Content uploaded by Fabian S Rudin
Author content
All content in this area was uploaded by Fabian S Rudin on Mar 20, 2018
Content may be subject to copyright.
The effects of the social environment and physical disturbance on
personality traits
Fabian S. Rudin
*
, Joseph L. Tomkins, Leigh W. Simmons
Centre for Evolutionary Biology, School of Biological Sciences, The University of Western Australia, Crawley, WA, Australia
article info
Article history:
Received 25 July 2017
Initial acceptance 6 September 2017
Final acceptance 29 January 2018
MS. number: 17-00598R
Keywords:
behavioural plasticity
disturbance
indirect genetic effects
interacting phenotypes
personality
social environment
The environment can have a considerable impact on behaviour. The social environment is predicted to be
a particularly important driver of behavioural variation and evolution through the indirect genetic effects
that arise whenever individuals interact with conspecifics. We used male Australian field crickets, Tel-
eogryllus oceanicus, to examine the effects of changes in the social environment (recorded acoustic sexual
signals of other males) on the expression and consistency of boldness, activity and exploration, and their
between-individual covariation. Switching from a silent environment to being exposed to male acoustic
sexual signals resulted in crickets becoming less bold, active and explorative. Switching from an acoustic
to a silent environment resulted in increased boldness and activity. We also looked at the effects of
changes in the nonsocial environment via a physical disturbance that mimicked the presence of a po-
tential predator (mechanical shaking). The effects of physical disturbance (and changes thereof) on
behaviour were far less pronounced than the effects of changes in the social environment. Neither the
repeatability of nor correlations between behaviours were affected by changes in physical disturbance.
Only the average level of exploration was affected significantly when crickets were moved from an
undisturbed to a disturbed environment, with crickets becoming less explorative. Although changes in
the social and the nonsocial environment affected the repeatability of and correlations between some of
the behaviours measured, changes in the social environment had the greater effect. We discuss the
ecological and evolutionary implications of our findings and how they relate to our current under-
standing of social and nonsocial environmental effects on behaviour.
©2018 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
The effect of environmental factors on animal phenotypes is
well established. In particular, the behaviour of an animal can be
profoundly influenced by its social environment. In honeybees, Apis
mellifera, for example, brood pheromone has been found to affect
the age at which workers start foraging (Le Conte, Mohammedi, &
Robinson, 2001) and in bank voles, Myodes glareolus, male expen-
diture on the ejaculate can be affected solely by the presence of
rival male pheromones in the environment (delBarco-Trillo &
Ferkin, 2004). Similarly, we have known for some time that
different levels of predation risk affect both nonbehavioural (Creel,
Christianson, Liley, &Winnie, 2007; Hawlena &Schmitz, 2010) and
behavioural traits (Briffa, Rundle, &Fryer, 2008; Lima &Dill, 1990;
Werner, Gilliam, Hall, &Mittelbach, 1983) that serve in predator
avoidance in both vertebrate and invertebrate taxa.
Investigating between-individual (animal personality) and
within-individual behavioural variation (phenotypic plasticity)
together (Dingemanse, Kazem, R
eale, &Wright, 2010) has attracted
increased interest in recent years. Different explanations for such
behavioural variation have been proposed. Besides adaptations to
endogenous attributes such as cognitive ability (Sih &Del Giudice,
2012) and metabolism (Wolf &McNamara, 2012), between-
individual behavioural variation may be shaped by exogenous
factors such as predation threats (e.g. Bell &Sih, 2007; Sih, Kats, &
Maurer, 2003) or social environments (Montiglio, Ferrari, &R
eale,
2013; Wolf &McNamara, 2013). When individuals interact with
conspecifics in a way that influences their own behaviour (inter-
acting phenotypes; Moore, Brodie III, &Wolf, 2009), indirect ge-
netic effects (IGEs) are predicted to arise, where the genes of
interacting individuals affect the expression of traits in one another
(Moore et al., 2009; Wolf, Brodie, Cheverud, Moore, &Wade, 1998).
A diverse range of selective pressures can therefore result from
social interactions which might prove to be especially important
drivers of behavioural variation (Bailey, Marie-Orleach, &Moore,
2017). Similarly, we can expect behavioural plasticity in the light of
*Correspondence: F. S. Rudin, Centre for Evolutionary Biology (M092), The
University of Western Australia, Crawley, Perth, WA 6009, Australia.
E-mail address: fabian.rudin@research.uwa.edu.au (F. S. Rudin).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
https://doi.org/10.1016/j.anbehav.2018.02.013
0003-3472/©2018 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Animal Behaviour 138 (2018) 109e121
predation risks. However, plasticity in response to such risks may
be more costly (and therefore lower) because incorrect decisions
often lead to death (which is not the case for plasticity in response
to social cues). Thus, behaviour may be optimized to maximize
survival in different environments. In a recent review, Bailey et al.
(2017) suggested that behaviour is particularly prone to variation
in the social environment. Such variation should therefore have a
greater impact on behavioural plasticity than other aspects of the
environment such as the presence of predators. Some recent
theoretical papers suggest that there are coevolutionary processes
that lead to the existence of both socially responsive and consistent
individuals as a result of negative frequency dependence
(Johnstone &Manica, 2011; McNamara, Stephens, Dall, &Houston,
2009; Wolf, Van Doorn, &Weissing, 2011). Although predators and
prey coevolve, the presence of different predators and a diversity of
prey may dilute these effects in comparison to social interactions
within species. Thus, we may expect social interactions to have
more pronounced effects on behavioural plasticity. Much remains
to be learned about the ways in which environmental cues shape
between- and within-individual behavioural variation. Here we
investigated the effects of different environmental cues (social
versus nonsocial) on behavioural plasticity within the same
experimental framework.
We used Australian field crickets, Teleogryllus oceanicus, to test
the hypothesis that the environment, and changes therein, can
affect behavioural expression (phenotypic plasticity), the repeat-
ability of behaviours (a phenomenon often referred to as 'animal
personality': Bell, Hankison, &Laskowski, 2009; Gosling, 2001) and
correlations between multiple behavioural traits ('behavioural
syndromes': Bell, 2007; Sih, Bell, Johnson, &Ziemba, 2004). We
manipulated two aspects of the environment, the social environ-
ment via acoustic cues from conspecifics, and the physical envi-
ronment via mechanical disturbance, and examined the effects of
the presence and absence of these cues on male behaviour. In
crickets, acoustic sexual signals have been found to affect aggres-
sion, dominance, female mate choice and alternative mating tactics
(Bailey &Zuk, 2008; Bailey, Gray, &Zuk, 2010; DiRienzo, Pruitt, &
Hedrick, 2012). Males from various cricket species have been
found to be attracted to conspecific song, forming clusters in which
individuals remain relatively stationary while broadcasting acous-
tic sexual signals (Campbell &Shipp, 1979; Simmons, 1988;
Tinghitella, Wang, &Zuk, 2009). Based on these findings, we pre-
dicted that males would be more likely to engage in searching
behaviour (emerge quickly from a shelter and be more explorative
and active in search of conspecifics) in the absence than the pres-
ence of conspecific calls. Different levels of predation or parasitism
risk can affect how cautiously individuals behave (Hedrick &Kortet,
2006; Lewkiewicz &Zuk, 2004). Therefore, we might expect the
presence of a physical disturbance to render crickets less active and
less bold than crickets that are not exposed to disturbances. How-
ever, we expected the effect of disturbances in the environment to
be small compared to changes in the social environment owing to
the special role that the social environment is imputed to have in
the evolution of animal behaviour (Bailey et al., 2017).
In a previous study (Rudin, Tomkins, &Simmons, 2017), we
found that changes in dominance status eroded the repeatability of
some behaviours, but that boldness (latency to emerge from a
shelter) remained relatively stable. Additionally, changes in social
status had a disruptive effect on the correlation between boldness
and activity, but not on the correlation between boldness and
exploration. Because of the links between social status and acoustic
sexual signals (e.g. Brown, Smith, Moskalik, &Gabriel, 2006;
Callander, Kahn, Hunt, Backwell, &Jennionsa, 2013; Simmons,
1986), we predicted that changes in such signals will similarly
affect the repeatability of and correlations between behaviours.
Previous studies have investigated environmental effects on the
repeatability and expression of behaviours by exposing animals to
different environments, measuring them repeatedly in the same
environment. There is a distinct lack of studies that have
investigated the effects of relatively short-term environmental
changes on the repeatability and expression of and correlations
between behaviours. Our experimental design allowed us to
investigate the effects of such changes. Additionally, comparing
individuals that experienced a switch in environments to those
that did not allowed us to infer the presence or absence of
between-individual variation in behavioural plasticity, or
individual-by-environment interactions (IxEs; Alonzo, 2015;
Dingemanse &Wolf, 2013; Mathot, Wright, Kempenaers, &
Dingemanse, 2012; Stamps, 2016). Although researchers have
recently begun to focus on IxEs, much remains to be learned
about them, especially in the light of changes in the social
environment (Bailey et al., 2017).
