============================================================================================================
START readme
============================================================================================================

This file focuses more on the details of the data package. 


============================================================================================================
## General
============================================================================================================

+ Author(s): Els M. van de Zande, Lina Ojeda-Prieto, Andreas Markou, Julia van Leemput, Joop J.A. van Loon, Marcel Dicke
+ Project: Data underlying the publication: Enhanced parasitisation of caterpillars and aphids on field-grown Brassica oleracea plants upon soil amendment with insect exuviae
+ Contact: els.vandezande@wur.nl OR marcel.dicke@wur.nl


============================================================================================================
## Title
============================================================================================================

Data underlying the publication: Enhanced parasitisation of caterpillars and aphids on field-grown Brassica oleracea plants upon soil amendment with insect exuviae

[ADD DOI Article]


============================================================================================================
## Methods
============================================================================================================

# Abstract

Multitrophic plant-insect interactions are mediated by plant volatiles. The emission of herbivore-induced plant volatiles is influenced by environmental conditions, such as soil microbes and nutrient composition, 
with consequences for aboveground trophic interactions. Here we investigated whether insect exuviae in the soil alter the plant’s volatile blend and attraction of parasitoids in the laboratory and whether this 
attraction also occurs in the field.
We studied the effects of soil amendment with exuviae originating from three insect species, Tenebrio molitor, Acheta domesticus, and Hermetia illucens, on the proportion of parasitised Plutella xylostella 
caterpillars and Brevicoryne brassicae aphids in the field in three consecutive years. In the laboratory, we collected and analysed the volatile blend of amended plants infested with caterpillars or aphids. The 
attraction of the parasitoids Diadegma semiclausum and Diaeretiella rapae, respectively, towards these volatile blends was assessed in an olfactometer.
Our study shows that insect exuviae amended soil enhanced the proportion of parasitised herbivores of two species in the field. Relative amounts of several components of the plant volatile blend were affected by 
soil amendment. Soil amendment with Acheta domesticus or Tenebrio molitor exuviae resulted in increased attraction of the two parasitoid species in the olfactometer.
Soil amendment with insect exuviae altered the plant volatile blend leading to enhanced attraction of parasitoids in laboratory assays. These effects were sustained under the complex and variable biotic and abiotic
conditions in the field. Our results underline the importance of belowground processes, such as the decomposition of insect exuviae, on aboveground volatile-mediated multitrophic interactions.




# Measurements and data collection

Soil treatment
The exuviae from three different insect species produced by commercial insect mass-rearing companies in the Netherlands (see publication) were used as soil amendments. These exuviae were inspected for the presence of insects and insect fragments, which were removed, dried at 60 ⁰C for 24 h and subsequently ground to a fine powder (SM 100 cutting mill, Retsch, Haan, Germany). To identify effects caused by adding nutrients or chitin 
when adding insect exuviae, a treatment with dried manure granulates (field experiment 2020 and 2021), or with mineral fertiliser (greenhouse experiments), and a treatment with purified shrimp chitin were included (field experiments 2020 and 2021). The amount of mineral fertiliser added was based on the release of NH4+ and NOx from the insect exuviae applied at 10 g/kg over time (Nurfikari, 2022) and was comparable with the 
nitrogen release by insect exuviae at 2 g exuviae per kg of soil in four weeks since amendment (Table 1). The different soil treatments were manually mixed through agricultural soil in bags of 10 kg soil. This agricultural soil was collected from the organically managed Droevendaal experimental field of Wageningen University (the Netherlands; 51.9899634° N, 5.6652231° E).

