Side effects of different pesticides used in citrus on the adult stage of the parasitoid Aphytis melinus DeBach ( Hymenoptera Aphelinidae ) and its progeny

Twelve pesticides commonly used in citrus in Spain were tested on adults of Aphytis melinus DeBach to determine their effects on parasitoid survival and fecundity, and the duration of the residue of each pesticide. Six of these pesticides were found to be harmless to moderately harmful to this parasitoid in a laboratory assay in closed Petri dishes: spinosad (bait formulation), azadirachtin, fenbutatin, fosetyl-Al, copper oxichloride, and mancozeb, with their scores on the reduction of beneficial capacity (RBC) index being between 21.4 and 94.6% after one week. The other six pesticides classified as harmful were tested on citrus plants to study their persistence over time under greenhouse conditions: Pirimicarb, pyriproxifen, paraffinic oil, abamectin, chlorpyrifos, and lambda-cyhalothrin. Most of these products reduced their negative effect on adults of A. melinus between one and six weeks after treatment, although lambdacyhalothrin was still harmful to parasitoids 11 weeks after application. This information can help growers and consultants to make decisions about pesticide selection and application timing in citrus in order to support IPM implementation when A. melinus is present. Additional key words: fecundity; IPM; mortality; parasitism; persistence. * Corresponding author: zamora@us.es Received: 01-10-12. Accepted: 08-05-13. Abbreviations used: IOBC (International Organization for Biological Control); IPM (integrated pest management); PIEC (predicted initial environment concentration); RBC (reduction in beneficial capacity). Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA) Spanish Journal of Agricultural Research 2013 11(2), 494-504 Available online at www.inia.es/sjar ISSN: 1695-971-X http://dx.doi.org/10.5424/sjar/2013112-3556 eISSN: 2171-9292


Introduction
Integrated pest management (IPM) philosophy stresses the use as little pesticide as possible and only those compounds that are reasonably compatible with the natural enemies that are important in a particular crop.Moreover, in some parts of the world, such as the European Union, pesticide regulations have reduced the number of active ingredients available (e.g., Council Directive 91/414/EEC; European Union, 2011).Determining the compatibility of pesticides with key biological control agents in a crop should be an ongoing activity, particularly in orchard crops such as citrus, where some key pests are under economically significant biological control (Kennett et al., 1999;Jacas & Urbaneja, 2010;Sorribas & García-Marí, 2010).
One of the key pests of citrus in Spain is the California red scale, Aonidiella aurantii Maskell (Hemiptera Diaspididae) (Franco et al., 2006;Jacas et al., 2010), and one of the most widespread and important natural enemies of this scale is Aphytis melinus DeBach (Hymenoptera Aphelinidae) (Kennett et al., 1999).This parasitoid, native to Pakistan and India, was introduced into California (USA) in the 1950s and since then in many citrus producing regions of the word (Kennet et al., 1999).In Spain it is present since the 1980s (Jacas et al., 2006) and it is well established in most citrus areas (Pina & Verdú, 2007;Sorribas et al., 2008).
Parasitoids of the California red scale can provide a good control of the pest if they are not killed or their activities impaired by pesticides (Grafton-Cardwell et al., 2008).However, classical biological control needs to be supplemented with inundative releases in cooler areas where they suffer a high overwintering mortality (Mazih, 2008;Zappalá et al., 2008;Grafton-Cardwell et al., 2011;Olivas et al., 2011).At present, A. melinus plays an important role in the control of the pest in many regions (Kennet et al., 1999) but in Spain neither him nor other parasitoids are able to keep A. aurantii populations below the damage threshold (Jacas et al., 2010).
Laboratory trials with pesticides are generally considered to be the first approach in order to detect their potential harm to beneficial species.Current protocols for detecting pesticides side effects, developed by organizations like the International Organization for Biological and Integrated Control of Noxious Animals and Plants (IOBC), recommend that products showing harmful effects in laboratory tests (run in the most negative conditions and with the most exposed stage of the beneficial species) should be further tested under more natural conditions, such as extended laboratory (to estimate residual life), semi-field, or field trials (Sterk et al., 1999).Under such conditions, harmful pesticides often are less detrimental to beneficial species than under laboratory ones.
Because of the importance of citrus as a crop, there are many studies that have determined the side effects of pesticides on parasitoids that are key natural enemies of citrus pests (Bellows & Morse, 1993;Bellows et al., 1993;Prabhaker et al., 2007Prabhaker et al., , 2011;;Suma et al., 2009;Vanaclocha et al., 2012).Continuous evaluation of new products allows the creation of data banks with the side effects of pesticides on natural enemies, as it happens with citrus (Jacas & Urbaneja, 2010;IVIA, 2012), which are of great importance to IPM implementation.
The active ingredients investigated in this study cover a wide range of products used in Spanish citrus, in both IPM and organic farming.We studied the side effects of these pesticides on A. melinus, first in laboratory trials with both fresh and aged residues, and afterwards in an extended laboratory trial to estimate the persistence and rate of degradation of the more harmful materials.The objective of this study was to provide information on the compatibility of pesticides and A. melinus, including the effect on progeny production, which can help to estimate potential effects on natural enemy populations.Information generated can be useful for farmers and consultants to take decisions on pesticide selection and application timing in citrus to support IPM implementation when A. melinus is present.