METHODS
Study Population
The animals used in this experiment came from a large outbred
laboratory stock population (>1000 individuals) which is restocked
annually with freshly collected individuals from Carnarvon
(Western Australia). Animals were reared with ad libitum access to
food and water and held at 26
C on a 12:12 h light:dark cycle. At
the final larval instar, males (N¼208) were taken from the stock
population and housed in individual clear plastic containers
(7 7 cm and 5 cm high). Individuals were checked daily and
placed into experimental treatments the day following their final
moult to adulthood.
Experiment 1: Sociosexual Environment
In our first experiment, crickets were exposed to the presence
and absence of acoustic sexual signals from conspecific males.
Four groups of 26 crickets each (total N¼104) were assigned to
four separate environmental chambers, two silent and two
acoustic. We clipped the tegmen of all crickets to ensure they
could not produce song. Within the acoustic chambers, 5 min re-
cordings of about 30 sexually mature males housed with an equal
number of females were played back continuously. These re-
cordings included a mixture of calling, courtship and aggressive
song. The playback devices were MP3 players (iPod nano 7th Gen
and iPod classic 6th Gen) and speakers (Logitech Z200 Multimedia
Speakers and Philips Speaker Dock SBD8000/79). The light:dark
cycles of all chambers were set to 12:12 h light:dark and all were
held at 26
C. After 1 week of exposure to either the silent or
acoustic environments, behavioural trials were conducted as
described below. After behavioural trials, half of the crickets were
returned to the same treatment they had been exposed to previ-
ously, while the other half switched treatment, either from the
silent to the acoustic or from the acoustic to the silent treatment.
Crickets were again exposed to these treatments for a week after
which behavioural trials were repeated. This resulted in four
groups of individuals at the end of the two trials: AA (acoustic
environment for first week, acoustic environment second week),
AS (acoustic environment for first week, silent environment sec-
ond week), SS (silent environment for first week, silent environ-
ment second week) and SA (silent environment for first week,
acoustic environment second week) (Fig. 1). Each of these groups
consisted of 26 individuals.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121110
Experiment 2: Physical Disturbance
In our second experiment, a different set of crickets (again
silenced by clipping of the tegmen) was exposed to the presence or
absence of disturbance stimuli. We modified a commercial ceiling
extraction fan by removing the fan blades and cable-tying a weight
(large steel nut) eccentrically to the spindle. When the fan was
turned on, the eccentric weight caused the fan to shake. The
modified fan was then cable-tied to the top of a transparent plastic
box (41 30 cm and 25 cm high), into which the crickets were later
placed in their individual containers. Thus, the plastic box was
shaken when the fan was turned on. The fan was attached to an
HPM Slim Digital Timer (D817SLIM) which was set to turn on for
3 min, five times a day, at random intervals (5e535 min) with the
random interval repeating every other day. Intervals were ran-
domized to reduce potential habituation to the cue. For the un-
disturbed environment, crickets, in their individual boxes, were
placed into an identical large transparent plastic box that was not
fitted with a shaker. Again each of these disturbed/undisturbed
treatments were replicated twice across two environmental
chambers held on a 12:12 h light:dark cycle and at 26
C. Of the 104
crickets used in this experiment, half were initially exposed to the
disturbance treatment for a week (N¼52, 26 for each replicate
chamber) while the other half were left undisturbed (N¼52, 26 in
each replicate chamber). Behavioural trials were then conducted
(see below). After behavioural trials, half of the crickets were
returned to the same treatment that they had been exposed to
previously, while the other half switched treatments (Fig. 1).
Crickets were again exposed to these treatments for a week after
which behavioural trials were repeated. This yielded four groups of
individuals: DD (disturbed environment for first week, disturbed
environment second week), DU (disturbed environment for first
week, undisturbed environment second week), UU (undisturbed
environment for first week, undisturbed environment second
week) and UD (undisturbed environment for first week, disturbed
environment second week).
Behavioural Trials
All behavioural trials were conducted in an environmental
chamber at 26
C during the dark phase. To ensure that the crickets
were not disturbed by the observer, trials were conducted under
dim red illumination. The experimental set-up (Fig. 2) consisted of
a plastic trough (31 cm deep, 38 52 cm at top, 32 46 cm at base)
into which a shelter cut from PVC pipe (height: 8.5 cm; diameter:
Environmental
chamber
Environmental
chamber
Environmental
chamber
Environmental
chamber
Behavioural
trials
Behavioural
trials
Week 1 Week 2
104 crickets
Treatment
N = 52
N = 52
Control stimuli
N = 26
N = 26
N = 26
N = 26
Figure 1. Diagram of experimental design for both experiments (N¼104 for each experiment). ‘Treatment’was either exposure to male acoustic sexual signals or physical
disturbance (speaker symbols are used to illustrate the acoustic manipulation used in experiment 1). ‘Control stimuli’was the absence of acoustic signals or physical disturbances.
After 1 week of being exposed to either the treatment or the control environment, crickets underwent the behavioural trials. After the trials, half of the crickets spent another week
in the same environment they were exposed to during week 1 and the other half were switched to the environment they had not previously experienced. After week 2 each cricket
underwent another behavioural trial.
Middle
Far
17 cm
46 cm
Movable door
Near
8 cm
Shelter
17 cm
32 cm
Figure 2. Diagram of arena in which behavioural trials took place.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121 111
8 cm) was placed. The shelter had an opening that was fitted with a
movable door which could be opened from outside the trough by
pulling a piece of string attached to the door. Fine sand was used to
cover the base of the trough (ca. 2 cm deep).
Motion analysis software (EthoVision v8.5, Noldus, Wageningen,
The Netherlands) was used to track the movement of individuals
within the arena, resulting in objective, quantifiable measurements
of their behaviour. Crickets were filmed with a video camera
(Panasonic WV-CL930) installed 80 cm above the base of the arena.
Within the software, we defined different areas: the area closest to
the shelter (17 cm radius around the corner of the arena closest to
the shelter) was defined as ‘near’. The area furthest from the shelter
(17 cm radius around the corner opposite the shelter) was defined
as ‘far’. The area between ‘near’and ‘far’was defined as ‘middle’.
The specific behaviours quantified through EthoVision were as
follows: the latency to emerge from the shelter (once the whole
body of the cricket was outside the shelter); the total distance
moved within the arena; the average velocity within the arena; the
latency to ‘far’; the time spent moving (any movement >0.3 cm/s);
the total time spent within the arena as well as within the ‘near’,
‘middle’and ‘far’zones.
After 1 week of exposure to the environmental treatments, each
cricket was placed inside the shelter with the door closed, and the
shelter was tapped with a plastic rod for approximately 10 s to
disturb the cricket. Thirty seconds later the door was carefully
opened. This was defined as the starting point of the trial. Each
cricket was given 10 min to emerge from the shelter. Those crickets
that failed to emerge from the shelter within 10 min resulted in the
termination of the trial for those individuals. Once a cricket
emerged, its movements were tracked by EthoVision for 10 min.
We did not include the time individuals spent inside the shelter
after exiting the shelter for the first time in the analysis because it is
the inverse of the total time spent in the arena.
Within 1.5 h of completing their first behavioural trial, crickets
were transferred to their subsequent environments (either
reversed or kept the same; Fig. 1) where they remained for a further
week. The second behavioural trail was then conducted as
described above. After the second behavioural trial crickets were
frozen, and later their size and weight measured.
Size and Weight Measurements
We measured pronotum width using Mitutoyo CD-6 ASX calli-
pers (precision: 0.01 mm) and weight using a KERN PLS 510-3 scale
(precision: 0.001 g). There was a high degree of correlation be-
tween the two measures across all crickets used for both experi-
ments (r
S
¼0.71, N¼208, P<0.001). Therefore, pronotum width
was used as an estimator of individual size for all statistical analyses
because it is a fixed measure of size that does not vary with recent
food or water intake. Mean ±SE pronotum width was
6.36 ±0.02 mm and mean ±SE weight was 0.556 g ±0.006 g.
Statistical Analyses
We grouped behavioural measures into different categories as
follows: the latency to emerge from the shelter following a startle
cue was defined as ‘boldness’(sensu Carter, Feeney, Marshall,
Cowlishaw, &Heinsohn, 2013). For all other behavioural mea-
sures we performed a single principal components analysis (PCA)
for each experiment on the correlation matrix of data pooled from
trials conducted after the first and second week. The axes of vari-
ation were Varimax rotated and scores on the axes with eigen-
values >1 were extracted. For both experiments, the number of
components extracted was 2 (Table 1). The number of components
to be extracted was confirmed by running a parallel analysis using
the fa.parallel function (psych package in R). In both cases, average
velocity, the total time spent in the arena, time spent ‘middle’, time
spent ‘far’and time spent moving loaded most strongly onto the
first principal component (PC1). The latency to ‘far’, distance moved
and time spent ‘near’loaded most strongly onto the second prin-
cipal component (PC2). Comparing the two sets of components
using the factor.congruence function (psych package in R) resulted
in the following coefficients: 0.852 for PC1 Acoustic ePC1 Distur-
bance; -0.854 for PC2 Acoustic ePC2 Disturbance. These co-
efficients suggest relatively low marginal similarity between both
sets of components (Lorenzo-Seva &ten Berge, 2006). Component
scores extraction was regression based. Henceforth, PC1 scores are
used as a measure of ‘activity’, while PC2 scores are used as a
measure of ‘exploration’. Since a high PC2 score equates to more
time ‘near’and a lower proportion of time spent moving (Table 1),
the score indicates how ‘unexplorative’a cricket is. We therefore
reverse signed PC2 to ease interpretation.
