Laboratory evaluation of effects of powdered sulphur on the oviposition, fruit detection and behaviour of Drosophila suzukii (Diptera: Drosophilidae) on strawberry

Drosophila suzukii (Matsumura, 1931) is an invasive pest from South-East Asia that can damage a wide range of softskinned fruit crops (especially berries) resulting in serious fi nancial losses. At present, control of D. suzukii is still mainly based on adult-oriented chemical pesticides that is likely to result in resistance and environmental pollution. In this context, alternative methods that are compatible with sustainable management are required. Sulphur and its different formulations are widely used for a long time as a fungicide and insecticide on many crops as it is environmentally-friendly and compatible with organic management. In this study, the effect that powdered sulphur-treated strawberries had on oviposition were evaluated under laboratory conditions in order to confi rm the fi ndings of a previous report on blueberries. No-choice and choice experiments were carried out to assess whether powdered sulphur affected the ability of the fl y to detect fruit, altered its behaviour or did both. The treatment of strawberries resulted in a reduction of over 79% in the total number of D. suzukii that developed. In the no-choice experiment, there was a reduction of over 70% in the time the fl ies spent on treated compared to untreated fruit. In the choice experiment, 57% of the females chose untreated and 43% treated strawberries, but this difference was not statistically signifi cant. Furthermore, females took over 75% longer to visit the treated fruit than the control. Sulphur had no noteworthy effect on fl y behavior, although oviposition only occurred on untreated fruit indicating that sulphur affects the acceptance and suitability of fruit for females. These results indicate that powdered sulphur could be an appropriate alternative to synthetic-organic insecticides for controlling D. suzukii and is compatible with organic crop management and waste reduction strategies.


INTRODUCTION
The Spotted Wing Drosophila, Drosophila suzukii (Matsumura, 1931) (Diptera: Drosophilidae), is a polyphagous invasive pest native to South-East Asia, which affects a wide range of wild plants and important crops in Europe and America. Because it has a serrated ovipositor this species can infest unbruised ripening soft fruit (Asplen et al., 2015). Larvae feed and develop inside the fruit of numerous crops, which becomes unsellable and results in a dramatic reduction in fruit production and fi nancial loss Walsh et al., 2011;Kinjo et al., 2013). Due to its ability to disperse and polyphagy D. suzukii has spread rapidly throughout the world since 2008 when it was fi rst detected in Southern Europe and North America Cini et al., 2012). Berries (especially raspberries, blueberries and strawberries) and stone fruit (e.g. cherries, using choice and no-choice experiments in order to determine whether powdered sulphur affects their ability to detect fruit, alters their behaviour on fruits or both.

Insects, fruit and insecticide
The D. suzukii adults used in the bioassays came from a colony established at the IFAPA Laboratory of Entomology "Las Torres" (Alcalá del Río, Seville, Spain) from larvae collected from a fi eld of raspberries at Huelva (southern Spain) in February 2014. The colony of D. suzukii was continuously reared on berries, mainly blueberries, Vaccinium corymbosum L. (Ericaceae) and raspberries, Rubus idaeus L. (Rosaceae). The D. suzukii adults were kept at a 22 ± 1°C, 65% RH, and 16L : 8D photoperiod in 29 × 29 × 29 cm plastic cages (BugDorm® 1, Bio-Quip Products Inc., Rancho Rodríguez, CA, USA) with a 25 mm plastic cup containing a small ball of cotton saturated with 10% w/v honey-dH 2 O as a food source. The female and male D. suzukii used in the experiments were between 5 and 12 days old and had no contact with fruit from the time they emerged until they were used in the experiments. The substance tested was powdered sulphur (98.5% DP formulation, Bago d'Ouro®, Sapec Agro, Valencia, Spain). Ripe commercial strawberries Var. "San Andrea" (FresDoñarosa, Superexport Cia. Agraria S.L., Huelva, Spain) were used in the experiments.

