Content uploaded by Martin J. Whiting
Author content
All content in this area was uploaded by Martin J. Whiting on Jun 03, 2019
Content may be subject to copyright.
Can behaviour explain invasion success? A comparison between
sympatric invasive and native lizards
Isabel Damas-Moreira
a
,
*
, Julia L. Riley
b
,
1
, D. James Harris
c
,
2
, Martin J. Whiting
a
,
3
a
Department of Biological Sciences, Macquarie University, Sydney, Australia
b
Evolution &Ecology Research Centre, University of New South Wales, Sydney, Australia
c
CIBIO-InBIO, University of Porto, Porto, Portugal
article info
Article history:
Received 19 July 2018
Initial acceptance 12 September 2018
Final acceptance 7 February 2019
Available online 16 April 2019
MS. number: 18-00510R
Keywords:
biological invasions
boldness
exploration
neophobia
Podarcis sicula
To reduce the impact of biological invasions, we need to understand the behavioural mechanisms that
enable some species to be successful invaders. Testing differences in behaviour between sympatric
congeneric species with different invasive potential is an opportunity to study specific behavioural traits
associated with invasion success. Using the invasive Italian wall lizard, Podarcis sicula, and a noninvasive
congeneric, the green Iberian wall lizard, Podarcis virescens, which live in sympatry in a location that is
novel for P. sicula, we tested their exploratory behaviour, neophobia and boldness: all traits that should
promote invasion success. The invasive P. sicula was more exploratory, bold and neophilic than the
sympatric native P. virescens. Native lizards had highly repeatable behaviour, whereas in P. sicula boldness
was the only behavioural trait that was repeatable. The behavioural traits of the native species, but not
the invasive species, were correlated. A lack of correlation between behavioural traits, as well as a lack of
repeatability in two of the three behavioural traits, suggests higher levels of behavioural plasticity in
P. sicula, which may also explain the success of this lizard during invasions. Our experiment highlights
the potential importance of behavioural traits in invasions and provides insight into why P. sicula is such
a successful invader.
©2019 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Biological invasions have enormous ecological and economic
costs and are of great concern worldwide. To prevent or limit the
impact of invasions, we need to better understand what makes
some species successful, and others unsuccessful, invaders (Carere
&Gherardi, 2013; Chapple, Simmonds, &Wong, 2012; Holway &
Suarez, 1999). Recent work has begun focusing on how behav-
ioural traits at the population and individual levels promote inva-
sion success (Carere &Gherardi, 2013; Chapple et al., 2012; Holway
&Suarez, 1999; Sih, Cote, Evans, Fogarty, &Pruitt, 2012; Wolf &
Weissing, 2012). In general, invasive species have been associated
with higher levels of aggression (Downes &Bauwens, 2002; Weis,
2010), exploration and boldness (Chapple, Simmonds, &Wong,
2011; Monceau, Moreau, Poidatz, Bonnard, &Thi
ery, 2015; Short
&Petren, 2008) than noninvasive species. They may be more
likely to disperse (Rehage &Sih, 2004) and they may be more
efficient at foraging (Pintor &Sih, 2009; Short &Petren, 2008).
These behaviours are likely to promote the progress and success of
a species during different stages of the invasion process (Chapple
et al., 2012). For example, high levels of boldness and exploration
can determine whether individuals leave their native range, enter a
human transport vector and exit in a novel location (Chapple et al.,
2011; 2012). Once in a new environment, the establishment of a
species is often associated with higher levels of boldness and
exploration (Chapple et al., 2012; Monceau et al., 2015; Short &
Petren, 2008), and lower levels of neophobia (Candler &Bernal,
2015; Griffin, Guez, Federspiel, Diquelou, &Lermite, 2016). These
traits could promote the exploitation of resources and also give
invasive species an advantage over native ones. During establish-
ment, higher aggression and foraging levels can also give invasive
species a competitive advantage over native species, which may
increase their establishment success (Downes &Bauwens, 2002;
Weis, 2010). After establishment, the expansion of a population's
range can depend on the individual's affinity for dispersal, its
boldness and exploratory behaviour, aggression levels and socia-
bility (Cote, Fogarty, Weinersmith, Brodin, &Sih, 2010; Gruber,
*Correspondence: I. Damas-Moreira, 205b Culloden Road, Sydney, NSW 2109,
Australia.
E-mail address: isabeldamas.m@gmail.com (I. Damas-Moreira).
1
Present address: Department of Botany and Zoology, Stellenbosch University,
Stellenbosch, South Africa. E-mail address:julia.riley87@gmail.com (J.L. Riley).
2
E-mail address:james@cibio.up.pt (D.J. Harris).
3
E-mail address:martin.whiting@mq.edu.au (M.J. Whiting).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
https://doi.org/10.1016/j.anbehav.2019.03.008
0003-3472/©2019 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Animal Behaviour 151 (2019) 195e202
Brown, Whiting, &Shine, 2017a; Michelangeli, Smith, Wong, &
Chapple, 2017; Rehage &Sih, 2004).
An individual's personality (i.e. repeatable behaviours across
time; R
eale, Reader, Sol, McDougall, &Dingemanse, 2007; Wolf &
Weissing, 2012) can be correlated across contexts, forming a
‘behavioural syndrome’(Chapple et al., 2012; Sih, Bell, Johnson, &
Ziemba, 2004). For example, an individual that is bold in terms of
its activity and use of habitat (e.g. active for long periods and/or
uses more open habitat) may also be bold when confronted with a
predator (e.g. flees late). If this correlation enhances a species' in-
vasion success it is termed an ‘invasion syndrome’(Chapple et al.,
2012), and can determine the invasion success of a species (Cote
et al., 2010; Dame &Petren, 2006; Michelangeli et al., 2017;
Monceau et al., 2015; Pintor, Sih, &Kerby, 2009, 2008; Rehage &
Sih, 2004). For example, the invasion success of the signal cray-
fish, Pacifastacus leniusculus, is because of a positive correlation
between aggression and foraging activity (Pintor et al., 2009). The
invasion success of the mosquitofish, Gambusia affinis (Cote et al.,
2010), and Asian hornet, Vespa velutina (Monceau et al., 2015),
can also be explained by a positive correlation between their
boldness, activity and exploratory behaviour. Interestingly, the
correlation found in invasive V. velutina was also found in the native
Vespa crabro (with the invasive species exhibiting greater boldness,
activity and exploration), yet the invasive species did not behave
consistently while the native species did (Monceau et al., 2015).