Both experiments had a sample size of 104 crickets and in-
dividuals underwent two trials each. This resulted in 208 trials for
each experiment. In the sociosexual experiment, crickets did not
emerge from their shelter in 18 of 208 trials (8.7%), 10 of them after
week 1 and eight after week 2. In the disturbance treatment,
crickets did not emerge in 16 of 208 trials (7.7%), eight after week 1
and eight after week 2. The crickets that did not emerge from their
shelter were assigned the maximum time allowed for emergence,
i.e. 600 s. If an individual failed to express a behaviour within ‘ac-
tivity’or ‘exploration’(which was the case if they failed to emerge
from the shelter or if, for example, they never moved to the ‘far’
area of the arena), they were considered missing values for those
two behavioural axes. This occurred in 53 of 208 (25.5%) trials (34
after week 1 and 19 after week 2) in the sociosexual experiment
and in 31 of 208 (14.9%) trials (15 after week 1 and 16 after week 2)
in the disturbance experiment. In both experiments, sample sizes
were therefore lower for activity and exploration (sociosexual: 155
overall, 70 after week 1 and 85 after week 2; disturbance: 177
overall, 89 after week 1 and 88 after week 2) than for boldness
(both treatments: 208, 104 after week 1 and 104 after week 2). For
some analyses we used differentials, which were calculated by
subtracting the week 2 score from the week 1 score (latency to
emerge for boldness and the principal component scores for ac-
tivity and exploration). In these cases, sample sizes were further
reduced for activity and exploration. This is because differentials
could not be calculated if a cricket did not emerge and express all
behaviours within ‘activity’and ‘exploration’at both week 1 and
week 2. Consequently, the sample sizes for changes in activity and
Table 1
Loadings of the original behavioural variables on the principal axes of variation (PC1
and PC2) for the sociosexual and the disturbance environments
Sociosexual
environment
Disturbance
environment
PC1 PC2 PC1 PC2
Eigenvalue 3.12 2.22 3.12 1.84
Variance explained (%) 38.98 27.68 39.02 23.04
Distance moved 0.271 0.756 0.439 0.700
Latency to ‘far’0.120 0.829 0.331 0.657
Total time in arena 0.860 0.411 0.939 0.077
Time ‘near’0.033 0.863 0.478 0.750
Time ‘middle’0.829 0.030 0.571 0.474
Time ‘far’0.861 0.143 0.642 0.453
Velocity 0.721 0.084 0.581 0.263
Time spent moving 0.584 0.117 0.620 0.254
All behaviours were measured twice for each individual (after week 1 and after
week 2 of the experiment) and the principal components analysis was run on both
measures for each individual.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121112
exploration between week 1 and week 2 were 62 for the socio-
sexual experiment and 76 for the disturbance experiment.
To check whether there was a difference between individuals
that spent week 1 exposed to the treatment stimulus versus those
that were not, we ran a series of univariate general linear models.
Within each, the fixed factor was the environment: Acoustic (A)
and Silent (S) for experiment 1; Disturbed (D) and Undisturbed (U)
for experiment 2. The dependent variables were the behavioural
measures (latency to emerge for boldness, principal component
scores for activity and exploration) and the covariate was pronotum
width. To look for changes in behavioural expression in week 2,
univariate general linear models were run foreach behavioural trait
(boldness and the principal component scores for activity and
exploration) using the treatment profiles (AA, AS, SS, SA in exper-
iment 1, or DD, DU, UU, UD in experiment 2) as the fixed factor. In
each case, the dependent variables were the differentials (calcu-
lated as described above). Pronotum width was again used as the
covariate. To determine whether individuals that switched envi-
ronments after week 1 (AS and SA or DU and UD) differed signifi-
cantly from those that experienced the same environment for 2
consecutive weeks (AA and SS or DD and UU) we used Fisher's least
significant difference post hoc comparisons. Equality of variance
assumptions were met for all traits.
To test the repeatability of behaviours across the two behav-
ioural trials (week 1 and week 2) we used the R (version 3.4.2)
package rptR (Nakagawa &Schielzeth, 2010; Stoffel, Nakagawa, &
Schielzeth, 2017) to calculate repeatabilities (or intraclass correla-
tion coefficients) from variance components obtained from linear
mixed models. By using individual identities as random effects,
repeatabilities explain the proportion of total variance accounted
for by differences between individuals. In other words, repeat-
ability represents both the variability of behavioural traits across
individuals and their relative consistency within individuals.
Looking at the individual variance components from which re-
peatabilities were calculated gave us an indication as to whether
changes in within- or between-individual variation drove changes
in repeatability (although these differences could not be formally
tested). Response variables were analysed using REML estimation
and Gaussian fit because they were all approximately normally
distributed. Repeatability estimates were calculated with 1000
bootstrappings and 1000 permutations. Besides calculating the
repeatabilities of the response variable (boldness, activity and
exploration) pooled across all treatment profiles, repeatabilities of
the response variables were also calculated for each treatment
profile (AA, AS, SA, SS and DD, DU, UD, UU) separately. To test
whether repeatabilities differed significantly from each other we
calculated 84% confidence intervals around the repeatability esti-
mates. Repeatabilities were considered significantly different from
each other when these confidence intervals did not overlap. Two
95% confidence intervals that do not overlap may not differ
significantly at the 0.05 level and the 0.05 significance level can be
visualized by the nonoverlap criterion when adjusting the confi-
dence levels to 1.39 times the standard error (or 84%), which is why
we used 84% (and not 95%) confidence intervals (Goldstein &Healy,
1995).
Spearman rank correlations (r
S
) were used to check for the ex-
istence of between-individual behavioural correlations (following
examples and suggestions in Bell, 2007; Huntingford, 1976).
Because activity and exploration were represented by principal
components scores (which erased any possible correlations), cor-
relations between these two traits could not be assessed. Correla-
tions were assessed separately for week 1 and week 2 measures.
Week 1 measures were split into individuals that were exposed to
the cue (male acoustic signals or disturbance) and those that were
not. Week 2 measures were split into four groups for each
experiment: those that were exposed to the cue for 2 weeks, those
that were exposed to the cue in week 1 but not in week 2, those that
were never exposed to the cue and those that were not exposed to
the cue in week 1 but exposed to it in week 2. To determine
whether differences between correlation coefficients were signifi-
cant we used Fisher's r-to-z transformation (Myers &Sirois, 2006).
RESULTS
Experiment 1: Sociosexual Environment
The effect of environment on behaviour
There were significant boldness, activity and exploration dif-
ferences between individuals exposed to the male acoustic signals
and individuals held in the silent environment during the first
week. Crickets that were not exposed to male calls were signifi-
cantly bolder (decreased latency to emerge; Fig. 3a; F
2, 104
¼4.104,
P¼0.019), more active (Fig. 3b; F
2, 77
¼7.171, P¼0.001) and more
explorative (Fig. 3c; F
2, 77
¼7.356, P¼0.001) than those that were
exposed to calls. Size did not have a significant effect on any of the
behaviours (boldness: F
1, 104
¼1.311, P¼0.255; activity: F
1,
77
¼3.361, P¼0.071; exploration: F
1, 77
¼0.413, P¼0.522).
The effects of changing the environment
At week 2 there were significant changes in boldness, activity
and exploration depending on whether crickets switched envi-
ronments (Fig. 4). General linear models were significant for all
three traits (boldness: F
3, 103
¼3.687, P¼0.015; activity: F
3,
63
¼5.685, P¼0.002; exploration: F
3, 63
¼6.545, P¼0.001). For
boldness and activity, these changes were driven by significant
differences between individuals that did not change environments
(AA and SS, respectively) and individuals that did (AS and SA,
respectively; Fig. 4), as revealed by post hoc analyses. Individuals
that switched from acoustic to silent (AS) were bolder and more
active than those that stayed in the acoustic environment (AA)
while individuals that switched from silent to acoustic (SA) were
less bold and active than individuals that stayed in the silent
environment (SS). For exploration, only individuals that spent both
weeks in the silent environment (SS) differed from individuals that
switched from silent to being exposed to male acoustic signals (SA).
Individuals that were switched from the silent to the acoustic
environment were less explorative than those that remained in the
silent environment. No difference was found between individuals
that spent both weeks in the acoustic environment (AA) and those
that switched from being exposed to male acoustic signals to the
silent environment (AS; Fig. 4). Size had no significant effect on any
of the behaviours (boldness: F
1, 103
¼0.433, P¼0.512; activity: F
1,
63
¼2.196, P¼0.144; exploration: F
1, 63
¼0.030, P¼0.863).
Repeatabilities
Fig. 5 shows the repeatabilities of all personality traits for the
different treatment profiles (AA, AS, SS and SA) and the overall
repeatabilities for each trait (pooled across all treatment profiles).