Effects on oviposition
A choice experiment was used to determine the deterrent effect of powdered sulphur on D. suzukii oviposition. Prior to the experiment strawberries were rinsed with distilled water and then allowed to dry for 2 h. Strawberries were placed in a 0.036 mm nylon mesh sieve (Filtra Vibración S.L.; Barcelona, Spain) and powdered sulphur was applied directly to them (about 1mg per fruit); then they were gently rolled by tilting the mesh to prevent damage and to ensure that the entire surface of each fruit was covered. The fruit was shaken carefully so that all surplus powdered sulphur fell through the mesh (Pérez-Guerrero & Molina, 2016). The dose per fruit was estimated by weighing 10 groups of 10 fruit beforehand, and this was 0.19 ± 0.002 g of sulphur per fruit after treatment. All treated fruit were kept under laboratory conditions (22 ± 1°C, 60 ± 5% RH, and a 16L : 8D photoperiod) for 1 h before exposure to D. suzukii. The fruit were exposed to D. suzukii in 29 × 29 × 29 cm plastic cages (BugDorm® 1, Bio-Quip Products Inc., Rancho Rodríguez, CA, USA). Two strawberries (treated and untreated) were placed on the floor of the cage in alternating order in each of the replicates (to eliminate any bias). The strawberries were spaced 4 cm apart and 5 cm from the food source (a 25 mm plastic cup containing a small ball of cotton saturated with 10% w/v honey-dH 2 O). Two females and two males per cage were exposed to two strawberries for 24 h. No adults died during the experiments. After this time, the strawberries were removed from the cages and each fruit was kept separately and incubated under laboratory conditions. D. suzukii eggs are diffi cult to see on structured fruit such as blackberries, raspberries and strawberries, which renders egg counts unreliable , so the strawberries were dissected seven days after exposure and the number of larvae and pupae counted. To quantify the larvae and pupae per fruit, each strawberry was placed in a plastic bag with salt water (5%-6% v/v). The strawberries were lightly crushed and left for at least 1h before the contents of the bag were transferred to a container with a dark background and the D. suzukii larvae and pupae were counted under a dissecting microscope (Van Timmeren & Isaacs, 2013;Pérez-Guerrero & Mateus, 2018). EU berry markets (European Comission, 2019). In recent years, biological methods of control (e.g. parasitoids) and other alternative methods (e.g. plant extracts, push and pull strategies, insect proof-nets, harvest schedule/sanitation) have been increasingly incorporated into integrated pest management programs (Leach et al., 2016(Leach et al., , 2018Rossi-Stacconi et al., 2017;Pérez-Guerrero & Mateus, 2018), in order to mitigate the consequences of using chemical insecticides and manage this pest in a sustainable manner. Numerous plant-derived extracts and other compounds have been used to deter D. suzukii from ovipositing on fruit. Gargani et al. (2013) report the negative effects of Sophora fl avescens extract on egg viability on blueberries, Erland et al. (2015) those of Lavandula latifolia and commercial avocado oils and Renkema et al. (2016) the deterrent effect of peppermint oil (among others) and the deterrent and lethal effect of thyme oil on D. suzukii males. Finally, Pérez-Guerrero & Mateus (2018) recently reported a signifi cant reduction in D. suzukii infestation six days after using S. fl avescens and garlic extracts, and suggest they affected egg viability and/or larval development.
Sulphur and its different formulations have been widely used as fungicides for a long time (e.g. on strawberries for controlling Podosphaera aphanis Wallr.) and as insecticides in a range of crops (e.g. in strawberries, vineyards and tomatoes for controlling Tetranychus urticae Koch and Tuta absoluta Meyrick) because they have little effect on the environment and are compatible with organic management (Williams & Cooper, 2004;Bloem et al., 2005;Zappalà et al., 2011). However, there has been little research on their effectiveness against D. suzukii. Some sulphurbased formulations signifi cantly reduce oviposition on blueberries and strawberries (Pérez-Guerrero & Molina, 2016;Andreazza et al., 2017b). In the laboratory powdered sulphur on blueberry fruit signifi cantly reduces oviposition and adult emergence by over 76% and 96% in no-choice and choice experiments, respectively, and over 40% adult mortality after a seven day exposure (Pérez-Guerrero & Molina, 2016). Andreazza et al. (2017b) report a 47.6% reduction in oviposition and greatly arrested immature development when plants were treated with lime-sulphur. In contrast, powdered sulphur results in low mortality (< 40%) when used as an insecticide for controlling adults and larvae of D. suzukii and Zaprionus indianus Gupta (Andreazza et al., 2017a). It is known that powdered sulphur deters D. suzukii from ovipositing (Pérez-Guerrero & Molina, 2016), but the mechanism by which sulphur deters D. suzukii is unknown. Thus, reduction in oviposition may be due either to a deterrent effect of sulphur or it may alter how females behave on fruit.
Our objective was to determine the effects of powdered sulphur on D. suzukii in terms of oviposition, fruit detection and time spent by adults on strawberries in the laboratory using video techniques. First, a choice experiment was carried out to determine whether powdered sulphur inhibits egg-laying on strawberries in order to confi rm previous results obtained using blueberries. Second, the time spent on fruit and the behaviour of fl ies on fruit were analysed