This suggests that invasive species might be more plastic in their
behaviour than native species, which can be a significant advantage
when dealing with new challenges and opportunities that arise
from novel environments (Sih et al., 2012). The native species are
likely to experience consistent selection pressure on behavioural
traits, which promotes repeatability across time (Sih et al., 2012).
Behavioural differences between a successful invader and a
congeneric native species can shed light on behaviour that pro-
motes invasion success (Holway &Suarez, 1999; Rehage, Barnett,
&Sih, 2005). If both congeneric species with different invasive
potential are sympatric, then any environmental effects are
minimized (as they live sympatrically), and there is an oppor-
tunity to better understand behavioural traits associated with
invasion success. This comparison not only gives insight into why
particular species are successful invaders, but also helps us un-
derstand any potential impacts on native species (Carere &
Gherardi, 2013; Holway &Suarez, 1999; Phillips &Suarez,
2012). The Italian wall lizard, Podarcis sicula, is an invasive spe-
cies native to the Italian Peninsula and the Adriatic coast but is
established in several countries outside its native range (CABI,
2018; Carretero &Silva-Rocha, 2015). It spreads mainly through
human transport vectors (Carretero &Silva-Rocha, 2015; Kraus,
2009; Silva-Rocha, Salvi, &Carretero, 2012; 2014), reaching high-
density populations and affecting native lizards in new locations
(Capula, 1993, 2002; Carretero &Silva-Rocha, 2015; Downes &
Bauwens, 2002; Kraus, 2009). About 20 years ago, a population
was accidentally introduced to Lisbon, Portugal, from Tuscany,
Italy (Gonz
alez de la Vega, Gonz
alez-García, García-Pulido, &
Gonz
alez-García, 2001; Silva-Rocha et al., 2012). This popula-
tion lives in sympatry with the congeneric green Iberian wall
lizard, Podarcis virescens, but not in syntopy (although they can
be found as close as 50 m from each other; Ribeiro &S
a-Sousa,
2018). Podarcis virescens is a noninvasive lizard that can occur in
transport hubs but has never been recorded in an established
population outside its native range. These two congeneric species
thus have different invasive potential and live under very similar
environmental conditions (i.e. similar predation pressure, ur-
banization level, habitat, humidity, temperature and prey avail-
ability) which makes them a model system to study the role
behaviour plays in facilitating invasion success.
We quantified three behaviours typically associated with a
species' invasive potential, exploration, neophobia and boldness
(Chapple et al., 2012; Griffin et al., 2016; Phillips &Suarez, 2012),
and tested how they differed between the invasive P. sicula and the
native P. virescens. We predicted that P. sicula would be more
exploratory, bolder and less neophobic, given its potential to travel
to new places, prosper there and displace native species (CABI,
2018). We also investigated the repeatability of the behavioural
traits in each species. We expected P. sicula individuals to be less
repeatable than P. virescens individuals in their behavioural traits,
indicating more plasticity in the invasive species. Finally, we
investigated whether behavioural traits were correlated within
each species, to explore the existence of ‘behavioural’and/or ‘in-
vasion’syndromes.
METHODS
Study Species and Captive Conditions
We collected 26 P. sicula and 29 P. virescens from the wild and
assayed their behavioural traits. All lizards were adult males with
complete or completely regenerated tails. To ensure all lizards were
adults, we only collected lizards with femoral pores, and that were
larger (snoutevent length, SVL) than the known minimum SVL at
sexual maturity for these two species (male P. sicula: 55 mm,
Capula, Luiselli, &Rugiero, 1993; male P. virescens: 40 mm, Geniez,
S
a-Sousa, Guillaume, Cluchier, &Crochet, 2014). During capture, we
also tried to control for lizard size as much as possible in order to
reduce this factor as a potential confounding effect in our study:
P. sicula ranged from 59.53 to 80.75 mm SVL (average SVL ¼74.09
mm, SE ¼1.09 mm) and P. virescens ranged from 47.87 to 61.05 mm
SVL (average SVL ¼55.54 mm, SE ¼0.77 mm). We also weighed
lizards (P. sicula ¼8.96 ±0.35 g and P. virescens ¼3.64 ±0.14 g),
and measured tail size (P. sicula ¼115.69 ±4.63 mm and
P. virescens ¼86.97 ±3.23 mm). We did not test females as the
reproductive status of wild-caught females (whether gravid or not
or recently postparturient) cannot be assessed with certainty, and
the different hormones acting in each reproductive stage could
greatly influence their behaviour. Male Podarcis spp. typically
copulate from March to July, and testosterone levels tend to be
synchronous within a locality and breeding season, reducing this
problem in males of our model species (Carretero, 2006).
We collected all lizards (both species) during June (spring) from
Parque das Naç~
oes, Lisbon, Portugal (38
45
0
41.7N, 9
5
0
43.8
00
W) on 2
different days, 2 weeks apart. We assigned these lizards to two
separate groups (1 or 2) based on collection date. Lizards were
transported to CIBIO-InBIO, at the University of Porto, and accli-
matized to captivity for 2 weeks, while being fed every other day
with three live mealworms. During the experimental period (2 days
at a time), lizards were fed the day before trials commenced and at
the end of the second day, after trials had finished (Fig. 1). This also
helped with controlling individual variability in mass during ex-
periments. Animals were kept in individual terraria
(200 300 mm and 200 mm high) at 20e22
C with constant ac-
cess to water and a small brick shelter. The room had indirect
natural light, as well as artificial lamps set for a photoperiod of 12 h
(0800e2000). A 50 W heat cable beneath the enclosure created a
thermal gradient.
Experimental Set-up
We separately measured exploration, neophobia and boldness.
