Content uploaded by Jonathan Pruitt
Author content
All content in this area was uploaded by Jonathan Pruitt on Mar 29, 2014
Content may be subject to copyright.
ORIGINAL PAPER
Boldness is influenced by sublethal interactions with predators
and is associated with successful harem infiltration
in Madagascar hissing cockroaches
Donna R. McDermott &Michael J. Chips &
Matthew McGuirk &Fawn Armagost &
Nicholas DiRienzo &Jonathan N. Pruitt
Received: 29 August 2013 /Revised: 16 November 2013 /Accepted: 19 November 2013 /Published online: 10 December 2013
#Springer-Verlag Berlin Heidelberg 2013
Abstract One of the contemporary challenges of the behav-
ioral syndromes literature is to identify how individual varia-
tion in behavior is determined, and whether this variation
impacts ecological success. Although variation in experience
is an obvious potential driver of intraspecific behavioral var-
iation, predicting the impact of experience on suites of corre-
lated behavioral traits is less intuitive. Specifically, if experi-
ence impacts traits that shape individuals' success in multiple
contexts (e.g., foraging and anti-predator behavior), then ex-
perience could generate cross-contextual performance trade-
offs associated with behavioral spillover. In the present study,
we explore how sublethal experience with predators impacts
various aspects of male behavioral tendencies in the
Madagascar hissing cockroach, Gromphadorhina portentosa.
First, we found that males' activity level and boldness were
correlated together in the form of a behavioral syndrome.
Second, we found that repeated sublethal interactions with
predators shifted male boldness but not activity level, thus
suggesting that the syndrome's constituent traits can respond
to experience at least semi-independently. Third, we discov-
ered that although predator exposure only influenced bold-
ness, we found that boldness was highly correlated with males'
ability to obtain rewarding positions in the harems of rival
males. Taken together, our data suggest (but do not yet con-
firm) that although sublethal exposure to predators influences
only a narrow subset of male's behavioral tendencies, these
effects could still have nonintuitive consequences for males'
success in functionally dissimilar ecological contexts (i.e.,
social and sexual encounters).
Keywords Behavior .Constraint .Phenotypic plasticity .
Personality .Temperament
Introduction
Optimality theory posits that selection should operate to max-
imize individual fitness, resulting in less variable and more
highly optimized traits (Fisher 1930). Yet, individual trait
differences are a pervasive feature of most populations
(Bolnick et al. 2003;Belletal.2009;Dalletal.2012).
Individual differences in behavior are a broadly appreciated
phenomenon in behavioral ecology (Sih et al. 2004a,2012)
and occur in virtually every conceivable taxon. Individual
differences in behavior are intriguing to ecologists and evolu-
tionary biologists alike because behavior can have large im-
pacts on individual fitness (Downes 2002; Reale and Festa-
Bianchet 2003; Smith and Blumstein 2008)andwithin-
population variation in behavior can shape a diversity of
higher order ecological phenomena, including population size
and growth rates (Pruitt and Riechert 2011; Modlmeier et al.
2012), species distributions (Storfer and Sih 1998; Sih et al.
2003), range expansion (Fogarty et al. 2011), species interac-
tions (Smith and Blumstein 2010; Pruitt and Ferrari 2011),
andevenextinctionrisk(Pruittin press). Thus, under-
standing both the mechanisms that underlie behavioral
variation and the selective pressures that shape it is
important for our understanding of a variety of ecological
and evolutionary phenomena.
Communicated by W. O. H. Hughes
D. R. McDermott (*):M. J. Chips :M. McGuirk :F. Armagost :
J. N. Pruitt
Department of Biological Sciences, University of Pittsburgh,
Pittsburgh, PA 15260, USA
e-mail: drm2676@gmail.com
N. DiRienzo
Department of Neurobiology, Physiology & Behavior,
University of California, Davis, CA, USA
Behav Ecol Sociobiol (2014) 68:425–435
DOI 10.1007/s00265-013-1657-8
Sih et al. (2004a) define behavioral syndromes as consis-
tent individual differences in behavior within populations that
are correlated across time, situation, or ecological context.
Behavioral syndromes are intriguing because of their potential
to generate performance trade-offs and/or constrain the
evolvability of behavior. Numerous mechanisms can underlie
the formation of behavioral syndromes, and characterizing
these mechanisms is key to predicting the extent of their
influence. Behavioral syndromes can arise from one or more
of the following mechanisms: correlated selection for partic-
ular trait combinations, phenotypic plasticity, microhabitat-
mediated directional selection, linkage disequilibrium, physi-
cal linkage, or pleiotropy (Bell and Sih 2007;Sinnetal.2008,
2010; Sweeney et al. 2013). In cases where behavioral traits
are tightly linked (e.g., via pleiotropy), syndromes have the
potential to constrain the evolvability of behavior, but, in other
cases, syndromes may themselves represent the adaptive out-
comes of selection (Dochtermann and Dingemanse 2013).
Variation in individual experience can dramatically in-
crease behavioral variation within populations and generate
behavioral syndromes (Stamps and Groothuis 2010a,b). For
instance, older and more experienced individuals may be more
likely to explore and sample their environment (Rockwell
et al. 2012) or variation in early exposure to social cues may
increase among-individual differences in aggressiveness
(DiRienzo et al. 2012). Such shifts in behavior can have far-
reaching implications for individuals' performance in subse-
quent ecological challenges. For instance, a shift in individ-
uals' social aggressiveness may subsequently impact their
probability of winning access to mates or preferred territory
(Riechert 1978;PruittandRiechert2009), or shifts in foraging
aggressiveness could shape the amount and quality of re-
sources that individuals will acquire (Pruitt and Krauel 2010;
Mishra et al. 2011). For example, individuals may shift their
behavior as a result of experience with predators, but this shift
in behavioral tendencies could have implications for other
ecological challenges like courtship, mating, or foraging.
Such cross-contextual spillovers are intriguing because
trade-offs across contexts may help to maintain variation
in individual differences in behavior (Sih et al. 2004a,b)and/
or in the mechanisms that govern them (e.g., plasticity)
(Dingemanse et al. 2010). Such cross-contextual spill-
overs are the hallmark of behavioral syndromes (Sih et al.