Study system
Seeds of Brussels sprouts plants (Brassica oleracea L. var. gemmifera cv. Cyrus) were obtained from Unifarm (Wageningen, the Netherlands). Three seeds were sown per pot and germinated in a greenhouse (21 ± 2 °C, 60-70 % RH and 16:8 h L:D), or outside under a protective shed (field experiment 2019). After germination two of the three seedlings were removed to retain one viable seedling per pot. Plants in the greenhouse were watered three times per week, by filling the saucer underneath each pot with tap water and allowing the soil to absorb it during 2 h.
Two insect species that commonly attack B. oleracea were used to induce defence responses in the plants: the leaf-chewing herbivore Plutella xylostella L. (Lepidoptera: Plutellidae), and the phloem-sucking herbivore Brevicoryne brassicae L. (Hemiptera: Aphididae). Both herbivores were reared on B. oleracea growing in unamended potting soil under greenhouse conditions (21 ± 2°C, 60–70 % RH and 16:8 h L:D) and were obtained from the stock rearing of the Laboratory of Entomology (Wageningen University, the Netherlands).
For the olfactometer essay, two species of parasitoids were used. Diadegma semiclausum Hellén (Hymenoptera: Ichneumonidae) was reared on P. xylostella-infested B. oleracea and Diaeretiella rapae McIntosh (Hymenoptera: Braconidae) was reared on B. brassicae-infested B. oleracea. Both parasitoid species were originally collected, and the rearing was supplemented with wild-catch once every year from agricultural fields in the vicinity of Wageningen University (the Netherlands). The colonies were maintained under greenhouse conditions (21 ± 2°C, 60–70 % RH and 16:8 h L:D).   Diadegma semiclausum individuals were collected as pupae and eclosed in a separate cage without plants or hosts under greenhouse conditions (21 ± 2°C, 60–70 % RH and 16:8 h L:D) and provided with honey-water (1:10) ad libitum. Two-day-old naïve D. semiclausum females were used in the olfactometer assay. Diaeretiella rapae eclosed in the rearing cage and were provided with honey-water (1:10) ad libitum. They were not naïve and between 1 and 5 days old at the start of the olfactometer assay.
 
Field preparation and after-care
Each year a different section of an organically managed ploughed and irrigated experimental site at the Droevendaal experimental fields of Wageningen University (the Netherlands; 51.9899634° N 5.6652231° E) was used for the field experiment. The experimental site has been used to grow various brassicaceous plant species since 2011, most recently black mustard (Brassica nigra).  Each year, three-week-old B. oleracea plants that were grown in agricultural soil with the different amendments were transplanted with clod in plots of five plants of the same treatment at a spacing of ca. 20 cm (Figure 1). Plots with different treatments were randomly alternated to prevent abiotic gradients or specific neighbour influences, and the middle plant of each plot was 2 m apart from the middle plants of neighbouring plots. In 2019 and 2021, the field was surrounded by a flowering edge of Brassica nigra plants to facilitate the naturally occurring parasitoid community. The field was fenced to keep larger herbivores out. During the first five weeks in the field, the plots were protected against herbivory by mesh tents (Ømesh: 0.1 mm, W: 30 cm, L: 30 cm, H: 90 cm), which were removed one day before the plants were manually infested with insects. Each field was split in two (2019 and 2020), or three (2021) batches, that differed in planting and harvesting time. For each soil treatment x herbivore combination, at least 10 plots were harvested per batch, resulting in 20 (2019 and 2020), or 30 (2021) plots per treatment per year (Table 2). No pesticides or fertilisers were applied in the field. Plants were irrigated regularly and the fields were weeded by hand after three weeks. In 2019 parasitoid recruitment was measured in early August (± 23 ℃, 8 mm precipitation) and early October (± 18 ℃, 10 mm precipitation), in 2020 this was done in early August (± 33 ℃, 4 mm precipitation) and late September (± 23 ℃, 6 mm precipitation), and in 2021 this was done in early July (± 20 ℃, 16 mm precipitation), early August (± 18 ℃, 14 mm precipitation) and late September (± 14 ℃, 20 mm precipitation). 