We quantified each lizard's exploratory behaviour within a novel
environment containing four shelters (Fig. 1) by measuring its ac-
tivity and exploration of the shelters. We measured neophobia as
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202196
the minimum distance a lizard would get to a novel object (Fig. 1);
to measure this distance, we drew a black circle in the centre of the
arena, and subsequent circles separated by 2 cm (all circles were
drawn beneath the arena). Lastly, we quantified boldness as the
amount of time the lizard would take to leave a suboptimal refuge
(Fig. 1). All trials were conducted in an experimental arena
(320 480 mm and 300 mm high); this was a clear plastic
container covered with white paper on the outside. Each trial was
carried out three times per individual and replicates were 1 week
apart. We randomly allocated lizards to one of four different
batches across the day (batch ¼1, 2, 3 or 4), because the number of
lizards that could fit in our experimental room was limited. We
measured exploration and neophobia on the same day (day 1) and
boldness on the following day (day 2; Fig. 1). At the beginning of
each experimental day, lizards were removed from their enclosure
and transferred to the centre of the experimental arena. After the
neophobia and boldness tests (on different days), lizards were
returned to their home enclosures.
Lizards always had access to shelters (black plastic containers:
80 120 mm and 50 mm high), with a small opening on one side
(40 25 mm). We cleaned all cage materials with 96% ethanol
between trials. All trials were remotely video-recorded with CCTV
cameras.
Exploration
At the beginning of the exploration trial, each lizard was placed
under a closed, opaque plastic shelter (circular, 100 mm diameter
and 70 mm high) for 5 min, to acclimatize. The arena consisted of
four shelters placed along the four sides of the enclosure with the
opening facing the centre of the arena (e.g. Gruber et al., 2017a). The
trial began when we remotely lifted the shelter using wire, so the
lizard could not see the researcher, and lasted for 30 min. For each
exploration trial (N¼3) we used a different substrate: first dark
pine bark, then white sand and then soil. We scored four measures
related to exploration: the time spent active (s), time in hiding (s),
the frequency of entering the shelters (count) and the number of
unique shelters visited during the trial (range 0e4; Table 1). We
used the program BORIS (Friard &Gamba, 2016) to measure the
time lizards spent active and in hiding. To create one exploration
score for analysis, we performed a principal component analysis
(PCA) summarizing our four exploration measures using the prin-
comp function (Jolliffe, 2002) in R version 3.4.2 (R Core Team, 2017).
Because these variables have different scales, the PCA used a cor-
relation matrix to standardize the data (Jolliffe, 2002). The first
principal component (PC1) explained 52% of the variance in these
four traits, and so we used PC1 in all statistical analyses as the
exploration score. The time spent active, frequency of entering a
shelter and the number of unique shelters visited negatively loaded
on PC1, while the time spent hiding positively loaded on PC1
(Table 2). Therefore, as our exploration score decreased, lizards
were more exploratory.
Neophobia
Once the exploration trial finished, we ushered the lizard into
the closed refuge. We then placed a novel object in the centre of the
arena (Fig. 1). After 5 min, we lifted the closed shelter using the
same method as before. We recorded the lizard's behaviour for
30 min and later scored the minimum distance between the lizard
and the novel object (e.g. White, Meekan, McCormick, &Ferrari,
2013) using the circles in the arena. If the lizard contacted the
object, it was given a score of 0 cm (Table 1). For each replicate of
the neophobia trial (N¼3) we used a different novel object pre-
sented in the order: white nonperfumed candles in foil, yellow
cupcake paper and blue plastic clothes pegs. We chose these objects
because lizards are unlikely to encounter them in the wild and
because they look very different. In each replicate, the new object
was placed on a novel substrate, because the preceding experiment
was exploration, which required novelty in the arena. To control for
a potential effect of this combination on the neophobia test, we
always used the same combination, and in the same order, for all
lizards. For the statistical analyses, we applied a rank trans-
formation to our neophobia score to normalize the data (Riley,
Noble, Byrne, &Whiting, 2017). As the neophobia score
Exploration Neophobia Boldness
Consecutive days Day 2)Day 1)
1 week between replicates
Ice pack
Figure 1. Diagram of the three behavioural trials, and their arena set-up, which were carried out on 2 consecutive days. Regardless of the trial, all arenas always had a 100 W
halogen light bulb suspended on one side. Each trial was replicated three times per individual in the same sequence, 1 week apart. Day 1 refers to the first experimental day in each
week during which we measured exploration and neophobia. The solid grey circle in the middle of the arena for the neophobia trial was where we placed the novel object. Day 2
refers to the second experimental day in which boldness was measured. All trials took 30 min.
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202 197
decreases, it reflects more neophilic behaviour, because the lizard is
closer to the novel object.
Boldness
For this trial, the experimental arena had white paper as a
substrate, one shelter under the heat lamp to create a hot
(‘optimal’) shelter and another at the opposite end of the enclosure
(Fig. 1). An ice pack was placed under this shelter, to create a cold
(‘suboptimal’) refuge (e.g. Riley et al., 2017). The mean substrate
temperature inside each shelter was T
hot
¼30
C and T
cold
¼9
C.
We placed the lizard in the arena and, after 20 min, we started the
boldness trial by lifting the hot refuge and, using a nitrile-gloved
hand, we gently tapped the lizard repeatedly to scare it inside the
cold shelter. Once inside the cold shelter, the hot shelter was placed
back in the arena. We then recorded the lizard's behaviour for
30 min and scored boldness as the latency from when the lizard
entered the cold shelter until it emerged (Table 1). We believe both
species would be similarly impacted by the suboptimal refuge
given both inhabit small refuges in the wild and have similar
preferred temperatures (Carretero, 2015).
Ethical Note
We collected lizards from the wild using noosing, which is a
standard method for catching many lizard species. A dental floss
noose is attached to the end of a telescopic pole and dropped over
the lizard's head, before lifting the lizard off the ground. We only
collected lizards that appeared healthy and had complete tails. We
endeavoured to keep stress to a minimum by only handling lizards
when we needed to move them to experimental arenas and by
keeping them individually, thereby avoiding any social conflict. We
did have to gently tap lizards on the tail base during the boldness
assay, but this only resulted in a short-term behavioural response
and lizards quickly resumed normal behaviour after they were
returned to their home enclosure. All lizards had access to shelter
and a heat source, even during behavioural trials. Also, all lizards
ate food (mealworms) when it was offered. After the experiment,
all P. sicula were euthanized by injection of sodium pentobarbital,
as required by the Portuguese Institute for Conservation of Nature
and Forests (ICNF) because they are an invasive species and all
P. virescens were returned to their exact capture location. Our
protocols and research were approved by the Macquarie University
Animal Ethics Committee (ARA2017/004) and by the Portuguese
ICNF (Licence 428/2017/CAPT).
