Rows were oriented along a north-south direction on an east facing slope. Three treatments , were included with ~0.056 ha plots. No pesticides were applied during the experimental period. Here, there were 28 GUM and buffer plots each and 18 UTC plots. GUM dispensers were applied on 10 October and ten berries were collected from each plot on this date. Sampling dates were 11, 14, and 18 October 2019.Trials were conducted in a commercial sweet cherry orchard located at the Mid-Columbia Agricultural Research and Extension Center , Hood River, Oregon, USA. A 1.12-hectare orchard was divided into twelve plots . UTC, buffer, and GUM plots were replicated four times. The GUM dispensers were deployed on day 0 . No insecticides were applied to the orchard for the duration of the experiment. Here, an additional 200 mated 8- to 12-day-old D. suzukii were released in the center of each plot on a weekly basis on 23 June, and 1, 8, and 15 July 2020 . Data were collected for 35 days from 16 June through 22 July 2019. Because of relatively large canopy size of cherry trees compared to the other crops, ten cherries were collected from the lower , middle , and upper portions of the central two trees in each plot weekly.This trial was conducted in Oxnard, California, USA on high bush blueberry plots during 2020. Plants were irrigated with three drip stakes per plot ten times a day for ten-minute intervals delivering 1.1 liters of water per hour. Screen houses were fully enclosed with screen material to prevent insects from entering.
There were three 70 m x 5 m screen houses with GUM or UTC treatment randomly assigned to the north or south end of each screen house for a total of 6 plots. Within each screen house, round plastic pots treatment plots contained twelve plants in two rows, and plots were separated by 45 m. one-hundred flies were released in each plot four times, once per week. Three GUM deployment plots were compared with three UTC plots. GUM dispensers were installed in every other plant with irrigation stakes placed directly through the pads. The GUM application was completed on 14 April. Plots were sampled every seven days from 14 April to 12 May. One sample consisted of 50 berries.Ten field trials were conducted from September to November 2020 across multiple coastal production regions in California, USA , at different ranches and on multiple varieties being grown under high tunnels. Each location was a replicate consisting of two plots and were randomly assigned at each ranch to GUM or to UTC. Plots within a ranch received similar irrigation, fertilizer, and insecticides. Each plot received a minimum of four spinosad sprays timed 7-10 days apart during the cropping period and based on monitoring trends from fruit collections. Additional peroxyacetic acid applications were applied at 2-3 day intervals after each spinosad application, followed by a C. subtsugae application 1-2 d after each peroxyacetic acid application. Throughout the experimental periods, GUM dispensers were distributed evenly throughout each plot and replaced every 21 days. GUM dispensers were staked directly under the drip line in soil plots, and irrigation stakes were placed directly through the dispenser in substrate plantings. Six fruit samples were collected from each treatment plot every week for 4 to 12 weeks.
Samples were collected at least 2 m from each edge of the tunnel as well as from the center of the tunnel approximately 20- 30 m from the edge of the tunnel and at ~0.75 m from the ground. Each sample consisted of 50 berries. Sample berries were incubated at room temperature for 2-4 days to allow for larval growth and facilitate detection. Samples were evaluated by crushing fruit and submerging them in a saltwater solution . The crushed fruit solution was then poured into a tray where D. suzukii larvae subsequently floated to the top of the solution and were counted.For Oregon blueberry cultivar 1, egg counts were mostly zero. Among 864 total records of three treatment levels, no eggs were found in 825 records, while eggs were reported only in 39 records. To address the excessive zeros in the data, a dummy variable was created to indicate the existence of eggs, and a logistic regression was used to model the infestation rate for each plot, namely the probability of each plot having eggs. In the logistic regression model, treatment, sampling date, treatment and sampling date interaction, and quadratic term of sampling date were included as explanatory variables. The quadratic date effect was included since it provided a better model fit with smaller AIC/BIC value. The sampling date and its quadratic term were treated as continuous variables instead of fixed effects. This was because there were 24 sampling dates in the experiment and treating it as a fixed effect would increase model complexity and result in a slightly overfitted model according to Akaike Information Criterion . A random effect was not included since the mixed logistic model failed to provide reliable inference due to the excess zeros in the egg counts. Furthermore, a firth correction was specified using FIRTH option in the MODEL statement to account for the imbalance of zero and nonzero counts observed in the data.