Parasitoid recruitment in the field
To measure the effect of soil amendment with insect exuviae on the proportion of parasitised P. xylostella caterpillars and B. brassicae aphids on B. oleracea plants in the field, we measured the natural parasitism of P. xylostella and B. brassicae as follows (Figure 2a): when the plants were eight weeks old (five weeks in the field), 15 L2-L3 instar P. xylostella larvae, or 30 B. brassicae adults were placed on a young leaf of the middle plant of each plot, equally dividing the soil treatments over the two herbivore treatments. The herbivores were exposed to the naturally occurring parasitoid community for three days, after which all five plants in each plot were harvested, by covering the plants with a plastic bag, cutting the stems as close to the soil as possible, and flipping the bag to gently slide the plants into the bag. A thin layer of the soil underneath the plant shoot was scooped into the bag as well to include herbivores that fell on the ground during the harvest. The bags were stored at 4 ⁰C until further processing. Within one week after harvest, all bags were searched for herbivores, which were collected in one Petri dish per plot and stored at -20 ⁰C until dissection. From the recollected herbivores, only P. xylostella caterpillars and B. brassicae aphids were dissected with the use of two needles and checked for the presence of parasitoid eggs or larvae using a stereomicroscope (Olympus SZ51 KL-300 LED). The parasitoid eggs or larvae were not identified to the species level. However, field studies at this location showed that the most common parasitoid of P. xylostella caterpillars is Diadegma semiclausum, and for B. brassicae this is Diaeretiella rapae (Bukovinszky et al., 2004). All aboveground plant material was dried at 105 ⁰C for three days to determine the dry shoot biomass per plot.
The proportion of parasitised herbivores was analysed for each year using a generalised linear model (GLM) or a generalised linear mixed model (GLMM) with a quasi-binomial distribution. We used a two-column matrix of the number of parasitised herbivores (column 1) and unparasitised herbivores (column 2) per plot as the response variable, soil treatment as the main factor, and batch as a random factor when contributing significantly to the model. Dry shoot biomass was analysed using a linear model (2019 and 2020), or a linear mixed model (LMM; 2021), including soil treatment as the main factor, and batch as a random factor. Models were validated by plotting residuals and, where necessary, homogeneity of variances and normality were confirmed using Levene’s test and the Shapiro-Wilk test, respectively. All statistical tests were performed using R (Version 3.6.3; R Core Team, 2022) and the packages car (Fox & Weisberg, 2019), emmeans (Lenth et al., 2023), nlme (Pinheiro et al., 2020), lme4 (Bates et al., 2015) and DHARMa (Hartig & Lohse, 2022). Where needed, the data was log10 or Tukey transformed to meet model assumptions. 

Attraction of parasitoid wasps using a Y-tube olfactometer
A Y-tube olfactometer (Steinberg et al., 1992) was used to investigate the odour preference of female parasitoids (Figure 2b) between host-infested control plants and host-infested plants growing in treated soil. The odour sources were 4-week-old B. oleracea plants grown in the differently treated soils (Table 1) and infested 48 h prior to the experiment with either 10 adult B. brassicae aphids (host of D. rapae), 5 L2 instar P. xylostella caterpillars (host of D. semiclausum) or left uninfested (control). The soil and pots of the plants were covered with aluminium foil, and the plants were placed in the Y-tube olfactometer set-up 10 min prior to the choice experiment. A flow of 2 L/min charcoal-filtered air was admitted into each arm of the system, and one female parasitoid at a time was introduced at the entrance of the central arm of the Y-tube. In total, 7 to 14 sets of plants were tested per treatment, and 10 parasitoids were tested per set of plants (Table 2). The glass chambers with the plants were switched after 5 parasitoids to compensate for unforeseen asymmetry of the setup. Each parasitoid individual was used only once and observed for a period of 10 min or until she made a choice. The behaviour was classified as “choice” when the insect crossed a predetermined point (10 cm from the junction with the central arm, 2 cm from the end of the arm) into either one of the arms. If the parasitoid did not reach the predetermined point in either of the arms within the 10-min time limit, it was recorded as “no choice” and excluded from the statistical analysis. 
The preference of the parasitoids for host-infested B. oleracea plants growing in treated versus untreated soil was analysed using a generalised linear mixed model (GLMM). This included the combination of treatments as fixed factor and the tested plant pairs as random factor. Because of the singularity effect of the random factor (intercept close to zero), the random effect was removed from the model and the statistical analysis was conducted with a generalised linear model (GLM) with a binomial distribution and zero intercept logit link (H0 = 50:50 distribution). This model used soil treatment as a fixed factor and a two-column matrix of the choice for either the infested control plant (column 1) or the infested plant grown in amended soil (column 2) as the response variable. Models were validated by plotting residuals. All statistics were performed using R and the packages lme4, car and performance (Lüdecke et al., 2021).