Statistical Analysis
All analyses were performed in R version 3.4.2 (R Core Team,
2017). Prior to analyses we explored our data to ensure it fitted
model assumptions using the protocol described in Zuur, Ieno, and
Elphick (2010). During this process, we found significant differ-
ences in SVL (R
2
¼0.789,
b
¼18.544 ±1.318, t
1,53
¼14.070,
P<0.001) and mass (R
2
¼0.794,
b
¼5.285 ±0.373,
t
1,52
¼14.160, P<0.001) between species. We therefore did not
include SVL or mass in our models, to ensure the assumption of no
collinearity between variables was met. Data for this study are
accessible at Figshare: https://doi.org/10.6084/m9.figshare.
7822100.v1.
Differences between species
We examined exploration, neophobia and boldness separately,
but used the same analyses. In each model, we tested whether the
behavioural trait differed between species using a linear mixed-
effects model (LMM, using the function lmer from the lme4 R
package; Bates, M€
achler, Bolker, &Walker, 2015). These models
incorporated the fixed effects of species (P. sicula or P. virescens), the
trial day (the day the trial occurred on, with 1 being the first
experimental day), the experimental group (collection day 1 or 2)
and the batch (1e4). We also initially tested for an interaction be-
tween the fixed effect of species and trial day, which was removed
and the model rerun if the effect was nonsignificant. The contin-
uous predictor variable trial day was z-transformed prior to anal-
ysis to standardize the variable and facilitate interpretation of
interactions if present (Schielzeth, 2010). We also included each
lizard's identity as a random effect (intercept only) in the model to
control for dependencies in the data due to repeated behavioural
trials on the same lizards. When we plotted our data (see Fig. 2), we
set the fixed effect of batch to intercept level values because we
were not interested in visualizing this effect (we chose group 2 for
visualization purposes because it contained more individuals to
ensure better estimation of variance components), and we did not
include the variance from the random effect of lizard identity.
Table 1
Behaviours scored in each trial
Trial Behaviours scored from videos Variables
Exploration Activity: time (s) spent moving in the arena (0e1800 s) Exploratory score (PC1):
more sedentary with values [
more exploratory with values Y
Hiding: time (s) spent inside a shelter (0e1800 s)
Shelter frequency: number of times entered all shelters (0 to unlimited)
Shelter number: number of shelters visited (0e4)
Neophobia Minimum distance: the minimum distance (cm) a lizard would get to the
new object during the trial (0e14
þ
cm)
Minimum distance (Rank transformed):
more neophobic with values [
more neophilic with values Y
Boldness Latency to leave shelter: time (s) to emerge from the cold shelter (0e1800 s) Latency to leave shelter:
shyer with values [
bolder with values Y
Behaviours were scored from remotely recorded videos of the exploration, neophobia and boldness trials. Explanations of each of the parameters we measured, as well as the
response variables used in statistical analyses, and how we interpreted them, are listed.
Table 2
Principal component analysis (PCA) to form an exploratory score
Exploratory behaviours PC1 loadings
Activity 0.5026
Hiding 0.2341
Shelter frequency 0.5864
Shelter number 0.5905
Exploratory behaviours were combined using a PCA to form an
exploratory score (N
sicula
¼78 and N
virescens
¼87). The first principal
component (PC1) explained 52% of the variation in these four be-
haviours and was used in statistical analysis as our exploratory score.
Higher values of the PC1 reflect less exploratory lizards.
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202198
Repeatability of behavioural traits
We estimated the consistency of behavioural traits by calcu-
lating adjusted repeatability (R
adj
) with 95% confidence intervals
(CIs) for each behavioural trait (exploration, neophobia and bold-
ness), for each species separately. We used the rpt function from the
R package rptR with 1000 permutations and accounted for the
same covariates used in our LMM models (Nakagawa &Schielzeth,
2010). To assess whether R
adj
was significant, we visually examined
whether the 95% CIs for each estimate included 0. We also
compared differences in repeatability between species by visually
examining overlap in 95% CIs.
Correlations between traits
We also investigated correlations between the behavioural traits
for each species separately. This analysis was restricted by the
sample size of our study, and we were unable to account for de-
pendencies within our data (i.e. repeated measures of the same
individuals) or additional explanatory variables (i.e. experimental
group, batch, etc.) in this analysis. However, this analysis may afford
some insight into the correlations between behavioural traits in our
study species. We calculated Spearman rank-order correlations
between all behavioural traits using the function cor in the R
package stats, and then used the function cocor.indep.groups from
the R package cocor to test for significant differences between
species in trait correlations using Fisher's ztests (Diedenhofen &
Musch, 2015).
RESULTS
Behavioural Differences Between Species
Podarcis sicula were significantly more exploratory than
P. virescens and became more exploratory as trials progressed,
while P. virescens became less exploratory (Table 3,Fig. 2a). Neither
group (collection day) nor batch significantly affected exploratory
behaviour (Table 3).
In the neophobia trial, P. sicula got significantly closer to the
novel object than P. virescens (Table 3,Fig. 2b) and was therefore
more neophilic than P. virescens. We found no effect of trial day or
group (collection day) on neophobia score, but we did find a sig-
nificant effect of batch; batch 4 was significantly less neophobic
than batch 1 (Table 3). We found substantial individual variation in
the response to a novel object: some animals never made contact
with the novel object and had long minimum distances, while
others passed by the object, briefly touching it without paying
much attention and some lizards explored the novel object through
tongue flicking and even climbed onto it. During the neophobia
trial, 21.8% (N
trials
¼19/87; N
individuals
¼29) of P. virescens and 37.2%
(N
trials
¼29/78; N
individuals
¼26) of P. sicula explored the novel
object.
During the boldness trials, latency to emerge from the cold
shelter after being scared was significantly shorter for P. sicula than
for P. virescens (Table 3,Fig. 2c). The model indicated a significant
effect of batch, where batch 3 was significantly bolder than batch 1.
There was no effect of group (collection day) or trial day (Table 3).