Material and methods
Tests followed, with adaptation, the principles of the IOBC/WPRS Working group ''Pesticides and Bene-ficial Organisms'' (Hassan et al., 1994;Sterk et al., 1999).Evaluations were done of (1) the contact toxicity to adult parasitoids after 24 h with fresh and 7-dayold residues, (2) duration of harmful activity (persistence) of residues of the more harmful products, and (3) effects on parasitization activity and sex ratio of parasitoid offspring.

Insect rearing
Adults of A. melinus used in the assays were reared in the facilities of the University of Sevilla, following the method developed by DeBach & White (1960) for rearing Aphytis lignanensis and later modif ied by others (Rose, 1990;Raciti et al., 2003).The procedure is based on first rearing the host (a parthenogenetic strain of Aspidiotus nerii Bouché) on butternut squash (Cucurbita moschata Duchesne ex Lamarck).When the host is in the second instar, the infested squash and adult parasitoids are placed together in a ventilated cage.Adult parasitoids emerge about 15 days later.Parasitoids produced in this manner were used in our experiments within 1 to 2 days of emergence.

Assay with fresh residues
This assay was carried out using residues on the inner walls of glass Petri dishes (5 cm diam.).Pesticides were applied to tops and bottoms of dishes at the laboratory concentrations (Table 1) with a potter preci-Side effects of pesticides on Aphytis melinus sion spray tower (Burkard Manufacturing Co. Ltd, Rickmansworth, Hertfordshire, UK), at 0.7 bar pressure.The tower was calibrated to leave 1.5 mg of solution cm -2 , and the concentration of the pesticides was calculated using the Predicted Initial Environment Concentration (PIEC) (Barret et al., 1994), with the expression: PIEC (µg cm -2 ) = maximum f ield dose (g ha -1 ) • f /100.This procedure left a residue on the glass surface that was equivalent in concentration to that left on leaves in field applications, considering a correction factor f depending of the type of crop (f = 1 for horticultural and arable crops, and f = 0.4 for trees).
Pesticides were left to dry at room temperature for an hour, and then 6 to 8 adult females of A. melinus that were 24-48 h old were introduced into each Petri dish, adding several drops of honey as food.Tops and bottoms of the Petri dishes were kept together using elastic bands.Petri dishes (4 to 6) were used for each pesticide as replications.Parasitoid mortality on each Petri dish was evaluated after 24 h.Surviving females of each Petri dish were transferred to a correspondent ventilated container with a piece of squash with drops of honey and an excess of A. nerii scales.Females were left with the scales for 48-72 h to parasitize the hosts offered, and the resulting parasitoid offspring were counted and sexed following their emergence.The experiment was carried out at 25 ± 1°C, a 16:8 L:D photoperiod and 60 ± 5% relative humidity.
Tests were run in four batches (designated as Experiments 1, 2, 3a, 3b) with each batch containing three-four pesticides (with the exception of 3b) and a control treated with distilled water.Test with fenbutatin was repeated in a separate experiment (3b) due to technical difficulties with the original trial.

Assay with 7-day-old residues
Eleven of the original 12 pesticides (fosetyl-Al was excluded from further study) were tested in a second assay, using 7-day-old residues.Materials were separated into two groups: harmful, when mortality in the fresh residues assay was 100%, and less harmful pesticides, when mortality in the fresh residues assay was lower than 100%.These two groups were run in separate experiments.
After products were sprayed on Petri dishes, these were left upside open, but covered with a filter paper, under room conditions (20 ± 2°C) for 7 days to allow residues to age and evaporate because the experimental units had no ventilation.After this period, 6 to 8 A. melinus females (24-48 h old) were introduced in each Petri dish, adding several drops of honey as food.Six Petri dishes were used for each pesticide and the control, as replications.Adult mortality and progeny production was evaluated as in the previous assay.

