Content uploaded by Caterina Funghi
Author content
All content in this area was uploaded by Caterina Funghi
Content may be subject to copyright.
© The Author 2013. Published by Oxford University Press on behalf of the
International Society for Behavioral Ecology. All rights reserved. For permissions,
please e-mail: journals.permissions@oup.com
Behavioral
Ecology
The ofcial journal of the
ISBE
International Society for Behavioral Ecology
Original Article
Personality traits are related to ecology across
a biological invasion
Carlos F.Carvalho,a Ana V.Leitão,a,b CaterinaFunghi,a,b Helena R.Batalha,a SandraReis,a Paulo
GamaMota,a,b Ricardo J.Lopes,a and Gonçalo C.Cardosoa
aCIBIO—Centro de Investigação em Biodiversidade e Recursos Genéticos, Universidade do Porto,
Campus Agrário de Vairão, 4485-661 Vairão, Portugal and bDepartamento de Ciências da Vida,
Universidade de Coimbra, 3001-401 Coimbra, Portugal
Behavioral differences among individuals are common and are organized into personalities in a wide variety of species. Hypotheses for the
coexistence of behavioral differences fall into 3 categories: variation in selection, frequency-dependent selection, and behavioral plasticity.
We tested predictions of those hypotheses regarding geographic covariation of behavior with ecology, using a recent (≈40years) biological
invasion of common waxbills (Estrilda astrild). Behavior in tests for exploration and social interaction covaried among individuals, suggesting
a behavioral syndrome, although we could only demonstrate within-individual repeatability in the test for social interaction. These 2 behav-
iors changed geographically with the ecology of sites (degree of climate variation) in an apparently adaptive way, rather than with the direc-
tion of invasion. We found behavioral plasticity but showed that short-term plastic effects do not explain geographic divergence. Differential
dispersal does not explain geographic divergence either, which is orthogonal to the direction of invasion. Results are best interpreted either
as evolved divergences, although a candidate-gene approach could not identify genetic correlates of behavior, or as long-term behavioral
plasticity (e.g., effects of rearing environment). In this recent invasion, geographic differences in ecology and behavior equate to repeated
and fast changes over time. Thus, fluctuations in ecological conditions, which are common in nature, may have a widespread role maintain-
ing behavioral and personality differences via selection and/or long-term behavioral plasticity.
Key words: behavioral syndrome, biological invasion, climate, DRD4, ecology, personality, range expansion. [Behav Ecol]
INTRODUCTION
Behavioral dierences among individuals exist in humans and in
a wide range of species (Boissy 1995; Gosling and John 1999; Sih
etal. 2004). Often these take the form of behavioral syndromes or
personalities, whereby dierences among individuals span dierent
behavioral contexts (Gosling and John 1999; Sih et al. 2004) and
may be heritable (Dingemanse etal. 2002; van Oers et al. 2005).
Explaining this phenomenon is challenging and attracts much theo-
retical work because fixed behavioral dierences superficially seem
a worse adaptive solution than behavioral plasticity (Dall et al.
2004; Wolf et al. 2011). Three broad and nonmutually exclusive
classes of hypotheses have emerged to explain individual dierences
in behavior: variation in selective pressures, frequency-dependent
selection, and dierences in individual state plus behavioral plas-
ticity. There is theoretical and empirical support for each of these
classes of hypothesis (reviewed in Réale et al. 2007; Dingemanse
and Wolf 2010). It has seldom been attempted to evaluate their
relative importance in geographic divergence of behavior, which in
turn can inform the related question of what maintains behavioral
diversity within populations. Here, we attempt this by testing pre-
dictions from each of the hypotheses regarding geographic associa-
tions between behavior and ecology.
Behavioral dierences may not be fixed adaptations, but rather
plastic responses reflecting dierences in state between individuals
(Luttbeg and Sih 2010; Stamps and Groothuis 2010; Wolf and
Weissing 2010). This predicts that behavior diers between individuals
in dierent states (e.g., age) and that behavior adjusts to ecological
circumstances, both short term (e.g., seasonal eects) and long term
(e.g., rearing environment). Another hypothesis is that behavioral
dierences are due to dierences in selection pressures and then
coexist in populations due to fluctuating selection or migration (Réale
etal. 2007). This predicts associations between behavior and putative
selective pressures, such as ecological or demographic dierences
among sites, and that those are not explained by behavioral plasticity.
