Test subjects and experimental setup

During June-September 2019, late instar nymph crickets were purchased in two batches of 100, from Cricket Express (http://www.cricketexpress.se), and reared in a laboratory colony at 23 ± 2 °C temperature at Linköping University. Food (oat flakes, and pieces of soaked dog kibble) and water were given daily ad libitum . The colony was checked daily for sexually matured crickets, identified by the presence of wings (Horch et al., 2017). Matured crickets were removed from the colony and underwent 4 behavioural assays (described below). These tests were carried out in a set order on separate days, and within 7 days after maturation and removal from the colony. During testing, crickets were housed in individual containers in which the first 2 behavioural assays took place (transparent plastic containers, 9 × 16 × 10.5 cm), containing tissue paper covering the floor. After the first behavioural assay in this container, each cricket was given a shelter (a toilet paper roll, ca 13 cm long), and food and water. During behavioural testing, all crickets had auditory and olfactory, but not visual contact with other crickets. After testing, each cricket was weighed using an electric scale (to the nearest 0.001 g), and body length measured with a ruler (from head to wing tip, and to the nearest mm). To obtain a measure of body condition, body weight of each cricket was divided by its body length. Seventy-six crickets (38 males, 38 females) underwent all behavioural assays.

 

Behavioural assays

Behavioural assays were conducted between 9AM and 5PM in the same room as the crickets were housed. Artificial light was used during testing, otherwise, natural light was available through windows. Personality assays were conducted by three different experimenters and the threatening stimulus test by one experimenter.

 

Activity in a novel environment

Activity in a novel environment was measured immediately after each cricket was removed from the colony and placed individually in a novel container (containing only tissue paper covering the floor). The cricket was filmed inside the container for 3 minutes and locomotory activity was measured using the video recordings. Activity was measured by superimposing a grid of 10 equal quadrants covering the container floor on top of the video and counting the total number of times each cricket crossed between quadrants. A cricket was considered to have made ‘a crossing’ when one half of its body was over a grid line onto another quadrant. Activity videos were scored by A. H., C. G. D. and R. B. 

 

Activity in a familiar environment (general activity)

Activity in a familiar environment was recorded in the same way as activity in a novel environment, but after the novel environment had become familiar (Réale et al., 2007). This was done by filming each cricket for 3 minutes, in its individual container approximately 20-24 hours after being initially placed in it. Before filming, shelter, food and water were removed from the container so the cricket’s movement could be observed. Most crickets did not freeze during the removal of objects, but all crickets were nevertheless allowed 1 minute to acclimatize to the absence of the objects before recording started. This was enough for all crickets to resume movements, as seen by the experimenter turning on the camera. Again, the number of crossings between quadrants was used as a measure of general activity (called ‘activity’). Videos were scored by A. H., C. G. D. and R. B. 

 

Boldness

Boldness was measured using a shelter emergence test, a commonly used measure of boldness (Réale et al., 2007). Each cricket was removed from its individual container by gently trapping it inside its shelter, which was then placed in a white plastic arena (40 x 30 x 10 cm). To prevent crickets from emerging near the safety of the wall, one end of the shelter was taped to the wall, allowing crickets to emerge only into the centre of the arena. Each cricket was allowed 20 minutes to emerge from its shelter and was considered emerged after it had four legs out of its shelter. Crickets that did not emerge during the allowed time were given the maximum score of 1200 seconds. 

 

Consistency of behavioural measures

To estimate consistency of the behavioural measures obtained or, in other words, whether these measures described variation in personality, mature crickets from the same colony underwent two sets of personality assays (as described above). For activity in a novel environment, each cricket’s activity was measured in two different containers containing floor paper of different colour and texture (thus ensuring the arena was novel both times), 24 hours apart. For activity in a familiar environment, each cricket’s activity was measured in a familiar container 24 hours and 48 hours after being placed in it. For boldness, each cricket’s latency to emerge from a shelter was tested on two consecutive days, at the same time of day, returning the cricket to its own individual container in between tests. A different group of crickets was used for each three personality measures.