Collection and analysis of headspace VOC
To explain possible differences in the attraction of parasitoid wasps, volatile organic compounds (VOC) were collected from B. brassicae and P. xylostella-infested B. oleracea plants (Figure 2b) that grew in soil amended with black soldier fly exuviae, mealworm exuviae, mineral fertiliser, or untreated soil (Table 1). Unfortunately house cricket exuviae were not available at the time. Ten four-week-old plants per treatment (Table 2) were infested on the youngest fully developed leaf with 10 adult B. brassicae aphids or 5 L2 instar P. xylostella caterpillars, 48 h before VOC collection. Prior to VOC collection, all plants were watered with 50 mL tap water, and the soil and pots were covered with aluminium foil. The herbivores were kept on the plants during VOC collection. VOC were collected daily at 11 AM to avoid possible effects of the circadian rhythm on the VOC emission. The VOC of four plants were collected simultaneously by placing them individually in 28 L glass jars and letting the system acclimate for 25 min. The jars were supplied with a constant flow of charcoal-filtered synthetic air (nitrogen 80 %, oxygen 20 %; Linde, Ireland) at 330 mL/min via PTFE tubing. VOC were trapped on stainless steel thermal desorption sorbent tubes (TD tubes) filled with 200 mg Tenax TA 20/35 mesh (Markes international, UK) by using an air-sampling pump (SKC) equipped with an inlet-protection filter at a flow rate of 300 mL/min. TD tubes were removed after 180 min of volatile trapping, dry-purged with a 15-20 mL/min helium flow for 15 min and stored at room temperature until analysis. All treatments were replicated 9 - 11 times. To eliminate background volatiles, VOC were also collected from pots containing unamended soil covered with aluminium foil in the absence of plants. Mixed samples were collected by collecting volatiles from one plant per treatment for 20 min each, to obtain a range of possible VOC for the analysis. Collection of VOC was carried out in the laboratory at room temperature. Following the VOC collection, the plant leaves were scanned and fresh aboveground biomass was determined. Caterpillar damage was determined on the P. xylostella infested plants by dividing the leaf surface by the leaf surface eaten as assessed in Adobe Photoshop (Adobe Inc., San José, USA). 
The collected VOC were analysed by Thermal Desorption – Gas Chromatography – Mass Spectrometry (TD-GC-MS). VOC were thermally desorbed from the TD tubes by an automated thermal desorption unit (Ultra TD and Unity modules, Markes International, UK) separated on a gas chromatograph (Trace GC Ultra, Thermo Fisher Scientific, USA) and detected by a mass spectrometer detector (EI-Single Quadrupole Trace DSQ, Thermo Fisher Scientific, USA). The instrument was controlled with Thermal Desorption System Control Program (Markes International, UK) and Thermo Xcalibur 3.0 Qual (Thermo Fisher Scientific, USA) software. 
The collected VOC were desorbed from the collection tubes at 250 ℃ for 10 min and re-adsorbed on an electrically cooled sorbent trap at 0 ℃. Compounds were desorbed from this trap during the secondary desorption by ballistic heating at 40 ℃/s to 280 ℃ which was then kept for 10 min, while all the volatiles were transferred to a ZB-5 MS analytical column (30 m × 0.25 mm ID × 1 mm F.T.) with 10 m built-in guard column (Phenomenex, Torrance, CA, USA), placed inside the oven of a Thermo Trace GC Ultra (Thermo Fisher Scientific, Waltham, MA, USA) for further separation of the plant volatiles. The temperature of the GC was initially held at 40 °C for 2 min and was then raised at 6 °C min−1 to a final temperature of 280 °C, where it was kept for 4 min under a constant helium flow of 1 mL min−1. A Thermo Trace DSQ quadrupole MS (Thermo Fisher Scientific) coupled to the GC was operated in electron impact (EI) ionisation mode at 70 eV in a full scan with a mass range of 35–400 amu at 4.70 scans s−1 . The MS transfer line and ion source were set at 275 °C and 250 °C, respectively.
The MetAlign-MSClust software pipeline was used to analyse the data (Lommen, 2009; Tikunov et al., 2012). Basically, MetAlign eliminates the noise and corrects the baseline of each GC-MS output file, using the background and mixed samples, and aligns the individual mass peaks in all files. Subsequently, MSClust clusters the aligned mass peaks to construct mass spectra of putative compounds. Only mass peaks with a retention time within 5 - 38 min and in the 55 - 400 m/z range were further processed. Additionally, only VOC whose peak intensity statistically differed from the background (Student t-test, α = 0.05) were included as plant headspace VOC and further processed. Peak intensity of the VOCs was not correlated with fresh shoot biomass, or with caterpillar damage on the P. xylostella infested plants (Pearsons product moment correlation, r2 < 0.02) and was thus not normalised for fresh shoot biomass, or herbivore damage. VOCs were provisionally annotated by comparing their mass spectra and their AI values, as they were experimentally determined by Moisan et al. (2020), with those of the reference Wageningen Mass Spectral Database of Natural Products and NIST library (National Institute of Standards and Technology). For each putatively identified VOC an individual aligned mass peak with a minimum peak height of 9.95 x 107 counts across all samples was selected. The intensity of the selected mass peak across samples relates to the abundance of the compound. 
The average and standard error of peak intensity values were calculated per treatment and are reported in Table S3. Differences in peak intensity between treatments for each VOC were statistically analysed for each VOC using ANOVA, or the nonparametric Kruskal-Wallis test (𝛼 = 0.05). Peak intensities were log-transformed and autoscaled (mean-centering and unit-variance scaling). Subsequently, a heatmap was plotted for each herbivore treatment, indicating the peak intensity of each plant VOC per soil treatment. To separate the blends of VOC emitted by the plants grown in differently amended soils, a partial least squares-discriminant analysis (PLS-DA) was performed. Differences in the volatile blend between the plants grown in differently amended soils were statistically tested using PERMANOVA (𝛼 = 0.05, H0 = centroid and spread of data points are equivalent for all treatments). All statistics were performed using R and the packages car, ropls (Thévenot et al., 2015), PERMANOVA (Vicente-Gonzalez & Vicente-Villardon, 2022) and pheatmap (Kolde, 2022).