Within boldness and activity, the 84% confidence intervals around
the repeatabilities of the individuals that changed from the silent to
the acoustic environment (SA) did not overlap with those of the
other treatment profiles (AA, AS and SS) and the repeatabilities
were therefore significantly different (the repeatability of SA being
lower than AA, AS and SS). For exploration, the 84% confidence
intervals around the repeatabilities of both treatment profiles in
which crickets spent the first week in the silent treatment (SS and
SA) did not overlap with and was lower than the other two profiles
(AA and AS). The within- and between-individual variance com-
ponents from which repeatabilities were calculated are shown in
Appendix Fig. A1.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121 113
Correlations between behaviours
Correlations between boldness and activity and boldness and
exploration for week 1 are compared in Table 2, and for week 2 in
Table 3. After week 1, correlations between all behavioural traits
(boldness eactivity and boldness eexploration) were significant.
After week 2, correlations between all traits were significant for
both cases in which individuals did not experience a change in
environments (AA and SS) but not for cases in which a change
occurred (AS and SA). The one exception was the correlation be-
tween boldness and activity, which was still significant for in-
dividuals that changed from the acoustic to the silent environment
(AS). Only the correlations between boldness and exploration
differed significantly between AA and AS individuals.
Experiment 2: Disturbance
The effect of environment on behaviour
Significant differences in behaviour between individuals
exposed to random bouts of disturbance and individuals that were
undisturbed were found only for exploration, with individuals that
were exposed to disturbance being less explorative than those that
were exposed (Fig. 6c; F
2, 89
¼5.073, P¼0.008). No significant ef-
fects were found for boldness (latency to emerge; Fig. 6a; F
2,
104
¼1.835, P¼0.165) and activity (Fig. 6b; F
2, 89
¼0.830,
P¼0.439). Pronotum width had no significant effect on any of the
behaviours (boldness: F
1, 104
¼0.397, P¼0.530; activity: F
1,
89
¼0.344, P¼0.559; exploration: F
1, 77
¼0.669, P¼0.416).
The effects of changing the environment
After week 2 there were significant changes in exploration
depending on whether a switch in the environment occurred (F
3,
75
¼3.383, P¼0.014; Fig. 7). The other general linear models were
nonsignificant (boldness: F
3, 103
¼2.201, P¼0.074; activity: F
3,
75
¼1.767, P¼0.145). Individuals that did not experience a change
midway through the experiment (DD and UU) did not alter their
behaviour in week 2. There was a significant difference between
individuals that spent both weeks in the undisturbed environment
(UU) and those that switched from the undisturbed to the
disturbed environment (UD; Fig. 7). Here, individuals became less
explorative (the differential became negative) when switching
from the undisturbed to the disturbed environment, whereas no
change was observed for individuals that spent both weeks in the
undisturbed environment. No such difference was observed be-
tween individuals that spent both weeks in the disturbed envi-
ronment (DD) and those that switched from the disturbed to the
undisturbed (DU). No such effects were observed for boldness and
activity. Size had no significant effect on any of the behaviours
(boldness: F
1, 103
¼1.663, P¼0.200; activity: F
1, 75
¼1.376 ,
P¼0.245; exploration: F
1, 63
¼1.548, P¼0.218).
Repeatabilities
Fig. 8 shows the repeatabilities for the three behavioural traits,
both pooled across all treatment profiles (DD, DU, UU and UD) and
within each profile separately. All the 84% confidence intervals
around the repeatabilities within each personality trait overlapped
and could therefore not be considered significantly different from
each other.
Correlations between behaviours
Tables 2 and 3 show the correlations between the three
behavioural traits after week 1 and 2, respectively. Correlations
between all behavioural traits were significant after week 1. All but
one of the correlations (boldness eexploration for DU individuals)
were also significant after week 2. Comparing DD to DU individuals
and UU to UD individuals, none of the correlations were signifi-
cantly different from each other.
DISCUSSION
The social environment is expected to have a particularly strong
effect on animal behaviour (Bailey et al., 2017). We found that when
crickets were switched from a silent environment to an environ-
ment in which male acoustic sexual signals were present, they
became less bold, active and explorative. When switched from an
acoustic to a silent environment, individuals became bolder and
more active. Changes in the social environment also affected the
repeatability of and correlations between some behaviours. In
contrast, changing between an undisturbed and disturbed envi-
ronment had either no or considerably weaker effects on behav-
ioural plasticity, repeatability and correlations. These findings
support the hypothesis that the social environment affects behav-
iour more markedly than the nonsocial environment and highlight
the importance of interacting phenotypes in understanding
behavioural consistency and plasticity.
600
500
400
300
200
100
0
3
2
1
0
–1
–2
–3
3
2
1
0
–1
–2
–3
AS AS AS
Latency to emerge (s)
Activity (PC1)
Exploration (PC2)
*** **
(a) (b) (c)
Figure 3. Differences between individuals that were exposed to male acoustic signals for 1 week (A) and those that were not exposed (S) using univariate general linear models. (a)
Boldness (latency to emerge), (b) activity and (c) exploration. The bottom and top of the box represent the first (Q1) and third quantiles (Q3) of the data, respectively (interquartile
region, IQR). The horizontal line within the box represents the median. The whiskers end at the largest and smallest nonoutliers. Outliers (1.5 IQR above Q1 and below Q3) are
represented by dots. *P<0.05; **P<0.01.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121114
Variation in the Social Environment
Balenger and Zuk (2015) showed that T. oceanicus from a pop-
ulation in which a genetic mutation rendered approximately 90% of
the males incapable of calling, exhibited increased locomotor
behaviour when reared in silence. They argued that, in the absence
of calling rivals, it would be beneficial for individuals to increase
overall movement if such behavioural alterations increase mating
opportunities. Furthermore, crickets have been found to cluster in
areas in which other crickets are calling; searching behaviour is
increased in the absence and decreased in the presence of other
calling males (e.g. Cade, 1981b; Campbell &Shipp, 1979). When
reared in the presence of male calling song, male T. oceanicus were
less likely to employ satellite behaviour (settling near calling males
to parasitize their song to gain access to females) as adults
compared to crickets reared in silence (Bailey et al., 2010). While
the population used for our experiment does not exhibit mutations
rendering them silent and our crickets were not reared as juveniles
in the absence of conspecific song, we can nevertheless draw par-
allels between these previous studies and our findings. Here,
crickets held in a silent environment for the first week of their adult
lives emerged more quickly from a shelter and displayed more
active and explorative behaviour than those that spent the week in
the presence of male acoustic sexual signals. Thus, the presence of
male song may be an indicator for an individual that it is within a
chorus, reducing its need to search for an aggregation of singing
conspecifics. Conversely, males that do not experience song during
the first week of their adult lives may increase their explorative
behaviour and activity and more readily leave a shelter and search
for a chorus. Females are attracted by male acoustic sexual signals
(Ulagaraj &Walker, 1973) and the choruses males form have been
compared to ‘lek’type mating systems (Alexander, 1975). In such
systems, females compare males and choose mates (e.g. Beehler &
Foster, 1988). Female T. oceanicus have been found to become more
choosy when reared in an environment that mimicked an aggre-
gation of calling males compared to females reared in a silent
environment (Bailey &Zuk, 2008). Thus, male behaviour may be
explained as a response to female behaviour; staying put in a
chorus (as simulated here through playback of male song) is ex-
pected to increase a male's mating success. Within calling aggre-
gations, males of multiple field cricket species space out relatively
evenly. This is thought to be a result of maintaining an exclusive
female attraction zone with calls acting as aggressive signals (Cade,
1981b; Campbell &Shipp, 1979; Simmons, 1988). Thus, acoustic
spacing within aggregations would further decrease exploration,
activity and boldness in crickets exposed to the acoustic sexual
signals of rivals.
We found considerable behavioural plasticity in response to
changes in the acoustic environment. Interestingly, both scenarios
(changing from the acoustic to the silent or silent to acoustic en-
vironments) elicited statistically significant changes in the average
level of all behaviours, except for exploration in crickets that
experienced a switch from the acoustic to the silent environment.
The fact that crickets become less bold, explorative and active when
they go from a silent environment to being exposed to male calls
further supports the notion that individuals are more likely to stay
put, rather than engage in searching behaviour, upon ‘joining’an
aggregation of singing males. This, again, is in line with studies that
found that males increased searching behaviour in the absence of
male acoustic sexual signals (e.g. Cade, 1981b; Campbell &Shipp,
1979).
We found significant repeatability in behaviours indicating
consistent between-individual variation in behaviour, or person-
ality. Repeatability can be high due to both between-individual
differences and within-individual consistency of behaviours
500
250
0
–250
–500
3
1.5
0
–1.5
3
AA AS SS SA
AA AS SS SA
AA AS SS SA
P = 0.045
P = 0.041
P = 0.028
P = 0.016
P = 0.001
P = 0.189
Latency to emerge differentials (s)Activity differentials (PC1)
Exploration differentials (PC2)
3
1.5
0
–1.5
3
(a)
(b)
(c)
Figure 4. Differences in behaviour at week 2 (behavioural score week 2 minus score at
week 1) for (a) boldness (latency to emerge), (b) activity and (c) exploration. Crickets
were either exposed to the same treatment for 2 weeks (acoustic: AA; silent: SS) or
experienced a change after the first week (AS or SA). The bottom and top of the box
represent the first (Q1) and third quantiles (Q3) of the data, respectively (interquartile
region, IQR). The horizontal line within the box represents the median. The whiskers
end at the largest and smallest nonoutliers. Outliers (1.5 IQR above Q1 and below
Q3) are represented by dots.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121 115
Pooled
AA
SS
AS
SA
Pooled
AA
SS
AS
SA
Pooled
AA
SS
AS
SA
BoldnessActivity
Exploration
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Re
p
eatabilit
y
Figure 5. Repeatabilities for boldness, activity and exploration pooled across all treatment profiles and for the different subsets (AA and SS i.e. no change in environment; AS and SA
i.e. change in environment). The 84% confidence intervals around repeatabilites are represented by the whiskers.