Repeatability of Behavioural Traits
Podarcis virescens was significantly repeatable in all three
behavioural traits (Fig. 3), whereas P. sicula was significantly
repeatable in their boldness (Fig. 3). The species were not signifi-
cantly different from each other in the repeatability of behavioural
traits (Fig. 3).
Correlations Between Traits
The behavioural traits measured in P. virescens were significantly
correlated; individuals that exhibited more exploratory behaviour
were also bolder and less neophobic (Table 4). We did not find any
significant correlation between P. sicula behavioural traits (Table 4),
and the difference between species’behavioural trait correlations
were all nonsignificant (Table 4).
DISCUSSION
Overall, our predictions for the behavioural differences between
an invasive and a native species were supported. The invasive
species, P. sicula, was more exploratory, bolder and less neophobic
than the native P. virescens.Podarcis virescens were highly repeat-
able in their behaviours, while only boldness of P. sicula was
repeatable. In the native P. virescens there were correlations be-
tween all behavioural traits, with more exploratory individuals also
being bolder and less neophobic. In contrast, there were no sig-
nificant correlations between behavioural traits in P. sicula.
–2
–1
0
1
2
110 20
Measure of exploration (PC1)
–0.5
0
0.5
1
110 20
1800
1200
600
0
21121
Distance to object (rank transformed)Latency to leave cold shelter (s)
Test da
y
s
Sedentary
Explorator
y
Neophobic
Neophilic
Shy
Bold
(a)
(b)
(c)
Figure 2. Behavioural differences between the invasive P. sicula (orange) and the
native P. virescens (blue) for (a) exploration, (b) neophobia and (c) boldness. We plotted
fitted lines predicted from our linear mixed-effect models with 95% confidence in-
tervals (shaded polygon).
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202 199
Being bolder, more exploratory and neophilic probably en-
hances the ability of P. sicula to be successful during all aspects of
the invasion process. For example, bolder and more exploratory
behaviour may increase an invader's likelihood of entering a
transport vector, and thereby colonize a new environment (Chapple
et al., 2012; Griffin et al., 2016). These behaviours might have been
similarly expressed and important for P. sicula when they were
introduced 20 years ago, as these behavioural traits are likely to be
heritable (Gruber, Brown, Whiting, &Shine, 2017b; R
eale et al.,
2007). We cannot exclude the possibility that selection might
have acted on these behavioural traits during the invasion process,
leaving only the bolder, more exploratory and neophilic individuals.
Nevertheless, the differences between traits we found in this study
might favour the invasive species when it has to compete for food
and habitat (by more easily finding and exploiting new resources,
such as food and shelter sites), and increase its likelihood of
interacting with other lizards (by being bolder for example), which
can promote aggressive encounters with native species (Candler &
Bernal, 2015; Gruber et al., 2017a; Rehage &Sih, 2004; Short &
Petren, 2008; Sol, Timmermans, &Lefebvre, 2002). Indeed, the
displacement of P. virescens from gardens inhabited by P. sicula
(Ribeiro &S
a-Sousa, 2018), or of other Podarcis species native to
other invaded sites, may be explained in part by higher levels of
aggression in this species (Downes &Bauwens, 2002), which can
result in competitive exclusion of native species (Nevo et al., 1972;
Valde
on, Perera, Costa, Sampaio, &Carretero, 2010).
Podarcis virescens was consistent in all behavioural traits, while
P. sicula was only consistent in their boldness. This was a similar
result to a study of hymenopterans, where the native wasp V. crabro
was repeatable for activity, boldness and exploration, but the
invasive V. velutina was not (Monceau et al., 2015). The invasive
P. sicula may benefit from being more plastic in its behaviour
because invasive species in general have to respond to changing,
novel environments (Griffin et al., 2016). Podarcis sicula is usually
unintentionally transported to new locations (Carretero &Silva-
Rocha, 2015) and very successful at adapting to new conditions
(Herrel et al., 2008; Vervust, Grbac, &Van Damme, 2007). Its
behavioural plasticity may thus partly explain this species’invasion
success. Interestingly, P. sicula did exhibit repeatability in boldness,
which potentially highlights the importance of boldness in all
stages of the invasion process in this species. For example, bold
individuals may also be more likely to be transported outside their
native range because they are more likely to enter containers or
vessels being prepared for transport (e.g. olive trees; Rivera,
0
0.2
0.4
0.6
0.8
1
Exploration Neophobia Boldness
Behavioural traits
Adjusted repeatability (Radj)
Figure 3. Adjusted repeatability (R
adj
) and 95% confidence intervals of behavioural
traits (exploration, neophobia and boldness) for the invasive P. sicula (N
obs
¼78,
N
ind
¼26, orange circles) and the native P. virescens (N
obs
¼87, N
ind
¼29, blue circles).
The dashed line represents 0.
Table 3
Outcomes of the linear mixed-effects models for each behavioural trait: exploration, neophobia and boldness
Exploration Neophobia Boldness
N
obs
¼165, N
ind
¼55 N
obs
¼165, N
ind
¼55 N
obs
¼165, N
ind
¼55
b
SE tP
b
SE tP
b
SE tP
Fixed effects
Intercept ¡1.089 0.298 ¡3.662 <0.001 ¡0.190 0.180 ¡1.054 0.292 650.455 172.600 3.769 <0.001
Species (virescens;
REF ¼sicula)
1.451 0.254 5.719 <0.001 0.582 0.151 3.858 <0.001 561.030 143.234 3.917 <0.001
Trial day ¡0.337 0.121 ¡2.797 0.005 ¡0.042 0.054 ¡0.777 0.437 32.293 32.925 0.981 0.327
Group (2; REF ¼1) 0.328 0.263 1.244 0.213 0.047 0.157 0.297 0.766 139.666 151.958 0.919 0.358
Batch (2; REF ¼1) 0.135 0.219 0.613 0.540 ¡0.058 0.142 ¡0.407 0.684 ¡10.293 116.764 ¡0.088 0.930
Batch (3; REF ¼1) 0.248 0.266 0.933 0.351 0.056 0.170 0.328 0.743 ¡245.284 124.279 ¡1.974 0.048
Batch (4; REF ¼1) 0.212 0.256 0.827 0.409 ¡0.566 0.167 ¡3.399 0.001 ¡75.716 131.157 ¡0.577 0.564
Species*Trial day
(virescens; REF ¼sicula)
0.597 0.166 3.592 <0.001 eeeee e ee
Random effects
s
2
SE
s
2
SE
s
2
SE
Lizard identity 0.470 0.053 0.139 0.029 207288.000 35.445
Residual 1.104 0.082 0.472 0.054 174750.500 32.541
When the interaction effect was nonsignificant, which is indicated by ‘¡’, we reran the model without this effect. Significant results are in bold.