2004a).
Here we explore how sublethal interactions with predators
influence the behavioral tendencies of male Madagascar
hissing cockroaches, Gromphadorhina portentosa (Insecta,
Blaberidae), and whether males' behavioral tendencies predict
their performance in other ecological contexts (i.e., social and
sexual interactions). The underlying hypothesis being that if
G. portentosa exhibits a strong signature of a behavioral
syndrome, behavioral shifts as a result of experience in one
context may compromise performance in others.
G. portentosa lives in large groups structured by a male
social dominance hierarchy. This hierarchy is maintained by
male–male contests that display a range of aggressive and
submissive behaviors (Clark and Moore 1994; Clark 1998).
Hissing cockroaches have been shown to exhibit behavioral
syndromes linking latency to move and latency to forage
(Logue et al. 2009) and also exploration of novel environ-
ments, latency to forage, and latency to recover from distur-
bance (Mishra et al. 2011). The latter study found that bold-
ness and activity were not correlated with aggression in male–
male contests.
In this study, we first independently test for a behavioral
syndrome between boldness, male–male aggression, and ac-
tivity level in a separate laboratory population of hissing
cockroaches. Second, we test how repeated sublethal interac-
tions with predators influence individuals' behavior in various
behavioral assays. Finally, we test whether variation in bold-
ness, aggressiveness, and activity level predict male perfor-
mance in a staged social–sexual encounter. In particular, we
are interested in predicting males' ability to infiltrate unfamil-
iar social groups, and whether males that entered groups
remained therein for extended time periods. Obtaining a po-
sition inside social groups is important for male G. portentosa
because grouping provides both (i) increased access to fe-
males and (ii) drastically reduced water loss (Yoder and
Grojean 1997; Clark 1998). In fact, Yoder and Grojean found
in 1997 that clusters of grouping cockroaches conserve water
twice as well as singleton individuals.
We ask the following questions: (1) Are individuals' laten-
cy to resume movement (boldness), activity in a container
(activity level), and aggressiveness in male–male contests
(aggressiveness) repeatable and/or correlated together in a
behavioral syndrome? (2) Does repeated sublethal exposure
to natural predators influence one or more of these behavioral
tendencies (boldness, activity level, aggressiveness)? (3) Is
variation in one or more aspects of male behavior associated
with their ability to join and remain in unfamiliar social–
sexual groups? Together these questions are designed to probe
whether shifts in behavioral tendencies as a result of experi-
ence in one ecological context might have cascading effects
into other, functionally dissimilar contexts.
Methods
Laboratory maintenance: focal males
Cockroaches were purchased from Fluker's Cricket Farm in
southern Louisiana, USA in the spring 2012. Males were
opportunistically selected for our study, depending on the
sex ratios of our commercial shipments and cockroach avail-
ability. Thus, we could not control for the selection of indi-
viduals from the commercial population, and we used all
426 Behav Ecol Sociobiol (2014) 68:425–435
males that appeared healthy and ambulatory upon arrival. In
captivity, hissing cockroaches can live up to 5 years. Our
specimens were shipped from Fluker’s Cricket Farm as ma-
ture reproductives, at 3–18 months of age (Fluker's Cricket
Farm, personal communications). Males were housed individ-
ually in round plastic containers (11.5 cm diameter, 11.0 cm
high), referred to hereafter as “home containers”. Home con-
tainers were lined with 1–2 cm of peat substrate. All cock-
roaches were maintained in a high humidity environmental
chamber (27 °C, 60–75 % humidity) on a 12:12 LD photope-
riod and provided a diet of freshly cut carrots, organic chick
feed (Kalmback), and water ad libitum. Prior to initiating their
trials, males were weighed using an electronic balance
(Denver Instruments PI-114).
Laboratory maintenance: social groups
We established six experimental social groups which we used
as stimuli for our cluster trials (described below). Groups were
each composed of four females and one resident male.
Resident males exhibited a range of masses (4.84–6.71 g)
and did not differ significantly in size from our focal males
(t=0.97, df=118, p=0.33). Groups were housed in large clear
Sterilite containers (60 cm×41.6 cm× 16.5 cm) which each
contained 2–4cmofpeatmossassubstrate.Thetopedgesof
groups' containers were lined with a thin layer of Vaseline to
prevent cockroaches from climbing up the side of the contain-
er. Groups were housed together for 9 weeks prior to the start
of the cluster trials. Additionally, to improve the visibility of
cockroaches during the filming of trials, we labeled males and
females with sex-specific colored sequins, which were glued
atop their prothorax. Females were labeled with a pink sequin
and males were labeled with a silver sequin.
Laboratory maintenance: stimulus predators
Three Malagasy tarantulas, Monocentropus lambertoni
(Araneae, Theraphosidae), were purchased commercially
and maintained individually in 15-quart Sterilite boxes
(43.2 cm× 28.3 cm×16.5 cm). These enclosures were lined
with approximately 3 cm crushed coconut substrate.
Tarantulas were each provided with a dome-shaped retreat
and water dish. Tarantulas were fed two 6-week-old crickets
twice per week, and their enclosures were sprayed copiously
with a water bottle during each feeding. Tarantulas were
housed in an incubator at 28 °C, 75 % humidity, and a 12:12
LD photoperiod.
Cockroach personality assays
For all behavior trials, cockroaches were removed from the
growth chamber in their individual containers and placed in a
separate, low-humidity (<15 % humidity) laboratory
environment. Two groups of males were run through the
personality assays. Personality assays were administered in
the order presented here for both groups, and 48 h was allotted
between trials. In the first group (N=33), males were run
through boldness, aggressiveness, and activity level trials,
three times each. These males were used to determine the
repeatability of each behavioral measure and to assess whether
there were correlationsacross contexts. These males were then
run through a single cluster trial, to determine whether males'
personalities were associated with their ability to obtain a
position within a foreign, unfamiliar social–sexual group.