Finally, behavioral dierences may coexist through frequency-
dependent selection (Wolf et al. 2011; Wolf and McNamara 2012).
This hypothesis refers primarily to the coexistence of individual
dierences within populations. It could also contribute to geographic
dierences in behavior if the fitness of some behavioral types depends
on population density (e.g., Cote etal. 2008), in which case it predicts
associations of behavior with demography acrosssites.
Address correspondence to G.C. Cardoso. E-mail: gcardoso@cibio.up.pt.
Received 6 February 2013; revised 4 March 2013; accepted 18 March
2013.
Behavioral Ecology
doi:10.1093/beheco/art034
Behavioral Ecology Advance Access published May 6, 2013
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Behavioral Ecology
We tested these predictions on behaviors that have been related
to animal personality, using a recent biological invasion as a natural
experiment where divergence over space can be used as a proxy
for divergence over time. The common waxbill (Estrilda astrild) is
a small finch native to sub-Saharan Africa, which is traded as a
pet bird and has invaded several regions of the world (Stiels et al.
2011). It is highly gregarious, gathers in flocks year-round, inhab-
its open habitats in proximity of water, and feeds on herbaceous
seeds (Clement et al. 1993). In Europe, common waxbills use an
ecological niche quite distinct from native passerines’ (Batalha etal.
2013), and their largest invasion started in the 1960s at locations
near the Portuguese coastline, initially progressing slowly, but then
expanding to most of Portugal and also spilling to some parts of
southwest and northwest Spain (Reino and Silva 1998; Martí and
Moral 2003; Reino 2005). The expansion in Portugal was well
documented (Reino 2005; Sullivan etal. 2012), and it now encom-
passes a range of abiotic ecological conditions (altitude, climate,
and climate seasonality; Figure 1B–F), as well as sites colonized
for longer and with higher population densities and sites colonized
more recently and with lower population densities (Figure1A). We
tested if these dierences among sites predict behavior and assessed
if behavior changes plastically with season or if it diers among
individuals of dierent sex and age classes. We also assessed within-
individual repeatability in behavior and if dierent behaviors
covary among individuals, suggesting a behavioral syndrome, and
sequenced a candidate gene implicated in avian personality (Fidler
etal. 2007; Korsten etal. 2010; Kluen etal. 2012) to probe for a
possible genetic underpinning of behavioral dierences.
METHODS
Field behavioraldata
We conducted fieldwork in 2 breeding seasons, from June to
September 2010 and March to July 2011, in 23 approximately
similarly spaced sites throughout the entire range of the current
distribution of waxbills in Portugal (Figure 1A). Sites were visited
in semirandom order to not correlate date with ecological dier-
ences. We captured on average 28 waxbills per site (±14, standard
deviation [SD]; Supplementary Table S1), using mist-nets baited
with recordings of waxbill vocalizations or a live waxbill in a cage.
This sampling design, with nonselective capture methods (Biro and
Dingemanse 2009) and semirandom order of visits to sites, allows
testing for seasonal and other plastic eects on behavior.
Birds were aged as juveniles or adults based on bill color (adults
have red bill and juveniles have black bills that change to red as they
mature; Clement et al. 1993) and on bill commissures (fleshier in
juveniles), and sexed molecularly (see below). Captured birds were
kept in individual opaque paper bags until the behavioral tests. We
performed 4 tests in the field, targeting dierent behavioral contexts:
1)handling test to assess boldness, 2)tonic immobility test to assess fear,
3)mirror test to assess social interaction, and 4)open-field test to assess
exploration. The open-field test was conducted only for sites visited
in 2011. Tests were applied to all waxbills, except for occasional
logistic impediments, and conducted between 0800 and 1400 h.