Table 4
Spearman rank order correlations and 95% confidence intervals between behavioural traits for each species
Exploration vs boldness Boldness vs neophobia Neophobia vs exploration
P. sicula 0.0793 (0.1439, 0.3039) 0.1481 (0.1023, 0.3676) 0.1917 (0.0358, 0.4041)
P. virescens 0.2452 (0.0432, 0.4250) 0.3043 (0.1045, 0.4949) 0.3708 (0.1534, 0.5501)
Species comparison (Fisher's ztest) z¼1.32, P¼0.185 z¼1.04, P¼0.299 z¼1.23, P¼0.219
Significant results are in bold.
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202200
Arribas, Carranza, &Maluquer-Margalef, 2011; Silva-Rocha et al.,
2012). After introduction in a novel location, bolder individuals may
gain greater access to resources and do better in social interactions
(Monceau et al., 2015; Pintor, Sih, &Bauer, 2008; Short &Petren,
2008). Moreover, P. sicula usually invades urbanized locations
perhaps because they are in or near transport hubs (the introduced
population in Lisbon inhabits city gardens), and boldness confers an
advantage in urban environments because it can translate into
higher foraging efficiency (Short &Petren, 2008).
Behavioural traits of P. virescens were correlated with more
exploratory individuals also being bolder and less neophobic,
which suggests a possible behavioural syndrome in this native
species (Sih et al., 2004). However, the same was not true for
P. sicula, for which we did not find any correlations between their
behavioural traits. The correlations between behavioural traits we
found in this study should be interpreted with caution, however,
because we did not control for dependency among variables or
additional sources of variation. It is also important to consider both
within- and between-individual correlations in behavioural traits,
to understand the selective forces acting on behavioural traits
within a population and their evolutionary significance
(Dingemanse &Wolf, 2013; Sih et al., 2012). Nevertheless, the lack
of correlation between traits, allied with the overall inconsistency
in P. sicula's repeatability, may be advantageous during biological
invasions. Variation in behavioural traits within a population in-
crease the likelihood of success in fluctuating environments and
novel habitats and allow for a population's persistence in novel
environments (Dingemanse &Wolf, 2013; Sih et al., 2012). Like-
wise, correlations between behavioural traits constrain a popula-
tion, because if selection acts on one trait, correlated behaviours are
also likely to be affected (Sih et al., 2012).
In conclusion, we have shown that congeneric invasive and
native lizard species differed in key behavioural traits, explora-
tion, neophobia and boldness, that could promote the invasion
success of P. si cul a . These behavioural traits are likely to be
important for the success of other introduced P. s icu l a populations
given that these populations share the same long-range transport
and establishment pattern (CABI, 2018; Carretero &Silva-Rocha,
2015). Increasingly, behavioural mechanisms are being appreci-
ated as playing an important role in determining invasion success
(Chapple et al., 2012). We also suggest that comparisons between
closely related species that are variable in their invasive ability
may provide important insights into the relationship between
plasticity and personality and their relative roles in determining
invasive success.
Acknowledgments
We thank Bernardino Silva, Bruno Pleno, Jo~
ao Damas and F
abio
Sousa for their invaluable help, advice and support in this research
and two anonymous referees for their valuable comments. This
research was funded by Macquarie University (PhD scholarship to
I.D.M.) and J.L.R. was supported by an Endeavour Postdoctoral
Fellowship.
References
Bates, D., M€
achler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects
models using lme4. Journal of Statistical Software, 67,1e48. arXiv:1406.5823v1.
CABI. (2018). Podarcis sicula [original text by Silva-Rocha, I.]. In Invasive Species
Compendium. Wallingford, U.K.: CAB International. www.cabi.org/isc.
Candler, S., & Bernal, X. E. (2015). Differences in neophobia between cane toads
from introduced and native populations. Behavioral Ecology, 26(1), 97e104.
https://doi.org/10.1093/beheco/aru162.
Capula, M. (1993). Natural hybridization in Podarcis sicula and P. wagleriana (Rep-
tilia: Lacertidae). Biochemical Systematics and Ecology, 21(3), 373e380. https://
doi.org/10.1016/0305-1978(93)90028-P.
Capula, M. (2002). Genetic evidence of natural hybridization between Podarcis
sicula and Podarcis tiliguerta (Reptilia: Lacertidae). Amphibia-Reptilia, 23(3),
313e321. https://doi.org/10.1163/15685380260449199.
Capula, M., Luiselli, L., & Rugiero, L. (1993). Comparative ecology in sympatric
Podarcis muralis and P. sicula (Reptilia: Lacertidae) from the historical centre of
Rome: What about competition and niche segregation in an urban habitat?
Bolletino Di Zoologia, 60(3), 287e291. https://doi.org/10.1080/
11250009309355825.
Carere, C., & Gherardi, F. (2013). Animal personalities matter for biological invasions.
Trends in Ecology &Evolution, 28(1), 5e6. https://doi.org/10.1016/j.tree.2012.10.
006.
Carretero, M. A. (2006). Reproductive cycles in Mediterranean lacertids: Plasticity
and constraints. In C. Corti, P. Lo Cascio, & M. Biaggini (Eds.), Mainland and
insular lizards: a Mediterranean perspective (pp. 33e54). Florence, Italy: Firenze
University Press.
Carretero, M. A. (2015). Examining the ecophysiology of Mediterranean lizards
with evolutio nary, biogeographic and co nservation eyes. In 18th European
Congress of He rpetology, Wrocla w (Poland), 7-12 September 2015.Invitedple-
nary lecture.