A second set of males (N=141) was used to assess the
effects of repeated, sublethal interactions with predators on
males' boldness, aggressiveness, and activity level and to test
whether exposure to predators changed the structure of males'
behavioral syndrome. We ran predator exposure males (which
were repeatedly exposed to tarantulas) and control males once
through each trial type (boldness, aggression, activity level) in
the sequence presented here, following the application of their
predator exposure treatment. This second set of males was not
assayed repeatedly either before or after their predator expo-
sure treatments in order to avoid undue stress on the animals
and to diminish undesired effects of exhaustion or senescence.
Preliminary studies had previously demonstrated that repeat-
edly testing male G. portentosa for weeks or months on end
resulted in an unacceptably high level of mortality
(McDermott DR and Pruitt JN, Pers Obs). The behavioral
assays and/or experimental exposures of both test groups were
completed within 65 days of male's arrival in laboratory. At
the end of each trial, all components were wiped down with
70 % isopropanol solution and a paper towel.
Boldness
Male cockroaches were removed from their home containers
and placed in the center of a larger container (diameter=
18 cm, height=8 cm) containing a 1 cm substrate of child's
play sand. Males were placed beneath a small black retreat
(diameter=7 cm, height=3 cm) and allowed to acclimate for
60 s before swiftly lifting the retreat and exposing the male
(suddenly) to bright light. We then recorded males' latencies to
(i) move their antennae, (ii) move their head, and (iii) initiate
locomotion. Trials ceased when all three actions were com-
pleted or after 20 minutes. This trial is meant to replicate an
acute, stressful experience that hissing cockroaches are likely
to endure in the wild: in nature, hissing cockroaches live
beneath and/or inside of rotting logs and other debris, which
are occasionally uncovered and/or disturbed by roving pred-
ators (Pruitt pers obs, Toliary Madagascar).
We interpret longer latencies to resume movement as “fear-
ful”or “shy”behavior, whereas we deem shorter latencies to
resume movement to be “bolder”behavior. Granted, it could
be reasoned that we have the relationship between latency to
Behav Ecol Sociobiol (2014) 68:425–435 427
resume movement and boldness reversed. In other words, one
could rightfully argue that individuals that resume movement
quickly are actually fleeing, whereas those that remain mo-
tionless are exhibiting bold/aggressive behavior. We disfavor
this interpretation, however, because 100 % of individuals
started this trial in a huddled position with their heads buried
within the substrate and with their antennae tucked inward.
After some time, cockroaches slowly initiated movement by
antennating their substrate in an exploratory fashion. After this
period, cockroaches crawled slowly around the arena and
never engaged in high velocity locomotion like those exhib-
ited by fleeing males during territorial interactions
(McDermott DR and Chips MJ, Pers Obs). Finally, our meth-
odology and interpretation are modeled after personality stud-
ies performed on dozens of other animals, e.g., field crickets
(Niemela et al. 2012b), kangaroo rats (Dochtermann and
Jenkins 2007), stickleback (Bell 2005), funnel-web spiders
(Riechert and Hedrick 1993), and social spiders (Pruitt et al.
2008), as well as the disciplines of ecological “landscapes of
fear”(Stankowich and Blumstein 2005) and tonic immobility
(Mills et al. 1997). This final point is important because a
shared methodology and interpretation helps to facilitate com-
parisons among divergent test systems and studies.
Aggressiveness
Aggressiveness trials took place in circular containers (diam-
eter= 18 cm, height=8 cm) lined with several centimeters of
crushed coconut fiber substrate. The substrate used in these
trials was pretreated with female chemical signals. Fiber was
first placed in a large enclosure (60 cm× 41.6 cm×16.5 cm)
containing approximately 50 females. Female cockroaches
were allowed to roam freely over the substrate for 1 h. We
then removed the females and placed the recently trodden
substrate into smaller containers, where the aggressiveness
trials took place.
Aggressiveness trials were initiated by placing two male
cockroaches within the circular arena, approximately 3 cm
apart. We then recorded all of the antagonistic behaviors
exhibited by focal males over the next 25 minutes. Focal
males were pitted against one of 15 ringer males of unknown
behavioral tendencies, which were reused among trials.
Ringer males ranged in body mass from 4.51 to 7.23 g and
did not differ significantly in mass from our focal male pop-
ulation (t=0.23, df=127, p=0.74). Ringer males were
allowed a minimum of a 1 day resting period among trials.
Male–male aggressive behaviors included: (i) lunge, charac-
terized by a male ducking his head and charging toward the
opponent; (ii) head butt, characterized by a male pressing his
head or body to make contact with the opponent's head; (iii)
abdomen push, characterized by a male pushing his head or
body to make contact with the opponent's body; and (iv)
abdomen flick, a noncontact move where a male extends his
abdomen and directs it upward in a flicking motion. Each
behavior was given a score of 1, and male aggressiveness was
calculated as the sum of all their aggressive behaviors exhib-
ited. Notably, the behavioral displays in our studies closely
resemble those described by other investigations on G.
portentosa (Clark and Moore 1994;Clark1998).
Activity
Male cockroaches were placed in a clear circular container
(diameter=18 cm, height=8 cm) where the activity trials took
place. These enclosures were then placed atop a white sheet of
computer paper to increase cockroach visibility. This set-up
was placed beneath a motion-logging computer camera
(Lorex MC6050B cameras, AVerDiGi NV Card Series),
which automatically records the number of 5-s intervals when
cockroaches were active (e.g., orienting, walking, and
antennating). Activity trials were 1 h in duration. This test is
distinguishable from the boldness assay above because it
measures movement over a long span of time, whereas bold-
ness assays ended as soon as the individual engaged in loco-
motor behavior.
Cluster trials
Focal males were introduced into one of our six stimulus
groups on the opposing side of the enclosure from the resident
group (≈50 cm away), which was virtually always clustered at
one end of the enclosure. The behavior of the focal males was
recorded for the next 3 h. Specifically, we noted (i) whether
the resident cockroaches were clustered (i.e., all individuals
were in direct contact); (ii) the time taken for focal males to
make contact with the cluster; (iii) whether the male entered
the cluster or not, as evidenced by a male aligning his body
against one or more group members for 30 s or more (1, 0);
and (iv) whether males that entered the cluster remained
therein for the next 15 minutes. We used 15 minutes as our
cutoff because all male evictions occurred within 15 minutes,
and thus, focal males which remained in the cluster for >15
minutes remained in the cluster for the duration of the trial.