Table 2
Spearman correlations (r
S
±SE) between the three behavioural traits (boldness, exploration and activity) after week 1 [lower, upper 95% confidence intervals]
Treatment Boldness eactivity PBoldness eexploration P
Acoustic (A) 0.453 ±0.146
[0.127, 0.684]
0.006 0.429 ±0.157
[0.083, 0.707]
0.009
Silent (S) 0.502 ±0.117
[0.226, 0.691]
0.001 0.396 ±0.126
[0.142, 0.625]
0.009
z between A and S 0.28 0.78 (NS) 0.18 0.86 (NS)
Disturbed (D) 0.317 ±0.152
[0.002, 0.585]
0.041 0.359 ±0.155
[0.041, 0.631]
0.020
Undisturbed (U) 0.324 ±0.130
[0.226, 0.691]
0.026 0.341 ±0.131
[0.142, 0.625]
0.019
z between D and U 0.04 0.97 (NS) 0.1 0.92 (NS)
Correlations were compared statistically using a z test.
Table 3
Spearman correlations (r
S
±SE) between the three behavioural traits (boldness, exploration and activity) after week 2 [lower, upper 95% confidence intervals]
Treatment Boldness eactivity PBoldness eexploration P
AA 0.552 ±0.168
[0.127, 0.792]
0.009 0.555 ±0.164
[0.156, 0.778]
0.009
AS 0.561 ±0.176
[0.144, 0.848]
0.007 0.086 ±0.224
[0.407, 0.485]
0.456 (NS)
z between AA and AS 0.04 0.97 (NS) 1.97 0.048
SS 0.527 ±0.211
[0.073, 0.872]
0.010 0.428 ±0.214
[0.038, 0.767]
0.042
SA 0.066 ±0.234
[0.372, 0.521]
0.789 (NS) 0.201 ±0.255
[0.321, 0.670]
0.409 (NS)
z between SS and SA 1.64 0.10 (NS) 0.8 0.42 (NS)
DD 0.452 ±0.183
[0.016, 0.700]
0.049 0.502 ±0.167
[0.099, 0.762]
0.017
DU 0.375 ±0.195
[0.042, 0.699]
0.071 (NS) 0.410 ±0.221
[0.152, 0.674]
0.161 (NS)
z between DD and DU 0.29 0.77 (NS) 0.36 0.72 (NS)
UU 0.467 ±0.192
[0.050, 0.761]
0.025 0.505 ±0.166
[0.143, 0.769]
0.014
UD 0.518 ±0.166
[0.113, 0.764]
0.023 0.500 ±0.207
[0.008, 0.816]
0.029
z between SS and SA 0.21 0.83 (NS) 0.02 0.98 (NS)
The correlations after week 2 were split into subsets (AA, AS, SS, SA and DD, DU, UU, UD); correlations were compared statistically using a z test.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121116
(Nakagawa &Schielzeth, 2010). A meta-analysis found that, overall,
the repeatability of animal behaviours is around 0.37 (Bell et al.,
2009). We found overall repeatabilities of 0.44, 0.39 and 0.42 for
boldness, activity and exploration, respectively, close to or slightly
higher than the average behavioural repeatabilities reported for
insects (0.36; Bell et al., 2009). A recent study on the field cricket
Gryllus campestris found repeatabilities of 0.06, 0.21 and 0.12 for
boldness, activity and exploration, respectively, in the wild (Fisher,
James, Rodríguez-Mu~
noz, &Tregenza, 2015). Although there is a
trend for repeatabilites to be higher in the wild than in the labo-
ratory (Bell et al., 2009), the repeatabilities found here were
noticeably higher than those found by Fisher et al. (2015).
The repeatability of behaviours was relatively stable across most
treatment profiles. However, the repeatabilities of boldness and
activity decreased considerably for individuals that spent their first
week of adult life in the silent environment before experiencing a
switch to the acoustic environment. For exploration, repeatabilities
were low when individuals stayed in the silent environment and
when they were switched from silent to acoustic. This may suggest
that boldness and activity exhibit between-individual variation in
plasticity, or individual-by-environment interactions (IxE; Alonzo,
2015; Dingemanse &Wolf, 2013; Mathot et al., 2012; Stamps,
2016). Exploration, on the other hand, did not appear to exhibit
IxE. Repeatabilities can be reduced by IxEs when the rank order of
individuals changes in response to variation in the environment,
making individuals more unpredictable in their behaviour (i.e. an
increase in within-individual variation). If the reduction in re-
peatabilities is driven by decreases in between- rather than in-
creases in within-individual variation, the presence of an IxE
becomes less likely. We could not test for IxEs formally since more
observations per individual would be needed to run the appro-
priate mixed-effects models. However, our data suggest that vari-
ation in the social environment might affect between-individual
variation in plasticity (IxE) because reductions in repeatability were
driven by increases in within- and not decreases in between-
individual variances (Appendix Fig. A1). The fact that certain
changes in the social environment can lead to changes in behav-
ioural repeatability is similar to our previous finding that changes
in social dominance can erode the repeatability of some behaviours
(Rudin et al., 2017). Fighting success (Gryllus bimaculatus;Wedell &
Tregenza, 1999) and the tendency to adopt alternative mating
tactics (Gryllus integer;Cade, 1981a) have both been found to
exhibit significant additive genetic variance in crickets. Environ-
mentally dependent genetic variation (GxE) in either of these traits
in T. oceanicus could result in individual variation in plasticity in the
light of increased competition. Being switched from a silent to an
acoustic environment may indicate increased competition and thus
explain the disruption in the repeatability of behaviours we
observed. Interestingly, changing from the acoustic to the silent
environment did not result in the same disruption of repeatabilites,
suggesting that all individuals reacted in a similar way when
switched from a competitive to a noncompetitive environment.
Even though individuals adjusted their behaviour in response to
changes in their social environment, between-individual behav-
ioural differences persisted when crickets spent their first week of
adult life in the acoustic environment and were then switched to
the silent environment. Some studies have indicated that the
development of personalities may be delayed if sufficient infor-
mation on the state of the environment cannot be acquired
(Fawcett &Frankenhuis, 2015; Fischer, van Doorn, Dieckmann, &
Taborsky, 2014). Furthermore, the development of between-
individual behavioural differences in a population may be driven
by environmental challenges associated with the social environ-
ment to a considerable degree (e.g. Edenbrow &Croft, 2013;
Rittschof et al., 2014). Here, we found that the repeatability of all
behaviours was lower when crickets spent week 1 in the silent and
week 2 in the acoustic environment, and for exploration when both
weeks were spent in the silent environment. This may suggest that
being exposed to silence early during adulthood delays or even
inhibits the establishment of certain repeatable, predictable be-
haviours. This would be especially true if male acoustic sexual
signals are important for behavioural differentiation. Indeed, this is
supported by our finding that the behaviours of crickets that spent
their first week in the acoustic environment did not exhibit reduced
repeatability, regardless of whether they experienced a switch.
Furthermore, some recent studies found that the acoustic rearing
environment affects behaviour later in life in both male and female
crickets (Bailey &Zuk, 2008; Bailey et al., 2010). Little is known
about the effects of the social environment on behavioural
repeatability at different ages. Our study cannot address whether or
how long these effects would persist past the relatively short
timeframe of 2 weeks adopted in our experiment. Longer-term
studies would be needed to fully explore this question. However,
our results indicate that the presence or absence of sexual signals
600
500
400
300
200
100
0
3
2
1
0
–1
–2
–3
3
2
1
0
–1
–2
–3
DU D U D U
Latency to emerge (s)
Activity (PC1)
Exploration (PC2)
(a) (b) (c)
**
Figure 6. Behavioural differences between individuals that were exposed to randomly applied bouts of disturbance for 1 week and those that were not exposed using univariate
general linear models. (a) Boldness (latency to emerge), (b) activity and (c) exploration. The bottom and top of the box represent the first (Q1) and third quantiles (Q3) of the data,
respectively (interquartile region, IQR). The horizontal line within the box represents the median. The whiskers end at the largest and smallest nonoutliers. Outliers (1.5 IQR above
Q1 and below Q3) are represented by dots. **P<0.01.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121 117
early in adult life may affect the degree of behavioural repeatability
later on, at least for some behaviours.