Persistence assay
The most toxic products found in the previous assays were selected to study their residue degradation over Copper oxichloride 50 (w/w) 1.0 • 10 3 8.0 • 10 3 Diseases a All suppliers are from Spain.b Field rates were calculated with the maximum label dose.a.i.: active ingredient.c Laboratory concentrations were obtained applying the PIEC calculation, using the field rate and a volume of 3,000 L of water ha -1 .
time.This experiment was an extended laboratory test, where products were applied with a portable sprayer (Florabest, Abraham Diederichs GmBH & Co., Wuppertal-Germany) on young cv.Navelina orange trees, and the evaluation of residue persistence was made under laboratory conditions.Three trees were sprayed with each pesticide at the maximum recommended field rate until run off at a pressure of 2 bars.Control trees were sprayed with tap water.The trees were kept in pots inside a greenhouse (located in the facilities of the Escuela Técnica Superior de Ingeniería Agronómica, University of Sevilla) for security reasons, irrigated every two days, and fertilized once a week throughout the course of the assay.The greenhouse was of 250 m 2 , covered with a polyethylene plastic film of thickness 200 µm.The UVA radiation was measured inside and outside the greenhouse in several dates, with a reduction of 38.4 ± 2.6% of UVA radiation in the interior of the greenhouse.This experiment started at the middle of March 2011, with the application of the the pesticides, and lasted until the beginning of June 2011.
Toxicity of residues was evaluated 1, 2, 4, 6, 9, and 11 weeks after application of the pesticides.For this evaluation, three to four leaves were selected at random from the treated trees of each product and taken to the laboratory at the designated times.Three to four ventilated Petri dishes were prepared for each pesticide and the control at each occasion as replications, placing one or two leaves in each Petri dish over a moistened piece of filter paper.Ventilated Petri dishes were made by replacing most of the bottom half with a piece of organdy for ventilation, and this became the upper part of the test unit.Eight to 10 A. melinus females (24-48 h old) were introduced into each Petri dish, adding several drops of honey as food.Adult mortality and progeny production was evaluated as in the two previous assays.

Statistical analysis
Data on adult parasitoid mortality, the mean number of offspring produced per surviving parasitoid, and the sex ratio of the offspring (as % female) were subjected to one-way ANOVA using the results of each Petri dish as replicate.Means were separated using Tukey's honest significant difference test when analysis of variance were significant at p < 0.05.All data needing to be normalized were transformed before being analyzed.
Percentages were subjected to arcsin square root (x) transformation, and the number of offspring to a log (x + 1) transformation.All analyses were performed using Statgraphics Centurion XVI (Stat Point Technologies, 2010).Levels of mortality were adjusted using the mortality level in the control and Abbott's formula (Abbott, 1925).An index "reduction of beneficial capacity (RBC)" was also calculated, as a parameter that integrates both mortality and offspring production, using the following formula (Overmeer & van Zon, 1982): where M c = corrected mortality of the treated A. melinus females, R t = reproductive performance of the treated A. melinus females, and R c = reproductive performance of control A. melinus females.
Fecundity of treated wasps was similar to the controls for some products where adults survived, as for Side effects of pesticides on Aphytis melinus example azadirachtin in Experiment 1 (Fig. 1a), spinosad and copper oxichloride in Experiment 2 (Fig. 1b), and fosetyl-Al in Experiment 3a (Fig. 1c).Significantly reduced fecundity compared to the control was seen with mancozeb in Experiment 1 (Fig. 1a), and fenbutatin in Experiment 3b (Fig. 3d).
Taking into account the RBC values of the fresh residues and the IOBC criteria, six products were classified as harmful, two as moderately harmful, and four as slightly harmful (Table 2).As fosetyl-Al produced very low mortality and had a reasonable low value of RBC (Table 2), it was excluded from further experiments with aged residues.

Assay with 7-day-old residues
For the assay with the 7-day-old residues, compounds were separated in two groups, based on their toxicity in the previous assay: Group 1, harmful products, and Group 2, moderately or slightly harmful products.Parasitoid mortality when exposed to 7-day-old residues of compounds in the Group 1 was 100% for all compounds (pirimicarb, paraff inic oil, abamectin, lambda-cyhalothrin, and chlorpyrifos) except pyriproxyfen (with 81.5 ± 11.9% mortality), all of which differed from the level of mortality in the control group (Fig. 2a).Parasitoids exposed to compounds of Group 1 (except for the control) neither survived nor produced offspring.Among offspring of the control group, 77.8% were female (Fig. 2a).
The six products included in the harmful group (1) obtained the highest RBC values, and so were classified as harmful (Table 2).Based on their RBC values, two of the products in the other group (2) were classified as harmless, two as slightly harmful, and one as moderately harmful (Table 2).

