Assessment of the Positive and Negative Effects of Earwigs in Apple Orchards

(i) Project Sites – One Gala and three Fuji commercial orchards were selected for study. They were located between Orondo and Royal City, WA. All were modern, high-density plantings on dwarfing rootstocks with a history of woolly apple aphid infestation.

(ii) Objective 1: Test for the effect of earwigs on aphid abundance and apple fruit damage

Sub-objective 1a: manipulating earwig densities – In May 2017, we divided each of the 3 Fuji orchard blocks (Fuji 1, Fuji 2, and Fuji 3) into 12 sections consisting of 8 trees in each of 2 adjacent rows. In each block, we designated four sections as “earwig release areas,” four as “earwig removal areas,” and four as “control areas” (4 orchards x 3 treatments x 4 replications = 48 total sampling areas). From May until November 2017, we counted earwigs in all sections using rolled strips of corrugated cardboard. Earwigs naturally move towards tight spaces, away from light, and are attracted to other earwigs, thus they aggregate in these cardboard shelters. One shelter (a 7.5 cm by 35.5 cm rolled strip of corrugated cardboard) was placed on each tree, and earwigs were shaken out into a bucket and counted weekly. In the earwig removal areas, all earwigs found were placed in plastic bags and transported to the laboratory. In the control areas, earwigs were counted and released onto the orchard floor. In the earwig release areas, earwigs were counted, released, and earwigs collected from our earwig removal areas and a different orchard were added. Between 31-May and 11-Jul, a total of 215 earwigs were released per tree at each augmentation section in Fuji 1, 140 per tree in Fuji 2, and 230 per tree in Fuji 3 orchards.

The Gala orchard only had 10 control sections, monitored as above, but in 2016, five of these sections were treated as removal and the other five as augmentation sections. The augmentation sections at the Gala orchard in 2016 received ca. 46 earwigs per tree over the year.

Sub-objective 1b: monitoring aphids – In each sampling area, we counted woolly apple aphid colonies on leaf axils of 10 twigs per tree, and all woody branch and trunk material (16 trees per sampling area) weekly from May until November 2017.

Sub-objective 1c: monitoring fruit damage – During the harvest period in 2017, we inspected up to 30 apple fruits per tree in all earwig removal areas and earwig release sections at Fuji orchards 1 and 2, and all sections of the Gala orchard. Types of damage were recorded as: splits on sides of apples, depressions (limb-rub symptoms), russet patches, rounded holes on the sides of apples, irregular holes on the sides of apples, bitter bit symptoms, stem bowl splitting, and stem bowl splitting with rounded holes. Splitting was defined as an oblong exposure of internal tissue.

Objective 1 statistical analysis – To assess the effect of our earwig abundance manipulations on earwig counts, we first calculated cumulative insect-days for the number of earwigs per shelter at each section of each orchard. Cumulative insect-days represent the amount and duration of insect occurrence in one expression as: ∑(D_{i + 1} – D_i)[(Y_i +Y_{i + 1}))/2] where Di and Di+1 are adjacent observation time points (days) and Yi and Yi+1 are insect counts on those days. We then compared cumulative earwig-days within each site by ANOVA, with an explanatory variable of treatment, and response variables of cumulative earwig-days and cumulative woolly apple aphid colony-days. For the three Fuji orchards, which had three treatment types (earwig augmentation, control, and earwig removal), we followed ANOVAs with linear contrasts to compare each augmentation and removal treatment to the control treatment within each site.

To more directly quantify the relationship between earwig abundance and woolly apple aphid abundance, we first divided cumulative insect-days of woolly apple aphid colonies and of earwigs by the number of days each section was monitored, to convert the values to a more relatable metric, essentially representing season-wide average insect counts. Then, for each orchard (and year for the Gala orchard), we plotted woolly apple aphid values by earwig values of each section and fitted a negative exponential function: Y = h + a*e^-bx. We obtained maximum likelihood coefficients for this model using the ‘mle2’ function, and compared the support to a linear model (Y = s*h) and a null model (Y = h) using Aikake’s Information Criterion corrected for sample size (AICc) with the ‘AICctab’ function in the R package ‘bbmle’.

To assess the effect of earwigs on fruit damage, we totaled the number of apples in each damage category into two groups: apples from earwig augmentation areas and apples from the earwig removal areas (from Fuji orchards 1 and 2) and and control areas (at the Gala orchard). We then compared incidence of each damage category between the two groups with Fisher’s exact tests.

 (iii) Objective 2: molecular analysis of earwig stomach contents

Sub-objective 2a: earwig collections – At each of the four sites (see Objective 1), we collected a sample of 20 earwigs within an hour of sunrise at four time points from 28 June to 3 October.

Sub-objective 2b: detection of woolly apple aphid DNA in earwig guts – We dissected earwig stomachs and extracted DNA contents from each using a QIAGEN DNeasy extraction kit. Using woolly apple aphid specific primers tested and designed during this study (forward primer: 5’-GGAATAATTGGTTCATCCTTA-3’; reverse primer: 5’-CTACAAATTATTATTATTAAAGAAGGG-3’), we tested each extraction for presence of woolly apple aphid DNA with a simple polymerase chain reaction assay.

Sub-objective 2c: earwig diet characterization with next-generation sequencing – First, for each earwig collection we pooled DNA into a single unit (4 sites x 4 earwig collections = 16 samples). Pooling was necessary because the cost of sequencing individual stomachs was too high. These samples were sent to a commercial laboratory, RTL Genomics (Lubbock, TX), for sequencing of arthropod, plant, and fungal DNA contents. Taxonomic assignments to resulting DNA sequences were made by the RTL Genomics laboratory following their standard methods.

(iv) Objective 3: quantify apple orchardist experiences with earwigs through in-depth interviews

Sub-objective 3a: interview apple industry professionals – we conducted recorded interviews with 16 apple industry professionals. They were a mix of of private owners and consultants agreed to in-depth interviews about their perceptions and experiences who were solicited in 2016 at extension and industry meetings in central Washington.

Sub-objective 3b: transcribe and analyze interviews – To analyze interviews, we first transcribed all audio to text. We then organized all relevant text into a table matrix by topic. In two areas, we summarized responses into categories. First, in interviewees’ “overall strategy for woolly apple aphid management”, perceptions on the role of biological control in general, and of each natural enemy, were categorized as either: ‘important’ (having a significant role), ‘mixed’ (having a role, but with uncertainty of its importance, or having an inconsistent role) ‘don’t know’ (lack of knowledge of its existence, or uncertainty that a potential role exists), or ‘not important’ (a confident assertion that there is no role). Second, we scored whether each interviewee had ever observed earwigs in damaged fruit, whether they had ever been in involved in a decision to chemically suppress earwigs, and whether they considered earwigs a pest. We rated interviewees’ importance of earwigs as a pest as ‘yes’ (caused significant economic damage), ‘slight’ (caused economic damage, but there was no concern about it), or ‘no’ (has not caused economic damage).