Carretero, M. A., & Silva-Rocha, I. (2015). La lagartija italiana (Podarcis sicula)enla
península ib
erica e islas Baleares (The Italian lizard (Podarcis sicula) in the
Iberian Peninsula and Balearic islands). Boletín de la Asociaci
on Herpetol
ogica
Espa~
nola, 26(2), 71e75.
Chapple, D. G., Simmonds, S. M., & Wong, B. B. (2011). Know when to run, know
when to hide: Can behavioral differences explain the divergent invasion success
of two sympatric lizards? Ecology and Evolution, 1(3), 278e289. https://doi.org/
10.1002/ece3.22.
Chapple, D. G., Simmonds, S. M., & Wong, B. B. (2012). Can behavioral and per-
sonality traits influence the success of unintentional species introductions?
Trends in Ecology &Evolution, 27(1), 57e64. https://doi.org/10.1016/j.tree.2011.
09.010.
Cote, J., Fogarty, S., Weinersmith, K., Brodin, T., & Sih, A. (2010). Personality traits
and dispersal tendency in the invasive mosquitofish (Gambusia affinis). Pro-
ceedings of the Royal Society B: Biological Sciences, 277,1571e1579. https://doi.
org/10.1098/rspb.2009.2128.
Dame, E. A., & Petren, K. (2006). Behavioural mechanisms of invasion and
displacement in Pacific Island geckos (Hemidactylus). Animal Behaviour, 71,
116 5 e1173 . https://doi.org/10.1016/j.anbehav.2005.10.009.
Diedenhofen, B., & Musch, J. (2015). cocor: A comprehensive solution for the sta-
tistical comparison of correlations. PLoS One, 10(4), e0121945. https://doi.org/
10.1371/journal.pone.0121945.
Dingemanse, N. J., & Wolf, M. (2013). Between-individual differences in behavioural
plasticity within populations: Causes and consequences. Animal Behaviour,
85(5), 1031e1039. https://doi.org/10.1016/j.anbehav.2012.12.032.
Downes, S., & Bauwens, D. (2002). An experimental demonstration of direct
behavioural interference in two Mediterranean lacertid lizard species. Animal
Behaviour, 63(6), 1037e1046. https://doi.org/10.1006/anbe.2002.3022.
Friard, O., & Gamba, M. (2016). BORIS: A free, versatile open-source event-logging
software for video/audio coding and live observations. Methods in Ecology and
Evolution, 7(11), 1325e1330. https://doi.org/10.1111/2041-210X.12584.
Geniez, P., S
a-Sousa, P., Guillaume, C. P., Cluchier, A., & Crochet, P. A. (2014). Sys-
tematics of the Podarcis hispanicus complex (Sauria, Lacertidae) III: Valid no-
mina of the western and central Iberian forms. Zootaxa, 3794(1), 1. https://doi.
org/10.11646/zootaxa.3794.1.1.
Gonz
alez de la Vega, J. P., Gonz
alez-García, J. P., García-Pulido, T., & Gonz
alez-
García, G. (2001). Podarcis sicula (Lagartija italiana), primera cita para Portugal
(Podarcis sicula (Italian lizard), first record for Portugal). Boletín de la Asociaci
on
Herpetol
ogica Espa~
nola, 12(1), 9.
Griffin, A. S., Guez, D., Federspiel, I., Diquelou, M., & Lermite, F. (2016). Invading
new environments: A mechanistic framework l inking moto r diversity and
cognitiontoestablishmentsuccess.InJ.S.Weis,&D.Sol(Eds.),Biolog ical
Invasions and An imal Behaviour (pp. 26e46). Cam bridge, U.K.: Cambridge
University Press .
Gruber, J., Brown, G., Whiting, M. J., & Shine, R. (2017a). Geographic divergence in
dispersal-related behaviour in cane toads from range-front versus range-core
populations in Australia. Behavioral Ecology and Sociobiology, 71(2), 38.
https://doi.org/10.1007/s00265-017-2266-8.
Gruber, J., Brown, G., Whiting, M. J., & Shine, R. (2017b). Is the behavioural diver-
gence between range-core and range-edge populations of cane toads (Rhinella
marina) due to evolutionary change or developmental plasticity? Royal Society
Open Science, 4, 170789. https://doi.org/10.1098/rsos.170789.
Herrel, A., Huyghe, K., Vanhooydonck, B., Backeljau, T., Breugelmans, K., Grbac, I.,
et al. (2008). Rapid large-scale evolutionary divergence in morphology and
performance associated with exploitation of a different dietary resource. Pro-
ceedings of the National Academy of Sciences, 105(12), 4792e4795. https://doi.
org/10.1073/pnas.0711998105.
Holway, D. A., & Suarez, A. V. (1999). Animal behavior: An essential component of
invasion biology. Trends in Ecology &Evolution, 14(8), 328e330. https://doi.org/
10.1016/S0169-5347(99)01636-5.
Jolliffe, I. T. (2002). Principal component analysis and factor analysis. In I. T. Joliffe
(Ed.), Principal Component Analysis (pp. 150e166). New York, NY: Springer.
https://doi.org/10.1007/0-387-22440-8_7.
Kraus, F. (2009). Impacts of alien reptiles and amphibians. In F. Kraus (Ed.), Alien
reptiles and amphibians (pp. 57e93). Dordrecht, Netherlands: Springer. https://
doi.org/10.1007/978-1-4020-8946-6_3.
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202 201
Michelangeli, M., Smith, C. R., Wong, B. B., & Chapple, D. G. (2017). Aggression
mediates dispersal tendency in an invasive lizard. Animal Behaviour, 133,29e34.
https://doi.org/10.1016/j.anbehav.2017.08.027.
Monceau, K., Moreau, J., Poidatz, J., Bonnard, O., & Thi
ery, D. (2015). Behavioral
syndrome in a native and an invasive hymenoptera species. Insect Science, 22(4),
541e548. https://doi.org/10.1111/1744-7917.12140.
Nakagawa, S., & Schielzeth, H. (2010). Repeatability for Gaussian and non-Gaussian
data: A practical guide for biologists. Biological Reviews, 85(4), 935e956. https://
doi.org/10.1111/j.1469-185X.2010.00141.x.
Nevo, E., Gorman, G., Soul
e, M., Yang, S. Y., Clover, R., & Jovanovi
c, V. (1972).