Our six stimulus groups were used in multiple trials. Thus,
stimulus groups were allotted a minimum of 24 h recovery
time between consecutive trials.
Sublethal exposure to predators
We split our second group of males into “predator exposure”
and “control”treatments (predator exposure treatment N=71,
control treatment N=70).Assignmenttotreatmentswasper-
formed randomly. Individuals in the predator exposure treat-
ment were run through a series of staged, sublethal interac-
tions with the Malagasy tarantula, M. lambertoni.
Cockroaches in the predator exposure treatment were exposed
428 Behav Ecol Sociobiol (2014) 68:425–435
to substrates containing M. lambertoni chemical cues and
physically interacted with M. lambertoni through aluminum
screens. In contrast, control individuals were set up in identi-
cal apparatuses for the same amount of time, but without the
presence of predator cues or the tarantula itself.
The exposure apparatus consisted of a large, clear plastic
Sterilite container (60 cm×41.6 cm× 16.5 cm) lined with≈
3 cm of crushed coconut fiber. For both the predator exposure
and control treatments, we misted the substrate with a spray
bottle of water to increase humidity. For the predator exposure
treatment, we (i) stocked the apparatus with substrate from an
enclosure where a M. lambertoni previously resided for 4 h
and (ii) placed a single M. lambertoni on the periphery of the
enclosure. Thus, the tarantula was free to roam around the
enclosure and approach male cockroaches. In the control
treatment, we stocked the apparatus with substrate from an
enclosure without a tarantula, and no M. lambertoni was
placed in the environment. Inside of our large enclosures, we
placed smaller round chambers (11.5 cm diameter and
11.0 cm height) containing a single male G. por tentosa .
These smaller enclosures had open bottoms which allowed
males to make direct contact with the substrate, and four
aluminum screens flanked the sides of the containers (each
screen≈5×5 cm area, 1-m
2
mesh size). Grates allowed male
cockroaches to antennate tarantulas as they passed by and
allowed tarantulas to attempt sublethal strikes at males. We
left males in their treatments in an incubator at 28 °C and 75 %
humidity for 1 h. We repeated the predator exposure and
control treatments on males three times per week for 5 weeks,
resulting in a total of 15 h of sublethal interactions with M.
lambertoni for males in the predator exposure treatment.
Statistical analysis
We estimated the repeatability of the boldness, aggressive-
ness, and activity assays using ANOVA to partition variance
into the proportion of variation explained by within- versus
between-individual differences. Repeatability is defined here-
in as the proportion of total variation explained by between-
individual differences (Boake 1989; Falconer and McKay
1996). We then tested for the presence of a behavioral syn-
drome in these Group 1 males by averaging males' latency to
initiate movement in a container, aggressiveness score, and
activity level scores across their three assays and tested for
correlations in average scores across assays using nonpara-
metric Spearman's correlations. Our syndrome analyses were
performed using the data collected from males of Group 1. In
an independent analysis, we tested for the presence of a
behavioral syndrome in our predator-exposed versus control
males using Spearman's correlations. We then qualitatively
compared the syndrome structures of control and predator
treatment cockroach groups.
To determine whether male behavioral tendencies were
associated with their ability to infiltrate social groups, we
constructed three models: one predicting the latency for males
to make contact with a mating cluster, one predicting whether
males joined the cluster or not, and one predicting whether
males remained in the cluster for the next 15 minutes or
whether the male was extirpated. For our models predicting
males' latency to make contact with the cluster and whether
males joined the cluster, we used males' latency to initiate
movement in a container, aggressiveness score, activity level
and mass as predictor variables. We used multiple regression
to predict males' latency to make contact with a cluster and
multiple logistic regression to predict whether or not males
joined clusters. For the model predicting whether males
remained in the cluster, we included males' latency to initiate
movement in a container, aggressiveness score, and mass as
predictor variables. We discarded activity level from the mod-
el predicting whether males remained in the cluster because of
a limited number of observations (N=14) and low power. This
sample size was small due to the nature of our experimental
outcome (i.e., only a small number of males actually entered
the cluster and were available for the analysis) and because of
a female-biased sex ratio in the cockroaches received from
Flukers. We used a multiple logistic regression model for
predicting whether males remained in the cluster.
We used three ttests to compare the boldness, aggressive-
ness, and activity levels of males in our predator exposure and
control treatments. All of our statistics were performed in JMP
9.0. We used a Bonferonni-corrected alpha when testing for
multiple correlations (only) in order to reduce the probability
of type I error. Three pairwise correlations were calculated for
each test group, and thus, we used a Bonferonni-corrected
alpha of 0.016 (α=0.05/3).
Results
We detected significant repeatabilities for our measures of
boldness and activity level (Table 1A); however, we failed to
detect a significant repeatability for males' aggressiveness
score (Table 1A). Additionally, all three of our boldness
measures were very highly correlated with each other (all
r=0.82–0.97, p<0.0001, corrected α=0.008). Thus, for the
remainder of our analyses, we used males' latency to initiate
locomotion as our sole measure of male boldness.
We detected the presence of a behavioral syndrome be-
tween male activity level in a container and males' latency to
initiate locomotion during our boldness assay for all three
groups of males (Table 1B–D,correctedα=0.016): Group 1
males (Table 1B), control males (Table 1C), and predator-
exposed males (Table 1D). However, we failed to detect a
significant association between either of these measures and
our measurement of male aggressiveness (all P>0.12).
Behav Ecol Sociobiol (2014) 68:425–435 429
Moreover, none of our behavioral measures were correlated
with male mass (all P>0.73).
Our combined models predicting males' latency to make
contact with a cluster was highly significant (F
4,28
=5.44, R
2
=
0.44, P=0.002). We detected a strong association between
male activity level and their latency to make contact with the
cluster (Table 2), where more active males made contact with
clusters more quickly than more sedentary males (Fig. 1). One
hundred percent of males made contact with the cluster during
the 3-h cluster trials.
Our combined model predicting whether males' joined mating
clusters was also significant (χ
4
2
=22.61, R
2
=0.45, P<0.001).