Side effects of pesticides on Aphytis melinus 499
Table 2. Mortality of adults of Aphytis melinus (corrected with the Abbott's formula) and reduction of beneficial capacity (RBC) indices for 12 pesticides tested under laboratory conditions, with fresh and 7-day-old residues, using the laboratory concentrations calculated with the PIEC expression.IOBC classes of toxicity are refered to the RBC values obtained with fresh and 7-days old residues aged in the laboratory

Persistence assay
In the persistence assay, parasitoid mortality for most products was very similar to the control over the period of study, but in each week one or two products differed significantly from the control (Fig. 3a).The most toxic compounds were lambda-cyhalothrin, which caused 100% mortality from weeks 1-11, and chlorpyrifos, which caused mortality higher than the control in weeks 2 and 4.
The per capita production of offspring by surviving parasitoids was very similar among those products where some adults did survive as compared with the control (Fig. 3b) in week 1, week 2, and week 4.In week 6 there were a general decline in offspring production (Fig. 3b), with differences between products.
The proportion of parasitoid progeny that were female varied among products in some weeks of the study (Fig. 3c), as in week 1 (for pyriproxyfen and chlorpyrifos) and week 4 ( for chlorpiryfos), but not in week 2 or week 6.Overall, the proportion of progeny that were female was 57.8%.
Mortality (corrected with Abbot's formula) in the persistence assay decreased over time, and even in the first week after pesticide application, parasitoid mortality was low for residues of most of the products with the exception of lambda-cyhalothrin, which caused 100% mortality through week 11, and chlorpyrifos, which caused a high mortality in weeks 1 to 4 (Fig. 4a).