Competitive exclusion between insular Lacerta species (Sauria, Lacertidae).
Oecologia, 10(2), 183e190. https://doi.org/10.1007/BF00347990.
Phillips, B. L., & Suarez, A. V. (2012). The role of behavioural variation in the invasion
of new areas. In U. Candolin, & B. B. M. Wong (Eds.), Behavioural responses to a
changing world: mechanisms and consequences (pp. 190e200). Oxford, U.K.:
Oxford University Press.
Pintor, L. M., & Sih, A. (2009). Differences in growth and foraging behavior of
native and introduced populations in an invasive crayfish. Biological Invasions,
11, 1895e1902. https://doi.org/10.1007/s10530-008-9367-2.
Pintor, L. M., Sih, A., & Bauer, M. L. (2008). Differences in aggression, activity and
boldness between native and introduced populations of an invasive crayfish.
Oikos, 117, 1629e1636. https://doi.org/10.1111/j.1600-0706.2008.16578.x.
Pintor, L. M., Sih, A., & Kerby, J. L. (2009). Behavioral correlations provide a mech-
anism for explaining high invader densities and increased impacts on native
prey. Ecology, 90(3), 581e587. https://doi.org/10.1890/08-0552.1.
R Core Team. (2017). A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing.
R
eale, D., Reader, S. M., Sol, D., McDougall, P. T., & Dingemanse, N. J. (2007).
Integrating animal temperament within ecology and evolution. Biological
Reviews, 82(2), 291e318. https://doi.org/10.1111/j.1469-185X.2007.00010.x.
Rehage, J. S., Barnett, B. K., & Sih, A. (2005). Foraging behaviour and invasiveness: Do
invasive Gambusia exhibit higher feeding rates and broader diets than their
noninvasive relatives? Ecology of Freshwater Fish, 14, 352e360. https://doi.org/
10.1111/j.1600-0633.2005.00109.x.
Rehage, J. S., & Sih, A. (2004). Dispersal behavior, boldness, and the link to inva-
siveness: A comparison of four Gambusia species. Biological Invasions, 6,
379e391. https://doi.org/10.1023/B:BINV.0000 034618.93140.a5.
Ribeiro, R., & S
a-Sousa, P. (2018). Where to live in Lisbon: Urban habitat used by the
introduced Italian wall lizard (Podarcis siculus). Basic and Applied Herpetology,
32,57e70. https://doi.org/10.11160/bah.101.
Riley, J. L., Noble, D. W. A., Byrne, R. W., & Whiting, M. J. (2017). Early social envi-
ronment influences the behaviour of a family-living lizard. Royal Society Open
Science, 4, 161082. https://doi.org/10.1098/rsos.161082.
Rivera, X., Arribas, O. J., Carranza, S., & Maluquer-Margalef, J. (2011). An introduction
of Podarcis sicula in Catalonia (NE Iberian Peninsula) on imported olive trees.
Butlletí de la Societat Catalana d'Herpetologia, 19,79e85.
Schielzeth, H. (2010). Simple means to improve the interpretability of regression
coefficients. Methods in Ecology and Evolution, 1,103e113. https://doi.org/10.
1111/j.2041-210X.2010.00012.x.
Short, K. H., & Petren, K. (2008). Boldness underlies foraging success of invasive
Lepidodactylus lugubris geckos in the human landscape. Animal Behaviour, 76(2),
429e437. https://doi.org/10.1016/j.anbehav.2008.04.008.
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., Cote, J., Evans, M., Fogarty, S., & Pruitt, J. (2012). Ecological implications of
behavioural syndromes. Ecology Letters, 15(3), 278e289. https://doi.org/10.1111/
j.1461-0248.2011.01731.x.
Silva-Rocha,I., Salvi, D., & Carretero, M. A. (2012). Genetic data reveal a multiple origin
for the populations of the Italian walllizard Podarcis sicula (Squamata: Lacertidae)
introduced in the Iberian Peninsula and Balearic Islands. Italian Journal of Zoology,
79(4), 502e510. https://doi.org/10.1080/11250003.2012.680983.
Silva-Rocha, I., Salvi, D., Harris, D. J., Freitas, S., Davis, C., Foster, J., et al. (2014).
Molecular assessment of Podarcis sicula populations in Britain, Greece and
Turkey reinforces a multiple-origin invasion pattern in this species. Acta Her-
petologica, 9(2), 253e258. https://doi.org/10.13128/Acta_Herpetol-14968.
Sol, D., Timmermans, S., & Lefebvre, L. (2002). Behavioural flexibility and invasion
success in birds. Animal Behaviour, 63(3), 495e502. https://doi.org/10.1006/
anbe.2001.1953.
Valde
on, A., Perera, A., Costa, S., Sampaio, F., & Carretero, M. A. (2010). Evidencia de
una introducci
on de Podarcis sicula desde Italia a Espa~
na asociada a una
importaci
on de olivos (Olea europaea) (Evidence of an introduction of Podarcis
sicula from Italy to Spain associated with an import of olive trees (Olea europaea)).
Boletín de la Asociaci
on Herpetol
ogica Espa~
nola, 21,122e126.
Vervust, B., Grbac, I., & Van Damme, R. (2007). Differences in morphology, perfor-
mance and behaviour between recently diverged populations of Podarcis sicula
mirror differences in predation pressure. Oikos, 116(8), 1343e1352. https://doi.
org/10.1111/j.2007.0030-1299.15989.x.
Weis, J. S. (2010). The role of behavior in the success of invasive crustaceans. Marine
and Freshwater Behaviour and Physiology,43(2), 83e98. https://doi.org/10.
1080/10236244.2010.480838.
White, J. R., Meekan, M. G., McCormick, M. I., & Ferrari, M. C. O. (2013).
A comparison of measures of boldness and their relationships to survival in
young fish. PLoS One, 8(7), e68900. https://doi.org/10.1371/journal.pone.
0068900.
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.
Zuur, A. F., Ieno, E. N., & Elphick, C. S. (2010). A protocol for data exploration to avoid
common statistical problems. Methods in Ecology and Evolution,1,3e14. https://
doi.org/10.1111/j.2041-210X.2009.00001.x.
I. Damas-Moreira et al. / Animal Behaviour 151 (2019) 195e202202