Tabl e 1 A. Repeatability esti-
mates of various behavioral mea-
sures taken on Group 1 male
G. portentosa (N=33). B–D.
Spearman's correlations testing
for correlations in male behavior-
al tendencies across ecological
contexts. Boldness was estimated
as a latency to resume movement,
where bolder individuals have
lower latencies; hence, the nega-
tive association between boldness
and activity levels means that
bolder individuals were also more
active: (B) of males from Group
1, which were not exposed to any
treatment and were run through
each assay three times (N=33)
(C) control males from Group 2,
which were run through each as-
say once (N=59); (D) predator-
exposed males from Group 2,
which were run through each
assay once (N=57)
*
A significant correlation at a
Bonferonni corrected α=0.016
A. Repeatability analyses
Activity level F
32,96
=1.71, P=0.008 r=0.38
Boldness: antennae F
32,96
=1.96, P<0.001 r=0.50
Boldness: head F
32,96
=1.94, P<0.001 r=0.50
Boldness: locomotion F
32,96
=1.97, P<0.001 r=0.51
Aggressiveness F
31,93
=0.89, P=0.63 r=0.04
B. Syndrome analyses: Group 1 males
Boldness: locomotion Aggressiveness
Activity level −0.424 −0.075
P=0.014
*
P=0.676
Boldness: locomotion 0.0272
P=0.881
C. Syndrome analyses: Group 2 control males
Boldness: locomotion Aggressiveness
Activity level −0.392 0.09
P=0.002
*
P=0.49
Boldness: locomotion 0.031
P=0.82
D. Syndrome analyses: Group 2 predator-exposed males
Boldness: locomotion Aggressiveness
Activity level −0.326 0.21
P=0.013 P=0.12
Boldness: locomotion −0.12
P=0.37
Tabl e 2 Combined model
predicting male G. portentosa
success in cluster trials using three
performance metrics: (A) the la-
tency for males to make contact
with the cluster (seconds)
(N=33), (B) whether males
joined the cluster or not (N=33),
and (C) whether males remained
in the cluster or were evicted
(N=14)
A. Latency to make contact with cluster
Estimate±SE DF Fratio P
Mass 0.55±3.91 1, 27.8 0.02 0.89
Activity level 22.96±5.13 1, 27.88 19.99 <0.001
Latency to initiate movement 0.03± 0.02 1, 27.27 2.73 0.11
Aggressiveness score 0.09±1.03 1, 27.77 0.01 0.93
B. Joined cluster
Estimate±SE DF Chi-square P
Mass −0.61± 4.88 1, 27.8 0.43 0.51
Activity level 2.32±0.95 1, 27.88 4.55 0.03
Latency to initiate movement −0.02± 0.01 1, 27.27 11.81 <0.001
Aggressiveness score −0.02±0.26 1, 27.77 0.01 0.94
C. Remained in cluster
Estimate±SE DF Chi-square P
Mass −14.65±27.95 1, 9.8 0.89 0.34
Latency to initiate movement −0.16± 0.25 1, 9.9 2.96 0.08
Aggressiveness score 7.15±0.2.12 1, 9.77 13.27 <0.001
430 Behav Ecol Sociobiol (2014) 68:425–435
Only 14 of the 33 (43 %) males that made contact with a cluster
attempted to join it by physically aligning their body with one or
more cluster members. The remaining 19 males made contact
with the cluster, antennated one or more cluster members,
reoriented, and moved away from the cluster. We found that
bolder and more active males were more likely to join clusters
than shyer and more sedentary males (Fig. 2,Table2).
Our combined model predicting whether males remained
within the cluster or not was also highly significant (χ
4
2
=
22.61, R
2
=0.45, P< 0.001). Of the 14 males which joined
clusters, only five remained therein for >15 minutes. Those
males which remained in the cluster for the next 15 minutes
remained within the cluster for the duration of the trial. In two
of these instances, we observed head butting behavior be-
tween the focal male and the resident, and the resident male
subsequently left the cluster. We found that the aggressiveness
of the focal male was associated with its probability of re-
maining in the cluster, where more aggressive males were less
likely to leave the cluster (Fig. 3, Table 2).
Finally, we detected a significant effect of our predator
exposure treatment on some aspects of male behavior but not
others. We failed to detect a significant effect of sublethal
predator exposure on male mass (t=−0.15, df=95, p=0.86),
activity level (t=0.28, df = 125, p= 0.78), and aggressiveness
score (t=0.16, df = 113, p=0.88) (Fig. 4). In contrast, we de-
tected a strong significant effect of predator exposure on males'
latency to initiate locomotion during our boldness assay (t=
2.67, df=136, p=0.008)(Fig. 4), where predator exposed males
exhibited 47 % lower boldness than control males.
Discussion
Consistent behavioral variation can often have far-reaching
and nonintuitive ecological impacts. Here, we tested for a
behavioral syndrome across contexts in the hissing cockroach
G. portentosa. We then explored how experience with
Fig. 1 The negative relationship between male activity level and latency
to discover foreign mating clusters during staged social–sexual encoun-
ters in the hissing cockroach G. portentosa (N=33)
Fig. 2 Relationships between male behavioral tendencies and their pro-
pensity to join foreign mating clusters during staged social–sexual en-
counters in the hissing cockroach G. portentosa (N= 33). To p The
relationship between male activity level and male propensity to join
foreign mating clusters. Bottom The relationship between male latency
to initiate locomotion during boldness assays and male propensity to join
foreign mating clusters
Fig. 3 The positive relationship between male G. portentosa aggressive-
ness score and their tendency to remain within foreign mating clusters
once they have entered (N=14)
Behav Ecol Sociobiol (2014) 68:425–435 431
predators shifted one or more aspects of the syndrome and
assessed whether various components of the syndrome pre-
dicted male performance in staged harem infiltration events.
Because our experiment did not test the impact of sublethal
interactions with a predator on harem infiltration directly, we
cannot conclude that experience in predator encounters will
actually diminish males' social–sexual performance.
