Current control of the invasive pest Drosophila suzukii relies primarily on insecticides, including the incorporation of phagostimulant baits. The impact of insecticidal bait sprays on beneficial insects in crops is largely unknown. Using a laboratory and field trial, we exposed non-target insects to insecticides with or without bait and compared these to non-insecticide controls. In laboratory arena tests, we assessed the impact on mortality. In the subsequent commercial raspberry field trial, 1 m width spray bands of bait were applied weekly, using alternating 25% or 50% field rates of spinosad and cyantraniliprole respectively.

Results from the laboratory assays separated the insects into three categories: 1) adult Eupeodes corollae, Forficula auricularia, and Orius laevigatus, 2) adult Drosophila melanogaster, and 3) larval Chrysopa sp. and Adalia bipunctata. In the first group, bait + spinosad or spinosad alone were equally detrimental to the life expectancy of insects. For D. melanogaster, bait + spinosad was faster acting than spinosad without bait. No detrimental impact of the treatments was observed on the third group of larval predators. Baits alone did not increase insect mortality. In the raspberry crop there were no observed impacts of baits with insecticides on the abundance of insect pollinators or natural enemies compared to the full foliar applications of insecticides.

This study is the first to test a range of non-target insects for toxicity to low dose insecticides combined with baits. Further field testing in commercial crops should explore the placement of bait droplets for optimal D. suzukii control, whilst minimising further impacts on non-target insecticides.

Two main experiments were done, the first on six non-target insects in a laboratory bioassay evaluating the toxicity of baits alone or in combination with an insecticide (spinosad) in comparison to the insecticide alone, or controls without insecticides. The second was a field trial in a commercial raspberry crop comprising a programme of alternating insecticides (spinosad and cyantraniliprole) as either a full foliar spray or as a reduced rate in combination with a bait spray. Impacts on non-target insects were assessed.

Laboratory Bioassays

Six non-target insects commonly encountered in fruit crops were tested (Table 1): 1) Orius laevigatus (Fieber) adults, a commonly used predatory biocontrol agent for the control of thrips; 2) Forficula auricularia L. adults, a common generalist predator of orchard pests; 3) Eupeodes corollae Fabricius adults, pollinators with larvae that predate aphids; 4) Drosophila melanogaster Meigen adults, common fruit fly from a wide range of habitats that contributes to organic matter degradation and is closely related to D. suzukii; 5) Adalia bipunctata (L.) larvae, and 6) Chrysoperla sp. complex (Stephens) larvae common natural predatory insects in fruit crops, feeding primarily on aphids.

Two baits were used in the laboratory experiment: 1) Combi-protec®, a proprietary mixture of plant extracts, proteins, and sugars (Dedetec, Freiburg, Germany), and 2) ProBandz®, a natural product containing 54-58% sugars (Russell IPM, Deeside, Flintshire, UK). The bait solutions were prepared according to the product labels.

Perspex boxes (22.5 x 12 x 8 cm) with ventilated lids (Plain Leaded Net Curtain Fabric Dunelm, 1 mm mesh size) contained two 90 mm discs of filter paper (Whatman®, 25 × 20 mm), dampened with 1 ml of distilled water. On each filter paper disc, there was a 50 mm Petri dish containing two freshly picked bramble (Rubus fruticosus agg.) leaves (approximately 30 mm width x 40 mm length). One dish had a sugar feeder (2.1 mL Graduated Transfer Pipette bulb, Samco™) containing 5% sucrose solution and a sponge cloth wick. The second dish had an alternative food source, provided ad libitum, to optimise insect life expectancy (Table 1).

There were six replicate arenas for each species with five treatments for each, except for O. laevigatus, which had an additional treatment (ProBandz + spinosad; Table 2). The insecticide, ‘Tracer’ (a.i. spinosad 480 g L^-1, Dow,) was compared either with or without baits to water only and bait only controls (Table 2). Each bramble leaf had six 10 μl droplets, three on each side of the central leaf vein (12 droplets per box total). Baits (Combi-protec or ProBandz) were used at 5%v/v. Spinosad was used at a concentration of 0.02% w/w, equivalent to the concentration if applied at the label rate of 96 g ha^-1 in a spray volume of 500 L. For O. laevigatus, ProBandz was also used with spinosad at 0.0016% w/w, equivalent to the concentration if applied at 8% of the label rate.

Finally, the non-target insects being tested were introduced into the centre of each box. The ventilated lids were sealed with tape on to the boxes and then incubated in a growth chamber at 23 °C, 60% RH, 16:8 L:D.