Correlations between all behaviours were strong when in-
dividuals did not experience a switch in their environment; the
only correlation that was strong when there was a change in the
environment was that between boldness and activity for in-
dividuals experiencing a switch from acoustic to silent. IxEs are
thought to cause changes in correlations between traits (Brommer
&Class, 2017), so it is surprising that the only correlation that
changed significantly was one for which none of the traits exhibited
an IxE; the correlation between boldness and exploration was
higher for individuals that stayed in the acoustic environment than
those that were switched from acoustic to silent. Strong genetic
correlations have been thought to constrain the evolvability of
behaviours (Dochtermann &Dingemanse, 2013). Although this
study was conducted at the phenotypic level, it has been argued
that phenotypic correlations can serve as a proxy for genetic cor-
relations (Dochtermann, 2011). The fact that some aspects of the
social environment (e.g. dominance status, social environment) can
weaken otherwise strong behavioural correlations leads us to
caution against neglecting such social interactions when estimating
the evolvability of different behaviours.
Variation in Physical Disturbance
The ability to adjust behaviour in response to risks, such as
predation, in the environment should be beneficial from an adap-
tive perspective (Dall, Houston, &McNamara, 2004; Lima &Dill,
1990; Wolf &Weissing, 2012). In general, experiencing predator
cues can be expected to lead to more cautious behaviour (Smith &
Blumstein, 2008). Our expectation that the presence of a distur-
bance might affect cricket behaviour was only partly met. Here, we
found that crickets were less explorative after exposure to distur-
bances for a week, and this is in line with the expectation of
predator avoidance behaviour when faced with predator cues.
Similarly, crickets that were switched from being exposed to
disturbance to an undisturbed environment became less explor-
ative (but not vice versa). Although we lacked the sample size for
formal statistical analyses, the behavioural response (especially
boldness and activity) to the disturbance stimuli appeared to be
weak and less pronounced than the response to the presence of
male acoustic sexual signals. It is possible that the stimuli pre-
sented to the crickets in this study did not elicit strong responses
because they were not perceived as a predation threat, although
crickets were visibly startled when being disturbed by the shaker.
The levels of boldness and activity we observed may show low
plasticity (with exploration being somewhat more plastic) to
maximize survival within the population from which these crickets
were derived. Plasticity in response to social cues may be less costly
(wrong decision ¼reduced reproductive success) than plasticity in
response to potential threats (wrong decision ¼death). Alterna-
tively, or in addition, the relatively weak plasticity of exploration
and the absence of plasticity in boldness and activity may have
been due to habituation, resulting from 1 week of exposure to
shaking without negative consequences for the crickets. A study on
Christmas tree worms, Spirobranchus giganteus, found such an ef-
fect, with behavioural plasticity only being detected over short
timescales (1 day) but not long ones (4 days; Pezner, Lim, Kang,
Armenta, &Blumstein, 2017). Thus, the treatment period in our
experiment may have been too long to detect behavioural plasticity
in relation to disturbance.
We found no significant differences between the repeatabilities
of each behaviour (boldness, activity and exploration) within each
subset (DD, DU, UU and UD). Thus, disturbance stimuli had less
pronounced effects than acoustic sexual signals on repeatabilities.
500
250
0
–250
–500
3
1.5
0
–1.5
3
3
1.5
0
–1.5
3
DD DU UU UD
DD DU UU UD
DD DU UU UD
Latency to emerge differentials (s)Activity differentials (PC1)Exploration differentials (PC2)
P = 0.052
P = 0.136
P = 0.549
P = 0.505
P = 0.055
P = 0.027
(a)
(b)
(c)
Figure 7. Differences in behaviour at week 2 (behavioural score at week 2 minus score
at week 1) for (a) boldness (latency to emerge), (b) activity and (c) exploration. Crickets
were either exposed to the same treatment for 2 weeks (disturbance: DD; undis-
turbed: UU) or experienced a change after the first week (DU or UD). The bottom and
top of the box represent the first (Q1) and third quantiles (Q3) of the data, respectively
(interquartile region, IQR). The horizontal line within the box represents the median.
The whiskers end at the largest and smallest nonoutliers. Outliers (1.5 IQR above Q1
and below Q3) are represented by dots.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121118
In response to changes in the environmental stimuli presented
here, between-individual behavioural differences and behavioural
correlations appear to be relatively stable, regardless of the direc-
tion of such changes. This was to be expected since we found that
the effects of disturbance on the levels of behavioural expression
were negligible (no effects on boldness or activity and only a small
effect on exploration). Again, this may be attributable to habitua-
tion and/or point to the fact that all individuals react similarly to the
disturbance cue. Our findings suggest that there is no between-
individual variation in behavioural plasticity (IxE) in response to
changes in physical disturbance. Likewise, Mathot et al. (2011),
investigated escape flight duration in red knots, Calidris canutus
islandica,finding no between-individual variation in behavioural
plasticity within flocks. They attributed this finding to the fact that
escape flight durations exhibit positive frequency-dependent pay-
offs, which in turn constrain individual variation in plasticity.
The Social Environment and Behavioural Plasticity
We found considerable behavioural plasticity in response to the
social environment but not to a nonsocial environmental distur-
bance. These findings address gaps in our understandingof social and
nonsocial effects on behaviour since there appears to be a distinct
lack of studies comparing such effects within a single framework
(Bailey et al., 2017). The evidence found here and in our previous
study (Rudin et al., 2017) shows that social cues (dominance status,
acoustic sexual signals) can have appreciable effects on behaviour
and that changes in these social cues can affect the repeatability of
behaviours and between-individual behavioural correlations. The
fact that such changes can potentially disrupt personalities and cor-
relations between different behaviours is intriguing and has, in our
opinion, been somewhat neglected in the literature.
Behaviour is thought to be the result of an individual's intrinsic
properties and external abiotic and biotic effects. Traditionally,
genetic and environmental influences have been considered sepa-
rate. However, a distinctive property of social interactions is that
they are simultaneously environmental and genetic. Thus,
environmental effects can be heritable if the environment is social
and heritable differences contribute to these effects. Bailey and Zuk
(2012) found that female T. oceanicus from different populations
show different levels of choosiness depending on whether male
acoustic signals were present or not, indicating the presence of
indirect genetic effects. We note that we could not explicitly test for
indirect genetic effects in our study, since it was conducted at the
level of the phenotype, and we only had measurements from the
receivers of the social cue and not its broadcasters. Future studies
could build on ours by measuring behaviours in response to
different calling males rather than to recordings of song (e.g.
Santostefano, Wilson, Araya-Ajoy, &Dingemanse, 2016). We pre-
viously demonstrated that the outcome of agonistic social in-
teractions affected an individual's behavioural phenotype, namely
the expression of boldness, activity and exploration (Rudin et al.,
2017). Here, we have shown that the social environment has
strong effects on behavioural plasticity, but the nonsocial envi-
ronment does not. This provides further support for the notion that
responses to the social, rather than the nonsocial, environment are
likely to be important drivers of rapid coevolutionary dynamics
(Drown &Wade, 2014). This is also supported by our finding that
there is individual variation in plasticity in response to a change
from the silent to the acoustic environment. Since the behaviours
considered here are likely to affect access to food or mates, carry-
over fitness effects are a likely consequence for interacting in-
dividuals, resulting in differential selection. The nonsocial
environment (physical disturbance) did not appear to be as
important as the social environment in shaping behavioural plas-
ticity. Future work should use a quantitative-genetic framework to
determine the extent to which changes in the social environment
have the potential to affect evolutionary changes in behavioural
plasticity and personality.
Acknowledgments
We thank Jacob Berson, Maxine Lovegrove and Soon Hwee Ng
for their help with cricket husbandry and Carly Wilson and Carl
Pooled
DD
UU
DU
UD
Pooled
DD
UU
DU
UD
Pooled
DD
UU
DU
UD
BoldnessActivityExploration
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Re
p
eatabilit
y
Figure 8. Repeatabilities for boldness, activity and exploration pooled across all treatment profiles and for the different subsets (DD and UU i.e. no change in environment; DU and
UD i.e. change in environment). The 84% confidence intervals around repeatabilites are represented by the whiskers.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121 119
Schmidt for their help in constructing the experimental set-up. We
are grateful to Dr Alexander Weiss, Dr Alex Weir and two anony-
mous referees for their valuable comments on the manuscript.
Funding for this study was provided by the School of Biological
Sciences, The University of Western Australia. L.W.S.
(DP110104594) and J.L.T. (FT110100500) were supported by Dis-
covery Project Grants from the Australian Research Council. We
also thank the Australian Government for its support through an
Australian Government Research Training Program Scholarship.
References
Alexander, R. D. (1975). Natural selection and specialized chorusing behavior in
acoustical insects. In D. Pimental (Ed.), Insects, Science &Society (pp. 35e77).
New York, NY: Academic Press. Retrieved from: http://agris.fao.org/agris-
search/search.do?recordID¼US201303001227.
Alonzo, S. H. (2015). Integrating the how and why of within-individual and among-
individual variation and plasticity in behavior. Current Opinion in Behavioral
Sciences, 6,69e75. https://doi.org/10.1016/j.cobeha.2015.09.008.
Bailey, N. W., Gray, B., & Zuk, M. (2010). Acoustic experience shapes alternative
mating tactics and reproductive investment in male field crickets. Current
Biology, 20(9), 845e849. https://doi.org/10.1016/j.cub.2010.02.063.
Bailey, N. W., Marie-Orleach, L., & Moore, A. J. (2017). Indirect genetic effects in
behavioral ecology: Does behavior play a special role in evolution? Behavioral
Ecology, 29(1), 1e11. https://doi.org/10.1093/beheco/arx127.