Eff ects on fruit detection and behaviour
A no-choice experiment was used to analyse the potential effects of powdered sulphur on the ability of the D. suzukii to detect fruit. The strawberries were treated as above and a treated or untreated strawberry was placed in 29 × 29 × 29 cm plastic cages (BugDorm® 1, Bio-Quip Products Inc., Rancho Rodríguez, CA, USA) with one female and one male D. suzukii, and kept at a 22 ± 1°C, 60 ± 5% RH, and a 16L : 8D photoperiod. The fl ies were observed for 30 mins and the time they took to fi rst visit the fruit, the number of visits and time they spent on the fruit were recorded for the males and females using a digital stopwatch. After 30 mins, the strawberries were removed from the cages and each of them was incubated separately under laboratory conditions. Each treatment in this experiment was replicated 25 times over 1-9 days by three observers.
Finally, a choice experiment was used to evaluate the effects of powdered sulphur on the behaviour of females. One treated and one untreated strawberry were placed on 4 mm-diameter plastic rings and placed in the experimental arena, which was a 12 × 12 × 12 cm glass cube. The strawberries were 4 cm from each other and one female was placed in the centre of the arena and fi lmed for 15 min using two digital video cameras (Canon HFS 21), which together were focused on the entire surface of the fruit. The images were analysed using the Adobe Premiere Pro CS5 video program. Four categories were used to defi ne the behaviour of the fl y : (1) walking, the fl y takes steps with interruptions < 1s; (2) cleaning, the fl y does not move and rubs its legs together or its head, wings or ovipositor; (3) oviposition, the fl y probes the skin of the strawberry with its ovipositor and (4) stationary, the fl y does not move or behave in any of the ways described above (Lemoyne et al., 2008). In addition, the time that elapsed until the fi rst visit to the fruit, the time spent on the fruit and the number of visits were recorded. There was a total of 25 replicates of this experiment.

Statistical procedures
Since residuals were not normally distributed in the choice and no choice experiments, generalized linear mixed models (GLMMs) were used to determine the effects of the treatment on the number of D. suzukii that developed using the "glmer" function in the R v.3.1.3 software package. The GLMMs included treatment as a fi xed factor and total D. suzukii larvae and pupae as dependent variables fi tted to a Poisson distribution with a log-link function. Since treated and untreated fruit were in the same cage, replicate (ID test) was included as a random factor and treatment as a random slope by adding it to the random effects structure. Therefore, the model estimates a random intercept, random slope, the correlation between the two and the fi xed effect of treatment (Harrison et al., 2018). The data from the second ("no choice") experiment were analysed in the same way: GLMMs with interaction terms, which included treatment (untreated and treated) and sex as fi xed factors and time spent on and off the fruit or t he time that elapsed until the fi rst visit to the fruit and from then until the end of the experiment (using the cbind function) as dependent variables fi tted to a binomial distribution with a logit link function or number of visits as dependent variables fi tted to a Poisson distribution with a log-link function. In order to incorporate male-female variability, replicate (ID test) was considered to be a random factor and sex was specifi ed as a random slope by adding it to the random effects structure. Data from the third experiment ("choice") were also analysed using GLMMs: fi rst, treatment was included as a fi xed factor and time spent on and off fruit or time that elapsed before the fi rst visit to the fruit and then to the end of the experiment (using the cbind function) were dependent variables fi tted to a binomial distribution with a logit link function.
Secondly, the treatment was included as a fi xed factor and time spent on and off the fruit for each behavioural category (using the cbind function) were dependent variables fi tted to a binomial distribution with a logit link function. As in the fi rst experiment, replicate (ID test) was considered as a random factor and treatment was specifi ed as a random slope by adding it to the random effects structure. We checked the models for overdispersion and residual distribution using the DHARMa package. When overdispersion was detected, the models were fi tted to a quasibinomial or quasipoisson distribution using the "glmmPQL" function. GLMMs procedures used the Wald statistic ("z") value and Pr([|z|]) or "t" and Pr([|t|]) (when dispersion is estimated) in order to analyse the effects each factor has on the response variable, testing the hypothesis that the corresponding parameter (regression coefficient) is zero (Crawley, 2005). Finally, the number of females that fi rst visited the treated and untreated fruit were compared using Pearson's χ 2 test.