However, our data are consistent with the hypothesis that
contextual spillover via experience could occur and, at pres-
ent, this notion does not have much in the way of empirical
support (Johnson 2013; Kralj-Fišer et al. 2013; Pruitt and
Keiser 2013). Hence, the data and arguments presented below
provide a basis for further testing of experience-mediated
cross-contextual spillover in behavioral syndromes.
Consistent with two previous studies on a different labora-
tory population of G. portentosa (Logue et al. 2009;Mishra
et al. 2011), we detected repeatable individual differences in
males' activity level and boldness, though we failed to detect a
repeatability for male aggressiveness in staged male–male
agonistic encounters. Also, we independently confirmed a
syndrome between male boldness and activity level; however,
we failed to detect an association between either of these
behavioral measures and males' behavior in staged agonistic
encounters. Thus, males' boldness and activity level appear to
be coupled together in a behavioral syndrome that is indepen-
dent of male body mass or aggressiveness. And importantly,
similar syndrome structures were recovered in all three of our
test groups (Table 1B–D), and thus, the expression of this
syndrome appears robust despite differences in males' recent
experiences. This result is somewhat at odds with the majority
of data emerging in other test systems (Bell and Sih 2007;
Sweeney et al. 2013) and instead demonstrates that this spe-
cies syndrome is remarkably robust despite variation in these
studies' laboratory environments, rearing environments, and
testing protocols (Logue et al. 2009;Mishraetal.2011).
One plausible reason why boldness and activity level are
correlated is that both traits are linked together in a general
syndrome of activity, locomotor tendencies, and/or metabolic
rate. In this way, the observed syndrome could represent an
artifact of our test protocols rather than a true cross-contextual
trait linkage. However, an equally valid counterargument can
be made that the central idea behind all of the behavioral
syndromes literature is that there is some central trait (e.g.,
like activity level or metabolic rate) than manifests in what
scientists only deem to be “independent”ecological contexts.
And, that this central underlying trait generates a degree of
interdependence in individuals' performance in different eco-
logical contexts in situ. Either way, the robustness of the
observed activity-boldness syndrome across multiple studies
from different laboratories bodes well for the G. portentosa
system's ability to address some of these central methodolog-
ical issues.
When focal males were placed in novel social–sexual
settings, different measureable aspects of the syndrome pre-
dicted different metrics of male success. In G. portentosa,
social structure is determined by a male dominance hierarchy
where dominant males obtain a disproportionately large num-
ber of mating opportunities and preferred abiotic conditions
(Yoder and Grojean 1997). Subordinate males, in contrast,
either adopt subordinate positions in the group or abandon
the cluster. Here, we found strong associations between male
behavioral tendencies and their success at harem infiltration.
Fig. 4 Pairwise comparisons
between male G. portentosa
which had repeated sublethal
interactions with a native predator
(predator expos ure, N= 71) versus
control males (control, N=70) in
various behavioral and
morphological attributes. Bars
exhibiting different letter
flaggings were significantly
different at a Bonferonni
corrected α=0.0125. Error bars
represent standard deviations
432 Behav Ecol Sociobiol (2014) 68:425–435
First, we found that highly active males discovered the novel
mating clusters in 1/3 of the time required for sedentary males
(Fig. 1). However, this finding is perhaps an artifact of our
experimental arenas, as intuitively speaking, we would predict
that more active males would stumble upon mating clusters
more rapidly in an enclosed environment. This might also
explain why activity level was the only behavioral trait corre-
lated with locating clusters. Importantly, we also detected
associations between males' behavioral tendencies and their
propensity to join and/or remain in a cluster once it was
discovered. Both activity level and boldness were correlated
with the cockroaches' propensity to join clusters. Here, highly
active males were three times more likely to join a cluster than
nonactive males, and males with intermediate or high bold-
ness universally joined clusters (Fig. 2). Finally, we found that
aggressiveness was linked with whether males remained in the
cluster, and in two such instances, the focal male actually
supplanted the resident males. This last result is somewhat
counterintuitive because we actually failed to detect a repeat-
ability of aggressiveness (r=0.04, P=0.63). How a
nonrepeatable element of behavior (i.e., aggressiveness) can
impact males' performance during harem infiltration remains
unknown. We reason that it is possible that our composite (i.e.,
average) measure of male aggressiveness may somehow be a
more informative indicator of male behavior than repeated
individual measures. Alternatively, there may some element
of individual variation in within-individual consistency that is
at play in this system (reviewed in Stamps et al. 2012;Biro
and Adriaenssens 2013). At present, we lack the data needed
to address these topics in any rigorous way in our test system.
Taken together, our findings demonstrate that male behav-
ioral tendencies are tightly linked with their behavior and
success in novel social encounters, which is particularly im-
portant in this species because obtaining a position in a cluster
both helps conserve water (Yoder and Grojean 1997) and
increases mating opportunities. In this regard, bold, active,
and aggressive males are predicted to be better infiltrators of
harems. Interestingly, an adaptive “harem infiltrator”strategy
is one plausible driver of the observed behavioral syndrome:
males with high activity levels and boldness scores might
represent a “harem infiltrator”strategy, whereas males with
low activity levels and low boldness scores might exhibit
some other, unknown tactic. How males with low activity
levels, low boldness, and low aggressiveness are maintained
within the population cannot be addressed with our present
data; however, we argue that the familiar trade-offs between
mating success and predation risk or mating success and risk
of injury could apply to this system (Magnhagen 1991).
Repeated sublethal exposure to a predator influenced some
aspects of male behavioral tendencies but not others. Males
that were iteratively exposed to the Malagasy tarantula M.
lambertoni exhibited increased latencies to initiate movement
in a container and were, thus, were deemed to be less bold.
However, repeated exposure to M. lambertoni had no detect-
able effect on male activity level or aggressiveness. Although
we cannot ruleout the possibility that the observed differences
among treatments were a consequence of preexisting differ-
ences in male boldness, we argue that the random assignment
of males to the predator exposure versus control treatments
makes this possibility unlikely. These findings are at odds
with the inference that male boldness and activity level are
locked tightly together in a rigid syndrome because if they
were, we would have expected a concomitant increase or
decrease in both boldness and activity levels as a result of
repeated exposure. Instead, we observed that one of the traits
shifted readily as a result of individuals' experience but not the
other. Thus, the syndrome's constituent traits appear to shift
semi-independently as a result of experience. Similar shifts in
behavioral tendencies have been noted in other species of
invertebrate, for instance, in the field cricket Gryllus integer
(Niemela et al. 2012a). Moreover, variation in boldness has
been linked with predation risk in a variety of vertebrate
(Downes 2002; Reale and Festa-Bianchet 2003;Biroetal.