Bailey, N. W., & Zuk, M. (2008). Acoustic experience shapes female mate choice in
field crickets. Proceedings of the Royal Society B: Biological Sciences, 275(1651),
2645e2650. https://doi.org/10.1098/rspb.2008.0859.
Bailey, N. W., & Zuk, M. (2012). Socially flexible female choice differs among pop-
ulations of the pacificfield cricket: Geographical variation in the interaction
coefficient psi (
J
). Proceedings of the Royal Society B: Biological Sciences,
279(1742), 3589e3596. https://doi.org/10.1098/rspb.2012.0631.
Balenger, S. L., & Zuk, M. (2015). Roaming romeos: Male crickets evolving in silence
show increased locomotor behaviours. Animal Behaviour, 101,213e219. https://
doi.org/10.1016/j.anbehav.2014.12.023.
Beehler, B. M., & Foster, M. S. (1988). Hotshots, hotspots, and female preference in
the organization of lek mating systems. The American Naturalist, 131(2),
203e219. https://doi.org/10.1086/284786.
Bell, A. M. (2007). Future directions in behavioural syndromes research. Proceedings
of the Royal Society B: Biological Sciences, 274(1611), 755e761. https://doi.org/
10.1098/rspb.2006.0199.
Bell, A. M., Hankison, S. J., & Laskowski, K. L. (2009). The repeatability of behaviour:
A meta-analysis. Animal Behaviour, 77(4), 771e783. Retrieved from: http://
www.ncbi.nlm.nih.gov/pmc/articles/PMC3972767/.
Bell, A. M., & Sih, A. (2007). Exposure to predation generates personality in
threespined sticklebacks (Gasterosteus aculeatus). Ecology Letters, 10(9),
828e834. https://doi.org/10.1111/j.1461-0248.2007.01081.x.
Briffa, M., Rundle, S. D., & Fryer, A. (2008). Comparing the strength of behavioural
plasticity and consistency across situations: Animal personalities in the hermit
crab Pagurus bernhardus.Proceedings of the Royal Society B: Biological Sciences,
275(1640), 1305e1311. https://doi.org/10.1098/rspb.2008.0025.
Brommer, J. E., & Class, B. (2017). Personality from the perspective of behavioral
ecology. In J. Vonk, A. Weiss, & S. A. Kuczaj (Eds.), Personality in Nonhuman
Animals (pp. 73e108). Berlin, Germany: Springer International Publishing.
https://doi.org/10.1007/978-3-319-59300-5.
Brown, W. D., Smith, A. T., Moskalik, B., & Gabriel, J. (2006). Aggressive contests in
house crickets: Size, motivation and the information content of aggressive
songs. Animal Behaviour, 72, 225e233. https://doi.org/10.1016/
j.anbehav.2006.01.012.
Cade, W. H. (1981a). Alternative male strategies: Genetic differences in crickets.
Science, 212(4494), 563e564. https://doi.org/10.1126/science.212.4494.563.
Cade, W. H. (1981b). Field cricket spacing, and the phonotaxis of crickets and
parasitoid flies to clumped and isolated cricket songs. Zeitschrift Für Tierp-
sychologie, 55(4), 365e375. https://doi.org/10.1111/j.1439-0310.1981.tb01278.x.
Callander, S., Kahn, A. T., Hunt, J., Backwell, P. R. Y., & Jennionsa, M. D. (2013). The
effect of competitors on calling effort and life span in male field crickets.
Behavioral Ecology, 24(5), 1251e1259. https://doi.org/10.1093/beheco/art059.
Campbell, D. J., & Shipp, E. (1979). Regulation of spatial pattern in populations of the
field cricket Teleogryllus commodus (Walker). Zeitschrift Für Tierpsychologie,
51(3), 260e268. https://doi.org/10.1111/j.1439-0310.1979.tb00688.x.
Carter, A. J., Feeney, W. E., Marshall, H. H., Cowlishaw, G., & Heinsohn, R. (2013).
Animal personality: What are behavioural ecologists measuring? Biological
Reviews of the Cambridge Philosophical Society, 88(2), 465e475. https://doi.org/
10.1111/brv.12007.
Creel, S., Christianson, D., Liley, S., & Winnie, J. A. (2007). Predation risk affects
reproductive physiology and demography of elk. Science, 315(5814). https://
doi.org/10.1126/science.1135918, 960e960.
Dall, S. R. X., Houston, A. I., & McNamara, J. M. (2004). The behavioural ecology of
personality: Consistent individual differences from an adaptive perspective.
Ecology Letters, 7(8), 734e739.https://doi.org/10.1111/j.1461-0248.2004.00618.x.
delBarco-Trillo, J., & Ferkin, M. H. (2004). Male mammals respond to a risk of sperm
competition conveyed by odours of conspecific males. Nature, 431(7007),
446e449. https://doi.org/10.1038/nature02845.
Dingemanse, N. J., Kazem, A. J. N., R
eale, D., & Wright, J. (2010). Behavioural reaction
norms: Animal personality meets individual plasticity. Trends in Ecology &
Evolution, 25(2), 81e89. https://doi.org/10.1016/j.tree.2009.07.013.
Dingemanse, N. J., & Wolf, M. (2013). Between-individual differences in behavioural
plasticity within populations: Causes and consequences. Animal Behaviour, 85(5),
1031 e1039. http://www.sciencedirect.com/science/article/pii/S0003347213000092.
DiRienzo, N., Pruitt, J. N., & Hedrick, A. V. (2012). Juvenile exposure to acoustic
sexual signals from conspecifics alters growth trajectory and an adult person-
ality trait. Animal Behaviour, 84(4), 861e868. https://doi.org/10.1016/
j.anbehav.2012.07.007.
Dochtermann, N. A. (2011). Testing Cheverud's conjecture for behavioral correla-
tions and behavioral syndromes. Evolution; International Journal of Organic
Evolution, 65(6), 1814e1820. https://doi.org/10.1111/j.1558-5646.2011.01264.x.
Dochtermann, N. A., & Dingemanse, N. J. (2013). Behavioral syndromes as evolu-
tionary constraints. Behavioral Ecology, 24(4), 806e811. https://doi.org/10.1093/
beheco/art002.
Drown, D. M., & Wade, M. J. (2014). Runaway coevolution: Adaptation to heritable
and nonheritable environments. Evolution, 68(10), 3039e3046. https://doi.org/
10.1111/evo.12470.
Edenbrow, M., & Croft, D. P. (2013). Environmental and genetic effects shape the devel-
opment of personality traits in the mangrove killifish Kryptolebias marmoratus.Oikos,
122(5), 667e681. https://doi.org/10.1111/j.1600-0706.2012.20556.x.
Fawcett, T. W., & Frankenhuis, W. E. (2015). Adaptive explanations for sensitive
windows in development. Frontiers in Zoology, 12(Suppl 1), S3. https://doi.org/
10.1186/1742-9994-12-S1-S3.
Fischer, B., van Doorn, G. S., Dieckmann, U., & Taborsky, B. (2014). The evolution of
age-dependent plasticity. The American Naturalist, 183(1), 108e125. https://
doi.org/10.1086/674008.
Fisher, D. N., James, A., Rodríguez-Mu~
noz, R., & Tregenza, T. (2015). Behaviour in
captivity predicts some aspects of natural behaviour, but not others, in a wild
cricket population. Proceedings of the Royal Society B: Biological Sciences,
282(1809), 20150708. https://doi.org/10.1098/rspb.2015.0708.
Goldstein, H., & Healy, M. J. R. (1995). The graphical presentation of a collection of
means. Journal of the Royal Statistical Society - Series a: Statistics in Society,
158(1), 175e177. https://doi.org/10.2307/2983411.
Gosling, S. D. (20 01). From mice to men: What can we learn about personality from
animal research? Psychological Bulletin, 127(1), 45e86. Retrieved from: http://
www.ncbi.nlm.nih.gov/pubmed/11271756.
Hawlena, D., & Schmitz, O. J. (2010). Herbivore physiological response to predation
risk and implications for ecosystem nutrient dynamics. Proceedings of the Na-
tional Academy of Sciences of the United States of America, 107(35), 15503e15507.
https://doi.org/10.1073/pnas.1009300107.
Hedrick, A. V., & Kortet, R. (2006). Hiding behaviour in two cricket populations that
differ in predation pressure. Animal Behaviour, 72(5), 1111e1118. https://doi.org/
10.1016/j.anbehav.2006.03.018.
Huntingford, F. A. (1976). The relationship between anti-predator behaviour and
aggression among conspecifics in the three-spined stickleback,. Gasterosteus
aculeatus. Animal Behaviour, 24(2), 245e260. https://doi.org/10.1016/S0003-
3472(76)80034-6.
Johnstone, R. A., & Manica, A. (2011). Evolution of personality differences in lead-
ership. Proceedings of the National Academy of Sciences of the United States of
America, 108(20), 8373e8378. https://doi.org/10.1073/pnas.1102191108.
Le Conte, Y., Mohammedi, A., & Robinson, G. E. (2001). Primer effects of a brood
pheromone on honeybee behavioural development. Proceedings of the Royal
Society B: Biological Sciences, 268(1463), 163e168. https://doi.org/10.1098/
rspb.2000.1345.