Discussion
The implementation of biological pest control in citrus in Spain is increasing.However, the use of diffe-rent types of pesticides is still necessary, making it important to study the side effects of such pesticides on the most prevalent natural enemies in citrus (Jacas et al., 2010).
Five (spinosad, azadirachtin, fenbutatin, fosetyl-Al, and copper oxichloride) out of twelve pesticides tested caused little harm to A. melinus in experiments with fresh or 7-day-old residues in the closed Petri dishes, so they were excluded of the extended laboratory assay.Toxicity of these compounds was low seven days after application, so it is reasonable to expect that their persistence in real applications must be even lower.The exception is mancozeb, which although caused only low mortality, greatly reduced progeny production by surviving parasitoids (resulting in a high RBC value) in the laboratory experiments.For this reason it should be necessary to study its effect in a longer term and with more real conditions.The products that were most harmful when presented as fresh residues (paraffinic oil, pirimicarb, pyriproxyfen, abamectin, lambda-cyhalothrin, and chlorpyrifos) were still harmful when presented as 7-dayold residues.Pirimicarb and pyriproxyfen are generally considered compatible with A. melinus in laboratory studies (Prabhaker et al., 2007;Rill et al., 2008;Zappalà et al., 2011), but our results showed harmful effects in closed Petri dish assays, even after seven days of aging and evaporation.Our results were obtained in the worst possible conditions for the insects.Pesticides concentrations used in the laboratory experiments were calculated to leave a pesticide deposit per surface unit similar to the used in field applications in citrus, which is a practice that few researchers follow in their works.Besides, experimental units were closed Petri dishes, and although a period of seven days was used to allow degradation and evaporation of components in the formulations, the results indicate that after this period pyriproxyfen (and other products considered compatible with A. melinus) was still harmful, if not completely in the mortality effect, at least in the overall evaluation (RBC) because no progeny was obtained from the surviving females.
Paraffinic oil was found to be harmful to A. melinus adults in our laboratory tests with closed Petri dishes.Also, when applied to larvae or pupae of several California red scale parasitoids, paraffinic oil caused very high mortality (Domínguez et al., 2003).Narrow-range mineral oil has also been found to be harmful to A. melinus (Zappalà et al., 2011).The other two products in the most toxic group, abamectin and chlorpyrifos, have been generally considered to be toxic in laboratory studies (but with low persistence) to A. melinus and other parasitoids found on citrus crops (Prabhaker et al., 2007;Campos et al., 2008;Suma et al., 2009), as they were here.We have not found studies on the effects of lambda-cyhalothrin on A. melinus.
The bait formulation of spinosad used in this work showed little harm to A. melinus, especially as 7-dayold residue, even though it was sprayed over the entire inside of the Petri dish.Its recomended use in field applications is spraying in localized spots.Although fresh residues of spinosad bait caused some mortality, it was not as high as reported by other works (Michaud, 2003).Spinosad baits (0.02% of active ingredient) showed no harm to Aphytis spp. or Comperiella bifasciata Howard populations in field trials (Thomas & Mangan, 2005), nor were harmful effects observed in laboratory trials for different natural enemies common in citrus orchards (Urbaneja et al., 2006).In another study, on the other hand, a more concentrated formulation of spinosad (44.2% of active ingredient) was harmful to A. melinus (Suma et al., 2009).
Most of the products (pirimicarb, pyriproxyfen, paraffinic oil, abamectin, and even chlorpyrifos) included in the persistence assay showed a decrease in their negative effects in the first week, under the more natural conditions (solar radiation, residues on leaf, dust) of this assay, in contrast to the effects of residues presented on glass in the laboratory (as in the first two assays).This was not true, however, for lambda-cyhalothrin, which remained highly toxic for at least 11 weeks.
Pirimicarb, pyriproxyfen, paraffinic oil, and abamectin can be considered slightly or moderately persistent, whereas chlorpyrifos and lambda-cyhalothrin were persistent.In spite of this, 6 weeks after treatment parasitoid mortality was very low for most compounds, with only lambda-cyhalothrin causing high mortality for a long period.These results agree with the recommendations of different insectaries regarding how long it is necessary to wait before releasing parasitoids after the application of these pesticides (Biobest, 2011;Koppert Biological Systems, 2011).The values of persistence of the compounds presented in this work must be considered, anyway, as maximum values, due the negative effect of the plastic cover used in the greenhouse on the compounds photodegradation.This is particularly right for abamectin, which is known to photodegradate very quickly due to UV radiation (Demchak & Dybas, 1997;Van de Veire et al., 2004), and whose residual effect on A. melinus in field conditions is very short (Morse et al., 1987).It is reasonable to expect that in normal outdoor conditions persistence values would be lower than the obtained in this work, but the practical consequence is that our persistence classification provide an extra security time for the parasitoids before their introduction in a particular grove.
Other compounds (as pyriproxyfen, mineral oil, and chlorpyrifos) have been also tested in field or semifield conditions to assess their effects on natural enemies, being pyriproxyfen the most compatible with various natural enemies, including A. melinus (Grafton-Cardwell & Reagan, 1995;Grafton-Cardwell et al., 2006).Horticultural mineral oil (paraffinic oil) has also shown to be compatible in field conditions with a wide range of natural enemies found in citrus orchards (Liang et al., 2010).
Mortality rates seen in tests were consistent and generally had a decreasing slope over time (except for lambda-cyhalothrin).In contrast, progeny production (included also within the RBC value) was more variable, especially in week 6, when there was a general decline in offspring production, which altered in some way the general trend.
Pirimicarb showed a different trend over time in its RBC value than other compounds.The negative values of this parameter for this compound mean that in some instances, progeny production of wasps exposed to this compound was actually higher than for the controls, although no significantly, as happened in weeks 1 and 2.
The work presented here provides information that growers and consultants can use to select pesticides compatible with A. melinus for use in a citrus IPM context.Information is presented not only on direct mortality of adults, but also on parasitoid progeny production, which is critical for the long term stability of the biological control of pests (Stark et al., 2007).Moreover, more information on pesticide residue persistence, as for example with mancozeb, is needed to help schedule inundative releases of A. melinus in citrus.

Figure 4 .
Figure 4.Toxicity of aged pesticides of several compounds to Aphytis melinus females in an extended laboratory assay using residues applied to young orange trees, where (a) is the corrected mortality using Abbott's formula and (b) is the reduction of beneficial capacity (RBC) index.Vertical bars represent the standard error.

Table 3 .
Mortality of adults of Aphytis melinus (corrected with the Abbott's formula) and reduction of beneficial capacity (RBC) indices for six pesticides tested under extended laboratory conditions using young orange trees.IOBC classes of toxicity and persistence are refered to the RBC values in the table and the time to reach such values c Persistence, with four evaluation categories according to IOBC: A, short lived (< 5 days); B, slightly persistent (5-15 days); C, moderately persistent (16-30 days); D, persistent (> 30 days).