Table 1. Non-target insects and their quantity, life stage, and alternative food source. All insects had a 5% sucrose solution feeder. There were six replicate arenas for each insect species. The Nutrimac™ product contained Ephestia eggs. Dried cat food was from www.iams.com/cat/cat-food.

Table 2. Laboratory bioassay treatments, active ingredients, and product concentrations. The percentage approved field rate of spinosad was 50% for treatments 1 and 4, and 8% for treatment 5. *This treatment was only tested for Orius laevigatus.

Numbers of individual insects, alive or dead, were recorded at regular intervals depending on the life expectancy of each species (Table 1). Insects were gently prodded with a small artists paintbrush and deemed alive if there was any movement. Missing insects, presumed consumed by surviving predatory larvae, reached up to 50% by 120 h and 144 h for A. bipunctata and Crysoperla sp. larvae respectively and were considered as dead for analyses.

Commercial Field Trial

A replicated field trial was set up at a soft-fruit farm in Kent in two tunnels of a commercial Primocane raspberry (Rubus idaeus L.; cv. Majestic) crop. Plots were divided by a barrier of insect exclusion mesh (Gromax industries ltd, Gro-Net AA/6, hole size: 0.8 x 0.8 mm) clipped to the steel tunnel hoops. Mesh was used to limit the movement of D. suzukii and beneficial insects between treatment plots and to reduce spray drift between compartments. There were three sections of insect exclusion mesh falling on either side and in between the posts and wire systems that supported the raspberry canes. Mesh was also deployed along the leg rows in-between the tunnels. The mesh was secured with sandbags on the ground to allow easy movement of field staff and researchers for spray applications and assessments. The compartments were erected two weeks before the first treatment application when the first pink fruit was visible on the crop. Once the fruit harvest commenced, the treatment applications began.

There were 4 replicate plots of 5 treatments (Table 3) distributed along the length of two 119 m polytunnels. Each 6.9 x 8.9 m (w x l) plot consisted of two rows of potted raspberries spaced at 3 m apart (15 pots in each row, 3 canes per pot). There was no untreated control as this was a commercial site and we wanted to avoid the risk of D. suzukii building up in the crop.

Table 3. Commercial raspberry field trial with alternating treatments of Tracer (a.i. spinosad) and Exirel (a.i. cyantraniliprole) at different field rates applied with and without Combi-protec (5% v/v) bait.

 Alternating sprays, beginning with Tracer, were applied on 07/09/2022, 14/09/2022, 20/09/2022, and 26/09/2022 at the crop fruiting stage, with a motorized knapsack with a handheld blower (Birchmeier REC 14 ABC, nozzle pressure 3 bar). For the first three treatments (Table 3) sprays were applied as a full foliar application at 500 L ha^-1 (Albuz ATR 80 orange nozzle at 10.72 ml sec^-1). The other two treatments combined with 5% Combi-protect bait were applied as a 1-meter band at 40 L ha^-1 (Lecher IDK-120-015 green at 13.68 ml sec^-1) across the centre of the crop canopy, which was 2.1 m high. Treatments 1, 2, and 3 (full field rate foliar) and treatments 4 and 5 (low rate with bait sprays) applications took 223 seconds and 22 seconds per plot, respectively. Spray applications measured from the start and end tank volumes were 87% to 105% of the target values, respectively.

A preassessment for pollinators and a tap sampling survey for natural enemies was done before treatment application on 26/08/2022. Surveys consisted of a 1-minute (30 seconds along each side of the central-facing side of the row i.e., the middle of the tunnel) crop walk in each plot, recording beneficials contacting the crop (landing on fruit, flowers, and foliage). For tap sampling, a white tray (46 x 36 x 3 cm) was held beneath the lateral stems of the crop, and then the corresponding canes were struck with a beating stick three times. The invertebrates that fell onto the white tray were counted. Six tap sample points were taken in each plot (three evenly spaced on each side of the central-facing side of the row). Both visual observations and tap samples were taken from a 1 m horizontal section on the canopy, relative to where the bait spray bands were applied in both full foliar and bait spray-treated plots.

The post-treatment application insect surveys were done on 20/09/2022 (6 days after first 2 treatments), 03/10/2022 (7 days after second 2 treatments), 05/10/2022 (9 days after second 2 treatments) and 18/10/2022 (22 days after second 2 treatments). These followed the same methods as the preassessment described above for pollinator counts and tap sampling.

During the experiment, the average temperature was 16.3ºC (range = 36.5 ºC and 3.0 ºC). The average relative humidity was 80.2% (range = 95% and 35.5%).