Lewkiewicz, D. A., & Zuk, M. (2004). Latency to resume calling after disturbance in
the field cricket, Teleogryllus oceanicus, corresponds to population-level differ-
ences in parasitism risk. Behavioral Ecology and Sociobiology, 55(6), 569e573.
https://doi.org/10.1007/s00265-003-0745-6.
Lima, S. L., & Dill, L. M. (1990). Behavioral decisions made under the risk of pre-
dation: A review and prospectus. Canadian Journal of Zoology, 68(4), 619e640.
https://doi.org/10.1139/z90-092.
Lorenzo-Seva, U., & ten Berge, J. M. F. (2006). Tucker's congruence coefficient as a
meaningful index of factor similarity. Methodology, 2(2), 57e64. https://doi.org/
10.1027/1614-2241.2.2.57.
Mathot, K. J., van den Hout, P. J., Piersma, T., Kempenaers, B., R
eale, D., &
Dingemanse, N. J. (2011). Disentangling the roles of frequency-vs. state-
dependence in generating individual differences in behavioural plasticity.
Ecology Letters, 14(12), 1254e1262. https://doi.org/10.1111/j.1461-
0248.2011.01698.x.
Mathot, K. J., Wright, J., Kempenaers, B., & Dingemanse, N. J. (2012). Adaptive
strategies for managing uncertainty may explain personality-related differences
in behavioural plasticity. Oikos, 121(7), 1009e1020. https://doi.org/10.1111/
j.1600-0706.2012.20339.x.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121120
McNamara, J. M., Stephens, P. A., Dall, S. R. X., & Houston, A. I. (2009). Evolution of
trust and trustworthiness: Social awareness favours personality differences.
Proceedings of the Royal Society B: Biological Sciences, 276(1657), 605e613.
https://doi.org/10.1098/rspb.2008.1182.
Montiglio, P.-O., Ferrari, C., & R
eale, D. (2013). Social niche specialization under
constraints: Personality, social interactions and environmental heterogeneity.
Philosophical Transactions of the Royal Society of London. Series B, Biological
Sciences, 368(1618), 20120343. https://doi.org/10.1098/rstb.2012.0343.
Moore, A. J., Brodie, E. D., III, & Wolf, J. B. (20 09). Interacting phenotypes and the
evolutionary process: I. Direct and indirect genetic effects of social interactions.
Society, 51(5), 1352e1362.
Myers, L., & Sirois, M. J. (2006). Spearman correlation coefficients, differences be-
tween. In Encyclopedia of Statistical Sciences. Hoboken, NJ: J. Wiley. https://
doi.org/10.1002/0471667196.ess5050.pub2.
Nakagawa, S., & Schielzeth, H. (2010). Repeatability for Gaussian and non-Gaussian
data: A practical guide for biologists. Biological Reviews of the Cambridge Phil-
osophical Society, 85(4), 935e956. https://doi.org/10.1111/j.1469-
185X.2010.00141.x.
Pezner, A. K., Lim, A. R., Kang, J. J., Armenta, T. C., & Blumstein, D. T. (2017). Hiding
behavior in Christmas tree worms on different time scales. Behavioral Ecology,
28(1), 154e163. https://doi.org/10.1093/beheco/arw140.
Rittschof, C. C., Bukhari, S. A., Sloofman, L. G., Troy, J. M., Caetano-Anoll
es, D., Cash-
Ahmed, A., et al. (2014). Neuromolecular responses to social challenge: Com-
mon mechanisms across mouse, stickleback fish, and honey bee. Proceedings of
the National Academy of Sciences of the United States of America, 111(50),
17929e17934. https://doi.org/10.1073/pnas.1420369111.
Rudin, F. S., Tomkins, J. L., & Simmons, L. W. (2017). Changes in dominance status
erode personality and behavioral syndromes. Behavioral Ecology, 28(1),
270e279. https://doi.org/10.1093/beheco/arw151.
Santostefano, F., Wilson, A. J., Araya-Ajoy, Y. G., & Dingemanse, N. J. (2016). Inter-
acting with the enemy: Indirect effects of personality on conspecific aggression
in crickets. Behavioral Ecology, 27(4), 1235e1246. https://doi.org/10.1093/
beheco/arw037.
Sih, A., Bell, A. M., Johnson, J. C., & Ziemba, R. E. (2004). Behavioral syndromes: An
integrative overview. The Quarterly Review of Biology, 79(3), 241e277. https://
doi.org/10.1086/422893.
Sih, A., & Del Giudice, M. (2012). Linking behavioural syndromes and cognition: A
behavioural ecology perspective. Philosophical Transactions of the Royal Society
B: Biological Sciences, 367(1603), 2762e2772. https://doi.org/10.1098/
rstb.2012.0216.
Sih, A., Kats, L. B., & Maurer, E. F. (2003). Behavioural correlations across situations
and the evolution of antipredator behaviour in a sunfishesalamander system.
Animal Behaviour, 65(1), 29e44. https://doi.org/10.1006/anbe.2002.2025.
Simmons, L. W. (1986). Female choice in the field cricket Gryllus bimaculatus (De
Geer). Animal Behaviour, 34(5), 1463e1470. https://doi.org/10.1016/S0003-
3472(86)80217-2.
Simmons, L. W. (1988). The calling song of the field cricket, Gryllus bimaculatus (de
geer): Constraints on transmission and its role in intermale competition and
female choice. Animal Behaviour, 36(2), 380e394. https://doi.org/10.1016/
S0003-3472(88)80009-5.
Smith, B. R., & Blumstein, D. T. (2008). Fitness consequences of personality: A meta-
analysis. Behavioral Ecology, 19(2), 448e455. https://doi.org/10.1093/beheco/
arm144.
Stamps, J. A. (2016). Individual differences in behavioural plasticities. Biological
Reviews, 91(2), 534e567. https://doi.org/10.1111/brv.12186.
Stoffel, M. A., Nakagawa, S., & Schielzeth, H. (2017). rptR: repeatability estimation
and variance decomposition by generalized linear mixed-effects models. In
S. Goslee (Ed.), Methods in Ecology and Evolution.https://doi.org/10.1111/2041-
210X.12797.
Tinghitella, R. M., Wang, J. M., & Zuk, M. (2009). Preexisting behavior renders a
mutation adaptive: Flexibility in male phonotaxis behavior and the loss of
singing ability in the field cricket Teleogryllus oceanicus.Behavioral Ecology,
20(4), 722e728. https://doi.org/10.1093/beheco/arp052.
Ulagaraj, S. M., & Walker, T. J. (1973). Phonotaxis of crickets in flight: Attraction of
male and female crickets to male calling songs. Science, 182(4118), 1278e1279.
https://doi.org/10.1126/science.182.4118.1278.
Wedell, N., & Tregenza, T. (1999). Successful fathers sire successful sons. Evolution,
53(2), 620e625. https://doi.org/10.1111/j.1558-5646.1999.tb03797.x.
Werner, E. E., Gilliam, J.F., Hall, D. J., & Mittelbach, G. G. (1983). An experimental test
of the effects of predation risk on habitat use in fish. Ecology, 64(6), 1540e1548.
https://doi.org/10.2307/1937508.
Wolf, J. B., Brodie, E. D., Cheverud, J. M., Moore, A. J., & Wade, M. J. (1998). Evolu-
tionary consequences of indirect genetic effects. Trends in Ecology &Evolution,
13(2), 64e69. https://doi.org/10.1016/S0169-5347(97)01233-0.
Wolf, M., & McNamara, J. M. (2012). On the evolution of personalities via frequency-
dependent selection. The American Naturalist, 179(6), 679e692. https://doi.org/
10.1086/665656.
Wolf, M., & McNamara, J. M. (2013). Adaptive between-individual differences in
social competence. Trends in Ecology &Evolution, 28(5), 253e254. https://
doi.org/10.1016/j.tree.2013.01.006.
Wolf, M., Van Doorn, G. S., & Weissing, F. J. (2011). On the coevolution of social
responsiveness and behavioural consistency. Proceedings of the Royal Society B:
Biological Sciences, 278(1704), 440e448. https://doi.org/10.1098/rspb.2010.1051.
Wolf, M., & Weissing, F. J. (2012). Animal personalities: Consequences for ecology
and evolution. Trends in Ecology &Evolution, 27(8), 452e461. https://doi.org/
10.1016/j.tree.2012.05.001.
Appendix
24 000
20 000
16 000
12 000
8000
4000
0
20 000
16 000
12 000
8000
4000
0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1.4
1.2
1
0.8
0.6
0.4
0.2
0
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
AA AS SS SA AA AS SS SA AA AS SS SA
DD DU UU UD DD DU UU UD DD DU UU UD
Treatment
p
rofiles
Between and within-individual variances
(d) (e) (f)
(b)(a) (c)
Fig. A1. Between- (light grey) and within-individual (dark grey) variances of the three behavioural traits for each treatment profile within both the (a, b, c) acoustic sexual signal
experiment (AA, AS, SS, SA) and (d, e, f) mechanical disturbance experiment (DD, DU, UU, UD). (a, d) Boldness, (b, e) activity and (c, f) exploration.
F. S. Rudin et al. / Animal Behaviour 138 (2018) 109e121 121