Effects on oviposition
There was a signifi cant difference in the number of D. suzukii larvae and pupae recorded in the treatments (z = 2.49; P = 0.0125). The number of larvae and pupae in treat-  ed fruit (0.64 ± 0.20) was reduced by 79% compared to untreated fruit (3.09 ± 0.65; Fig. 1).

Effects on fruit detection and behaviour
There were no signifi cant differences (z = 0.95; P > 0.05) in the mean number of visits to treated (0.72 ± 0.14) and untreated fruit (0.52 ± 0.08) in the no-choice experiment. The GLMMs also revealed no signifi cant differences (z = -0.62; P > 0.05) between males (0.52 ± 0.10) and females (0.72 ± 0.12) and in the interaction between these factors (z = -0.04; P > 0.05). Likewise, there were no signifi cant differences in the time that elapsed before the fi rst visit to the fruit among the treatments (t = 0.85; P > 0.05; Fig. 2), by males and females (t = -1.66; P > 0.05) and the interaction between these factors (t = -0.67; P > 0.05). In contrast, the GLMMs revealed signifi cant differences in the time spent on the fruit among treatments (t = 5.61; P < 0.001). There was a reduction of over 70% in the time the fl ies spent on treated than on untreated strawberries (229.02 ± 65.12 s and 755.72 ± 117.37 s, respectively; Fig. 2). In addition, there were no signifi cant differences in the time spent on the fruit by males and females (t = 1.50; P > 0.05) and their interactions (t = -0.70; P > 0.05).
As with the no-choice experiment, no differences were recorded in the choice experiment regarding which fruit (42.86% for treated and 57.14% for untreated or control) was visited fi rst by females (χ 2 (1) = 0.86; P > 0.05). The females took over 75% longer (175.75 ± 74.68 s) to visit treated fruit (z = -279; P = 0.00 5; Fig. 3) than the control (44.00 ± 14.54 s). In contrast, there were no signifi cant differences in the time spent on the untreated and treated strawberries (z = 1.02; P = 0.049; Fig. 3). Furthermore, the GLMMs revealed no signifi cant differences between the treatment and control in terms of the time spent on the fruit in each behavioural category (Table 1). There were no signifi cant differences between the control and treated fruit in the time the fl ies spent walking, stationary and cleaning (z = -0.69; P > 0.05; z = 1.19; P > 0.05 and z = -1.01; P > 0.05, respectively). Finally, although oviposition only occurred on control fruit (Table 1), GLMMs revealed no signifi cant differences with respect to the treated fruit (t = -0.009; P > 0.05).

DISCUSSION
The results for strawberries support the previous report by Pérez-Guerrero & Molina (2016) for blueberries in which the infestation of fruit treated with powdered sulphur was signifi cantly less. Treatment of strawberries resulted in a reduction of over 79% in the number of D. suzukii that developed. There are few studies on how sulphur affects D. suzukii (Pérez-Guerrero & Molina, 2016;Andreazza et al., 2017a, b). Das et al. (2010) report that the sulphur-based fungicide Thiovit affects several life history traits of Drosophila melanogaster resulting in a slightly lower infestation. However, it is comparable to the effect of the treatment of blueberries with same sulphur formulation on eggs laid and adults affected (98 and 96%, respectively) under similar conditions (Pérez-Guerrero & Molina, 2016). In this study, the reduction estimates for strawberries were based on the number of larvae and pupae per fruit after seven days. Powdered sulphur is more effective than other formulations (Andreazza et al., 2017b). In fact, Andreazza et al. (2017b) report a signifi cant 47.6% reduction in oviposition and marked slowdown in immature development when plants were treated with lime-sulphur. However, when powdered sulphur was used as an insecticidal bait it did not control the adults and larvae of D. suzukii and Z. indianus as the mortality was less than 40% (Andreazza et al., 2017a).
Powdered sulphur is highly effective in controlling D. suzukii attacking strawberries and could be used instead of pesticides (e.g. organophosphorates, pyrethroids and spinosyns) in conventional production. In addition, powered sulphur could help control this pest in organic production (powdered sulphur is currently authorized as a fungicide for organic production of susceptible crops such as strawberries, and farmers use it at least once or twice a year at a rate of 20-30 kg per hectare). Furthermore, fi eld observations indicate that the lower effect of D. suzukii on strawberry crops in southern Spain (grown intensively in large plastic tunnels from October-November and May-June) may be due to the widespread use of powdered sulphur  Table 1. Average (± SE) and percentage (%) of the time females of Drosophila suzukii spent in each of the defi ned behavioural categories (see Material and Methods section) when simultaneously exposed to treated and untreated (control) fruit in a choice experiment for 15 min. Different letters denote a significant difference based on generalized linear mixed models (GLMMs) between control and treatment (alpha = 0.05).