2004) and invertebrate systems (Riechert and Hedrick 1990;
Riechert and Hall 2000; Pruitt et al. 2012), indicating that the
shifts in boldness observed in predator-exposed G. portentosa
may have performance consequences for subsequent preda-
tor–prey encounters.
Although boldness shifts independently from activity level
in predator-exposed G. portentosa , a shift in boldness alone
could still have adverse spillover effects in other ecological
contexts. Male G. portentosa in our predator-exposed treat-
ment were 47 % less bold after repeated sublethal exposure to
M. lambertoni, and our data here suggest that this shift could
have adverse effects on male performance during harem infil-
tration events. Specifically, we found that bolder individuals
were more likely to join established mating clusters, and in
two instances, bold, aggressive males were actually able to
displace resident males. However, males that adopt more risk-
averse behavioral tendencies as a result of experience in preda-
tors may, as a consequence of behavioral spillover, be less prone
to join unfamiliar mating clusters (i.e., as a result of their
decreased boldness). Thus, boldness itself could generate a
predationriskversusmatingperformance trade-off and help to
maintain phenotypic variation in wild populations. Granted, our
studies here have not directly tested whether predator-exposed
males are less prone to join foreign mating clusters; however,
that fact that both (i) predator avoidance and (ii) social–sexual
interactions are linked to male boldness suggests that such cross-
contextual spillovers could be at work in this system.
Acknowledgments We are indebted to the University of Pittsburgh,
Department of Biological Sciences (to JNP) and the Dietrich School of
Arts and Sciences Undergraduate Research Award (to DM) for
supporting this research. Additionally, would like to thank Carl Nick
Keiser, Andreas Modlmeier, and two anonymous reviewers for their
helpful comments on previous versions of our manuscript.
Behav Ecol Sociobiol (2014) 68:425–435 433
References
Bell AM (2005) Behavioural differences between individuals and two
populations of stickleback (gasterosteus aculeatus). J Evol Biol 18:
464–473
Bell AM, Sih A (2007) Exposure to predation generates personality in
threespined sticklebacks (gasterosteus aculeatus). Ecol Lett 10:828–834
Bell AM, Hankison SJ, Laskowski KL (2009) The repeatability of
behaviour: a meta-analysis. Anim Behav 77:771–783
Biro PA, Adriaenssens B (2013) Predictability as a personality trait:
consistent differences in intraindividual behavioral variation. Am
Nat 182:621–629
Biro PA, Abrahams MV, Post JR, Parkinson EA (2004) Predators select
against high growth rates and risk-taking behaviour in domestic
trout populations. Proc R Soc Biol Sci Ser B 271:2233–2237
Boake CRB (1989) Repeatability—its role in evolutionary studies of
mating-behavior. Evol Ecol 3:173–182
Bolnick DI, Svanback R, Fordyce JA, Yang LH, Davis JM, Hulsey CD,
Forister ML (2003) The ecology of individuals: incidence and
implications of individual specialization. Am Nat 161:1–28
Clark DC (1998) Male mating success in the presence of a conspecific
opponent in a madagascar hissing cockroach, Gromphadorhina
portentosa (Dictyoptera: Blaberidae). Ethology 104:877–888
Clark DC, Moore AJ (1994) Social interactions and aggression among
male madagascar hissing cockroaches (Gromphadorhina
portentosa) in groups (Dictyoptera: Blaberidae). J Insect Behav 7:
199–215
Dall SRX, Bell AM, Bolnick DI, Ratnieks FLW (2012) An evolutionary
ecology of individual differences. Ecol Lett 15:1189–1198
Dingemanse NJ, Kazem AJN, Reale D, Wright J (2010) Behavioural
reaction norms: animal personality meets individual plasticity.
Trends Ecol Evol 25:81–89
DiRienzo N, Pruitt JN, Hedrick AV (2012) Juvenile exposure to acoustic
sexual signals from conspecifics alters growth trajectory and an
adult personality trait. Anim Behav 84:861–868
Dochtermann NA, Dingemanse NJ (2013) Behavioral syndromes as
evolutionary constraints. Behav Ecol 24:806–811
Dochtermann NA, Jenkins SH (2007) Behavioural syndromes in
merriam's kangaroo rats (Dipodomys merriami): a test of competing
hypotheses. Proc R Soc B-Biol Sci 274:2343–2349
Downes SJ (2002) Does responsiveness to predator scents affect lizard
survivorship? Behav Ecol Sociobiol 52:38–42
Falconer DS, McKay TF (1996) Introduction to quantitative genetics.
Prentice-Hall, Longman
Fisher RA (1930) The genetical theory of natural selection. Clarendon,
Oxford
Fogarty S, Cote J,Sih A (2011) Social personality polymorphism and the
spread of invasive species: a model. Am Nat 177:273–287
Johnson JC (2013) Debates: challenging a recent challengeto the aggres-
sive spillover hypothesis. Ethology 119:811–813
Kralj-Fišer S, Schneider JM, Kuntner M (2013) Challenging the aggres-
sive spillover hypothesis: is pre-copulatory sexual cannibalism a
part of a behavioural syndrome? Ethology
Logue DM, Mishra S, McCaffrey D, Ball D, Cade WH (2009) A behav-
ioral syndrome linking courtship behavior toward males and females
predicts reproductive success from a single mating in the hissing
cockroach, Gromphadorhina portentosa. Behav Ecol 20:781–788
Magnhagen C (1991) Predation risk as a cost of reproduction. Trends
Ecol Evol 6:183–185
Mills AD, Crawford LL, Domjan M, Faure JM (1997) The behavior of
the japanese or domestic quail Coturnix japonica.Neurosci
Biobehav Rev 21:261–281
Mishra S, Logue DM, Abiola IO, Cade WH (2011) Developmental
environment affects risk-acceptance in the hissing cockroach,
Gromphadorhina portentosa . J Comp Psychol 125:40–47
Modlmeier AP, Liebmann JE, Foitzik S (2012) Diverse societies are more
productive: a lesson from ants. Proc. R. Soc. Biol. Sci. Ser. B
Niemela PT, DiRienzo N, Hedrick AV (2012a) Predator-induced changes
in the boldness of naive field crickets, Gryllus integer, depends on
behavioural type. Anim Behav 84:129–135
Niemela PT, Vainikka A, Hedrick AV, Kortet R (2012b) Integrating
behaviour with life history: boldness of the field cricket, Gryllus
integer, during ontogeny. Funct Ecol 26:450–456
Pruitt JN (in press) A real-time eco-evolutionary dead-end strategy is
mediated by the behavioral traits of lineage progenitors and interac-
tions with colony invaders. Ecol. Lett
Pruitt JN, Ferrari MCO (2011) Intraspecific trait variants determine the
nature of interspecific interactions in habitat forming species.