 

Response to repeating simulated predator attacks

To measure how crickets changed their behaviour when under attack by a predator, I conducted a test in which crickets experienced multiple successive threats simulating predator attacks. Each cricket was placed in the centre of a white plastic test arena (40 x 30 x 10 cm) with the floor covered by tissue paper to make it easier for them to walk. After the first head movement, each cricket was allowed 1 minute to acclimate and explore the arena. Following this minute, the first threatening stimulus trial was conducted by tapping the arena floor once, approximately 5 cm behind the cricket, with a pen. During this, the experimenter was sitting close to the table where the arena was located, taking care to not hover over the cricket and only move when conducting a trial. The latency a cricket remained frozen after this threatening stimulus, was measured, in seconds, from first becoming still to the first head movement. This procedure was repeated for 10 successive trials, and crickets were given 20 seconds to rest between the end of each freezing response and the next threatening stimulus. If a cricket did not move its head within 10 minutes after a stimulus, the arena was gently and slightly moved to encourage the cricket to move, and the cricket was given 20 seconds to rest before the next stimulus was given. Crickets that did not start moving within 10 minutes of a stimulus were given the maximum latency of 600 seconds for that trial. All crickets underwent 10 trials.

 

Data analyses

 

All statistical analyses were conducted using R software (version 3.6.1; R Development Core Team, 2019). Because data did not follow the assumptions for parametric data (normal distribution of residuals; visual inspection), non-parametric tests were used.

 

Consistency of behavioural measures

Consistency of behavioural responses were explored using Spearman rank correlation tests to analyse the correlation between the first and the second test occasion for activity, exploration and boldness. 

 

Exploration

My experimental measure of activity in a novel environment was likely to be influenced both by the cricket’s exploratory tendency (reaction to a novel environment) and general activity (tendency to move in general) (Réale et al., 2007), and therefore not a good measure of either single personality trait. Because of this, the experimental activity in a novel environment score was not used in the analysis. Instead, a measure of exploration was derived from the two activity measures by subtracting activity in familiar environment (general activity, hereafter the measure is referred as ‘activity’) from activity in a novel environment (exploration + general activity), thus excluding the effect of general activity level and only leaving the effect of the novel environment. I used this derived score in the analysis as a measure of exploration without the effect of activity. 

 

Co-variation among behavioural measures

To explore inter-relatedness among exploration, activity, and boldness, Spearman rank correlation tests were used.

 

Sex differences

To determine sex differences, Mann-Whitney U tests were used to compare male and female measures of exploration, activity, and boldness.

 

Effect of personality on response to repeated threatening stimuli

I explored the effect of each of the three different personality measures on the response to repeated threatening stimuli. The relationship between personality measures and the response to repeating threatening stimulus was investigated using generalized linear mixed effects models (GLMM) with Poisson error distribution (‘glmer’ function of ‘lme4’ package Bates et al., 2015). The length of the freezing response after threatening stimulus (in seconds, for trials 1-10), was used as the response variable, and ID was added as a random effect. To investigate the effect of personality on the change of freezing response over the 10 trials, Trial number (1-10) and the interaction terms Trial x Exploration, Trial x Activity, and Trial x Boldness were added as fixed effects. Males and females were analysed separately because of detected sex differences. Because models were over-dispersed, I included an observation-level random intercept to account for the extra variance in the residuals (Harrison, 2014). I included body condition and batch (June, September) as predictors. I simplified each model using the ‘drop1’ method (Bates et al., 2015), removing variables with nonsignificant effects until the minimal model with only significant predictors (i.e. p < 0.05) was reached. F statistics were obtained using ‘anova’ (Bates et al., 2015)

 

Additionally, to get a simple measure and visualisation of the change of freezing response, I calculated the rate of change (slope) of the regression line of each cricket’s freezing responses through the 10 trials using Excel. I used Spearman rank correlation tests to explore the relationships between the slopes and the personality measures.