Behaviour
Actual time spent ( during the fruiting season for deterring other pests and preventing plant diseases (Molina, 2015). However, it must be stressed that further research is required in order to evaluate how the residue of sulphur dust affects the pre-harvest and sale of treated fruit in order to defi ne commercially viable strategies for managing the different crops attacked by D. suzukii (Pérez-Guerrero & Molina, 2016).
Results obtained in the no-choice experiment showed that D. suzukii spent signifi cantly less time on treated fruit than on the control, which may be due to the toxic and deterrent effects of sulphur (Williams & Cooper, 2004). It is reported that powdered sulphur remains lethal for adults for seven days (not tested here), especially males confi ned in small experimental chambers (Pérez-Guerrero & Molina, 2016). Thus, it seems that powdered sulphur acts as a repellent and reduces the ability of D. suzukii to detect and ovipo sit on treated fruit. Furthermore, females in the choice experiment took signifi cantly longer to detect fruit treated with powdered sulphur than control fruit. Like most frugivorous insects, D. suzukii seems to depend on visual and olfactory cues for detecting suitable hosts for oviposition (Keesey et al., 2015;Little et al., 2018). In addition, some authors suggest that D. suzukii are mainly attracted by the colour red (Renkema et al., 2014;Kirkpatrick et al., 2016;Rice et al., 2016). Sulphur-covered fruit could affect these visual and olfactory cues, thereby hindering the fl ies ability to detect fruit. Lemoyne et al. (2008) report changes in the intensity (from darker to lighter) and hue of the colour (from indigo-blue to turquoise-blue) of blueberries treated with kaolin, which affects the ability of blueberry maggot (Rhagoletis mendax Curran) to detect and oviposit on blueberries. Sulphur did not signifi cantly affect how D. suzukii behaved on fruit. Lemoyne et al.'s (2008) study of how coming into contact with kaolin affects R. medax searching for blueberries report it did not affect cleaning behaviour and oviposition. According to these authors, oviposition in the choice experiment was little affected, which would explain why no signifi cant differences were recorded. However, it should be noted that they only oviposit on untreated fruit, which indicates that sulphur infl uences whether females consider fruit to be acceptable or suitable. This may account for the reduction in infestation reported here for strawberries and previously for blueberries (Pérez-Guerrero & Molina, 2016). Nevertheless, more extensive studies are needed to confi rm that powdered sulphur directly affects D. suzukii oviposition. Finally, it must be stressed that, since the exposure of fl ies to sulphur may vary greatly under fi eld conditions, the current experiment may be considered a best-case situation as sulphur completely covered the fruit in our laboratory experiment.
In conclusion, laboratory experiments revealed that powdered sulphur can signifi cantly reduce the infestation of strawberries, thereby confi rming previous results for blueberries. This sulphur formulation also affected the ability of the D. suzukii to detect fruit and the time they remained in contact with it; hence the lower probability of oviposition on treated fruit. Powdere d sulphur seems to be a suitable alternative to synthet ic-organic and neurotoxic insecticides. In addition, its use is compatible with organic crop management and waste reduction strategies. Field research is needed to confi rm the fi ndings of this study and determine how best t o use sulphur as an insecticide for protecting strawberries (where sulphur is currently used as a fungicide) and other berries.