Ecology 92:1902–1908
Pruitt JN, Keiser CN (2013) Debates: the aggressive spillover hypothesis:
existing ailments and putative remedies. Ethology 119:807–810
Pruitt JN, Krauel JJ (2010) The adaptive value of gluttony: predators
mediate the life history trade-offs of satiation threshold. J Evol Biol
23:2104–2111
Pruitt JN, Riechert SE (2009) Sex matters: sexually dimorphic fitness
consequences of a behavioural syndrome. Anim Behav 78:175–181
Pruitt JN, Riechert SE (2011) How within-group behavioral variation and
task efficiency enhance fitness in a social group. Proc R Soc Biol Sci
Ser B 278:1209–1215
Pruitt JN, Riechert SE, Jones TC (2008) Behavioural syndromes and their
fitness consequences in a socially polymorphic spider, Anelosimus
studiosus. Anim Behav 76:871–879
Pruitt JN, Stachowicz JJ, Sih A (2012) Behavioral types of predator and prey
jointly determine prey survival: potential implications for the mainte-
nance of within-species behavioral variation. Am Nat 179:217–227
Reale D, Festa-Bianchet M (2003) Predator-induced natural selection on
temperament in bighorn ewes. Anim Behav 65:463–470
Riechert SE (1978) Games spiders play—behavioral variability in terri-
torial disputes. Behav Ecol Sociobiol 3:135–162
Riechert SE, Hall RF (2000) Local population success in heterogeneous
habitats: reciprocal transplant experiments completed on a desert
spider. J Evol Biol 13:541–550
Riechert SE, Hedrick AV (1990) Levels of predation and genetically
based antipredator behavior in the spider, Agelenopsis aperta .
Anim Behav 40:679–687
Riechert SE, Hedrick AV (1993) A test for correlations among fitness-
linked behavioral traits in the spider Agelenopsis aperta (Araneae,
Agelenidae). Anim Behav 46:669–675
Rockwell C, Gabriel PO, Black JM (2012) Bolder, older, and selective:
factors of individual-specific foraging behaviors in steller's jays.
Behav Ecol 23:676–683
Sih A, Kats LB, Maurer EF (2003) Behavioural correlations across
situations and the evolution of antipredator behaviour in a sunfish–
salamander system. Anim Behav 65:29–44
Sih A, Bell A, Johnson JC (2004a) Behavioral syndromes: an ecological
and evolutionary overview. Trends Ecol Evol 19:372–378
Sih A, Bell AM, Johnson JC, Ziemba RE(2004b) Behavioral syndromes:
an integrative overview. Q Rev Biol 79:241–277
Sih A, Cote J, Evans M, Fogarty S, Pruitt JN (2012) Ecological implica-
tions of behavioral syndromes. Ecol Lett 15:278–289
Sinn DL, Gosling SD, Moltschaniwskyj NA (2008) Development of shy/
bold behaviour in squid: context-specific phenotypes associated
with developmental plasticity. Anim Behav 75:433–442
Sinn DL, Moltschaniwskyj NA, Wapstra E, Dall SRX (2010) Are behav-
ioral syndromes invariant? Spatiotemporal variation in shy/bold
behavior in squid. Behav Ecol Sociobiol 64:693–702
Smith BR, Blumstein DT (2008) Fitness consequences of personality: a
meta-analysis. Behav Ecol 19:448–455
Smith BR, Blumstein DT (2010) Behavioral types as predictors of sur-
vival in trinidadian guppies (Poecilia reticulata). Behav Ecol 21:
919–926
434 Behav Ecol Sociobiol (2014) 68:425–435
Stamps J, Groothuis TGG (2010a) The development of animal personal-
ity: relevance, concepts and perspectives. Biol Rev 85:301–325
Stamps JA, Groothuis TGG (2010b) Developmental perspectives on
personality: implications for ecological and evolutionary studies of
individual differencese. Philosophical Transactions of the Royal
Society B-Biological Sciences 365:4029–4041
Stamps JA, Briffa M, Biro PA (2012) Unpredictable animals: individual
differences in intraindividual variability (iiv). Anim Behav 83:
1325–1334
Stankowich T, Blumstein DT (2005) Fear in animals: a meta-analysis and
review of risk assessment. Proc R Soc B-Biol Sci 272:2627–2634
Storfer A, Sih A (1998) Gene flow and ineffective antipredator
behavior in a stream-breeding salamander. Evolution 52:558–
565
Sweeney K, Gadd RDH, Hess ZL, McDermott DR, MacDonald L, Cotter
P, Armagost F, Chen JZ, Berning AW, DiRienzo N, Pruitt JN (2013)
Assessing the effects of rearing environment, natural selection, and
developmental stage on the emergence of a behavioral syndrome.
Ethology
Yoder JA, Grojean NC (1997) Group influence on water conservation in
the giant madagascar hissing-cockroach, Gromphadorhina
portentosa (Dictyoptera: Blaberidae). Physiol Entomol 22:79–82
Behav Ecol Sociobiol (2014) 68:425–435 435