============================================================================================================
## FolderStructure
============================================================================================================

"There is only one parent directory present containing all files."


============================================================================================================
## FolderContents
============================================================================================================

- parent_folder/
Y-tube_raw_data
Field_raw_data
VOC_SelectedCT




============================================================================================================
## Software
============================================================================================================

SoftwareRequirements: 

Excel


============================================================================================================
## FileFormats
============================================================================================================

 .xlsx


============================================================================================================
## CodeBook
============================================================================================================

The Y-tube_raw_data.xlsx uses the columns:

+ Date = The data on which the measurement was performed
+ Treatment = the code used for the different soil amendments 
+ Replicate = replicate number of the two plants that were used in the Y-tube
+ wasp = replicate number of the individual wasp
+ Measurement_ID = Code indicating which treatment and replicate were measured
+ Plant0 = The plant for which the choice of the wasp is scored as 0
+ Plant1 = The plant for which the choice of the wasp is scored as 1
+ Time(min) = the time in minutes that the wasp needed to make a choice
+ Score = The choice of the wasp scored as 0 or 1

The file contains three sheets. The first is the choice of wasp between aphid infested plants, the second is the choice of wasp between caterpillar infested plants and the third sheet depicts the code used for soil amendment (treatment) and herbivores.

The Field_raw_data.xlsx uses the colums:

+ Field = The field number of the different fields that were measured in that field season
+ Treatment = the code used for the different soil amendments
+ Replicate = replicate number of the plot that was measured
+ Plot = Plot number of the plot that was measured, consisting of the treatment and replicate number
+ Herbivore = number indicating the herbivore with which the plants were infested
+ Parasitised = the number of parasitised herbivores on the plot
+ Unparasitised = the number of herbivores on the plot that were not parasitised
+ No_herbivores = the total number of herbivores recollected on the plant
+ No_leaves = the total number of leaves of all the plants on one plot at time of harvest
+ stem_width = the avergae stem width of the plants on one plot at the time of harvest
+ dry_weight = the total dry shoot biomass of all the plants on one plot at the time of harvest

The file contains 5 sheets. The first two sheets contain the information of the field season in 2019, the first on the parasitisation and number of recollected herbivores and the second sheet on the plant measurements at the time of harvest. The third sheet contains all the data of the field season in 2020, and the fourth sheet contains all the data of the field season in 2021. The fifth sheet depicts the code used for soil amendment (treatment) and herbivores.

The VOC_SelectedCT.xlsx uses the columns:

+ sample = code of the sample taken from the headspace of differently treated Brussels sprouts plants
+ Treatment = the code used for the different soil amendments 
+ Herbivore = letter indicating the herbivore with which the plants were infested
+ Treatment_herbivore = combination of the treatment and herbivore code
+ Biomass = fresh shoot biomass (g) as measured directly after VOC collection
+ Leaf damage (number of aphids or loss of leaf surface) = a measure of leaf damage. measured either as number of recollected aphids, or loss of leaf surface.
+ CT = centrotype number
columns G - AE = number codes, followed by compound names of the measured and selected VOC

The file contains 2 sheets. The first depicting the results of the VOC analyses and the second depicting the code used for soil amendment (treatment)




============================================================================================================
## Other
============================================================================================================

[describe any other attention points that will help understandability of your data package; delete this 
explanation]


============================================================================================================
END readme
============================================================================================================