Multiple independent trials were conducted to test the treatment effect. To effectively combine results from multiple analysis and increase the generalizability of the analysis results, we employed Stouffer’s method to combine a total of 10 pvalues from trials testing the difference in mean number of larvae in terms of GUM vs UTC, specifically those from California tails, including raspberry, blackberry, and strawberry . The null hypothesis for the meta-analysis is that all of the individual null hypotheses are true, indicating no significant difference between GUM and UTC, while the alternative hypothesis is that at least one of the individual alternative hypotheses is true. The test was implemented using the “stouffer” function in the R package “poolr.The buildup of D. suzukii populations was modeled under four scenarios i.e.; no intervention , GUM only, insecticide , and GUM and insecticide . The model parameters were obtained from experimental work and iterations of the model have been used in previous studies . Recorded D. suzukii population levels and weather data were used as model inputs. Outputs from the model were directly compared with D. suzukii infestation data of the blueberry field trial 1 . This trial was selected because of its relatively long duration and is most suitable for describing population build-up. Ambient temperature influences the fecundity rates, mortality rates, and maturation delays of the four principal life stages . The simulations were based on daily mean temperature data recorded at Aurora, Oregon, USA, between June and September 2020. We assumed that the flies had access to unlimited fruit and that no other factors affected population dynamics . Parameter values, including for fecundity rates, mortality rates, and maturation delays, were obtained from laboratory experiments on blueberry . The simulations were initialized on 30 June with a population composed equally of adult males and females. The model simulations track relative population densities, hydroponic bucket and the initial adult density was chosen so that the simulated egg density matched the final eggs/berry in the UTC treatment. The GUM dispensers were assumed to reduce D. suzukii fecundity by 49%, according to data from Tait et al. . Insecticide induced mortality rates caused by GS were calculated from laboratory data . Spinosad was used as insecticide model. Details on the model and on how the GS and GUM treatments were implemented can be found in the Supplementary Material. The simulations were implemented using Wolfram Mathematica 13.0 . The code for the simulations is available online1 .The current study supports findings from previous laboratory and small-scale field cage trials. Here we show through fieldcollected and modeled data that food-grade gum use can reduce D. suzukii fruit damage . The aim of this work was to acquire detailed knowledge about limitations of food-grade gum in a range of commercial cropping systems including blueberry, blackberry, cherry, raspberry, strawberry, and winegrape. These studies were conducted in two key production regions i.e., California and Oregon in the USA. The overall results supported initial findings and provided additional evidence that this tool can reduce D. suzukii crop damage especially when applied together with the grower standard. Both field-collected data and model simulations indicates that there is a synergistic effect of food-grade gum when used in combination with a conventional insecticide. For most of the experiments , field plots receiving the food-grade gum resulted in either numerical or statistical differences in D. suzukii damage compared to untreated control plots. This was not recorded for the cherry, strawberry, and blackberry trials. Reasonable hypothesis about these data are discussed below. In trials where D. suzukii infestations were measured in buffer plots , there was evidence of a reduction in damage, but not at the same level as in plots treated by the food-grade gum. Overall, considering all the trials, crop damage was reduced up to 78% over a period of up to 21 days post application of the food-grade gum. The results from the current study indicate that the food-grade gum can be used in combination with standard insecticides , and in some cases as a stand-alone treatment to reduce the infestation level of D. suzukii.
Similar reductions in D. suzukii damage were reported under laboratory and controlled semi-field conditions , suggesting that the food-grade gum resulted in lower damage due to oviposition. These findings support earlier results where the effects of semiochemical volatiles emanating from the food-grade gum resulted in significant behavioral changes . In several trials, data lower oviposition and fruit infestation in the presence of the food-grade gum under field conditions. Reasons of why in multiple trials a statistical difference was not reached, can be explained by multiple parameters observed by scientists and growers such as animals removing the cottons pads, water-irrigation issues, and wind. These factors are addressed in a future publication . In the Hood River cherry trial, constant windy conditions may have resulted in dispersion of volatiles, ultimately resulting in impacts that were less pronounced. There is little doubt that efficiency of the food-grade gum can vary depending on production conditions and crop . Host preference of D.suzukii was ranked 4th for cherry, followed by blueberry and winegrape . Such differences in host preference should be considered when applying food-grade gum. Synthetic blends can be less attractive compared to the actual fruit; thus, additional adjustments may be required to minimize egg-laying in the fruit. Results showed that the application of the food-grade gum in grape shows clear impacts to protect berries from D. suzukii attack. Considering the vulnerability of several winegrape cultivars towards D. suzukii and the encouraging results collected, we have reasons to believe that the food-grade gum can be a useful tool for the winegrape production. For the food-grade gum applications in blueberry in open field experiments, the infestation rate for the food-grade gum and grower standard were 70% and 85% lower than that for untreated control respectively, with the food-grade gum treatment resulted in a significantly lower infestation rate compared with the control. Open and semi-field experiments conducted in California provide similar outcomes to those in Oregon. Blueberry experiments conducted in California within a screen house provided 45.5% egg reduction. There were sequential applications with differing timing and the results indicated that early applications resulted in lower egg reductions . A potential hypothesis for this phenomenon could be related to environmental conditions including temperature and humidity that could significantly change the emission of plant volatiles . Egg reduction in raspberry and blackberry varied from 42-90% and 24-70% respectively. Two cultivars of raspberry have been subjected to the trial and in both cases there was reduction in egg infestation. For blackberry the same cultivar has been evaluated but in three different farms. Results were consistent between the different locations. For strawberry, in several cases results showed numerically increased larval levels compared in the food-grade gum treatments. A potential hypothesis for this phenomenon could be related to either unreported production practices or environmental conditions that could significantly change the emission of plant volatiles or the food-grade gum. Other reasons that can justify the negative results, range from lack of irrigation to rodents removing food-grade gum within a day of placement .