, p<0.001) and population (F 2,18 =172.9, p<0,001), while their interaction did not result in a statistically significant inf luence on the r i values (F 2,18 =3.0, p=0.074). Rates of increase of treated females N, compared to control females, were significantly lower, by 4.23 %, while the rates of increase in populations B and D were not significantly reduced (Table 5). Both treated and control females of population B (Bt:Bc = 0.283 females/day:0.286 females/day) had the highest rate of increase.

DISCUSSION
The imidacloprid product applied in our present study directly to parasitoid pupae at the concentrations of 30 mg/l (B), 20 mg/l (N) and 25 mg/l (D) significantly affected female survival after pupal treatment, extended juvenile development, significantly reduced total parasitism of females D, total adult emergence of females B and D, and significantly reduced the instantaneous rate of increase in females N.
Only sporadic studies have examined imidacloprid sublethal effects by quantifying life history parameters and/or population growth of parasitoids (Saber et al., 2009;Saber, 2011;Kheradmand et al., 2012;Sohrabi et al., 2012Sohrabi et al., , 2013 and predators (James, 1997). Sohrabi et al. (2012) used female longevity, fecundity, gender ratio and population parameters such as r m , to evaluate sublethal effects of imidacloprid at population level of the parasitoid wasp E. inaron. They immersed larvae and pupae of E. inaron together with cotton leaves in an imidacloprid solution (LC 25 = 62.5 ppm). The study showed that treatment of that parasitoid at the pupal stage caused net reproductive rate (R0) and mean generation time (T) of E. inaron to decrease significantly, while intrinsic (r m ) and final (l) rates of increase did not change significantly. Similarly, Saber (2011) found imidacloprid to reduce significantly the longevity, fecundity and population parameters of adults of the egg parasitoid Trichogramma cacoeciae Marchall that survived treatment at the pupal stage.
Although the treated females of population D had the highest number of emerged adults, females B lived longer, laid eggs longer, and had higher values of instantaneous rate of increase. Treated females N had less desirable results of all parameters, compared to commercial females D.
Based on high acute toxicity of imidacloprid to the pupal stage of E. formosa in our study and a significant reduction in its life history and population parameters, imidacloprid cannot be considered as compatible for integrated use with this parasitoid. Precise estimation of risks, including an assessment of different ways in which the parasitoid may be exposed to the compound, require further testing in the laboratory, as well as in the field. Before more data have been obtained in field trials under more realistic conditions, our conclusions need to be considered with rational caution. Imidacloprid may exert different undesirable effects on E. formosa and be coactive with other factors apart from its insecticidal effect. Although these laboratory data should not be immediately used in field testing, they provide a sound illustration of potential negative effects that the insecticide may have on the parasitoid in practice and commercial use, while also providing a basis for careful selection of the most compatible insecticide for complementary use with this beneficial organism.