1) In the handling test, always the first conducted, we noted
whether birds pecked or vocalized during 2 periods of stan-
dardized handling by the same researcher: a shorter period
(≈2 min) of banding and photography and a longer period
(≈10 min) encompassing blood sampling, morphometrics,
and spectrophotometry. Photography, morphometry, and
spectrophotometry data are not used in this article and are
not described here. Birds were given a score of 1 for either
pecking or vocalizing in either of the 2 periods, obtaining
an index that varied from 0 to 4.We chose these behaviors
because they were unambiguous and reflect the overall activ-
ity of birds during handling. This type of assay is used to
assess boldness (Garamszegi et al. 2009) because, from a
bird’s perspective, handling by a human for the first time is
similar to an interaction with a predator.
2) In the tonic immobility test, always the last conducted, one of
us removed the bird from the paper bag, gently placed it on
a flat surface (the same spot in the roof of a car) lying dor-
sally sideways to the observer, and retreated approximately 2
m while timing for how long the bird would remain in tonic
immobility before overturning and flying away, up to a maxi-
mum of 120 s. Tests where the bird flew o instantly (<0.5
s) were considered null, as these birds may not have entered
the state of tonic immobility. Tonic immobility is a state of
apparent paralysis triggered by overturning the body plan,
widely used as an index of fear in studies involving dierent
species, especially in animal welfare studies of birds (Gallup
1979; Jones 1986; Forkman etal. 2007; Hazard etal. 2008).
3) In the mirror test, each bird was placed in a small birdcage
(24.5 × 17 × 15 cm) placed at 43 cm from the ground with
3 equally distanced perches. To minimize external influ-
ences, the cage was in the middle of a square (90 × 90 cm)
of 90-cm-high cloth walls (Supplementary Figure S1A). One
end of the cage was covered with cardboard and the other
end had a similar but removable cardboard covering a mir-
ror. The behavior of the bird was filmed with a fixed digital
video camera for 10 min: in the first 5 min with the mirror
covered and in the following 5 min with the mirror exposed.
The mirror was exposed pulling the cardboard with a string
from a distance. From the video recordings, we quantified
7 behaviors that comprehensively characterize the birds’
activity in the cage, then tested which behaviors diered
between the periods with the mirror covered and uncovered,
and selected for further analyses the behaviors that diered
in the direction predicted by increased social response. The
7 candidate behaviors are described below. All were quanti-
fied separately for the 5-min period with the mirror covered
and with the mirror uncovered.
a) Position relative to mirror. It was categorized by dividing
the cage in 5 areas from mirror to opposite end (1: closest
to mirror, 2: close to first perch, 3: close to second perch,
4: close to third perch, and 5: closest to side opposite of
mirror; Supplementary Figure S1B), and calculating
the weighted average time (sum of area codes times the
duration of permanence there, divided by total duration).
This gives an index of proximity to the mirror that can
vary from 1 (always near the mirror) to 5 (always near the
opposite side).
b) Time facing the mirror. It was calculated as the length of
time the head was oriented toward the mirror (i.e., within
a 90° angle centered on the direction of the mirror).
c) Duration of grooming. It was calculated as the length
of time the bird was involved in cleaning the bill or the
feathers.
d) Duration of resting. It was calculated as the length of
time the bird was in typical resting position with the
plumage bulked.
Page 2 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Carvalho etal. • Personality traits are related to ecology
Figure1
Distribution range of waxbills in western Iberia and bioclimatic variation. (A) Distribution of waxbills in the western Iberian Peninsula (in gray), and location of
study sites indicating waxbill densities and time since colonization. Maps showing (B) altitude, (C) mean temperature, (D) precipitation, (E) variation in precipitation
(coecient of variation of monthly values), and (F) variation in temperature (SD of monthly means) across the distribution range. See text for data sources.
Page 3 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Behavioral Ecology
e) Number of vocalizations. Waxbills very commonly vocal-
ize when at close range (Clement etal. 1993). In waxbills,
and in the Estrildidae at large, song is generally used in
the context of mate attraction or stimulation (Hall 1962),
and birds in these tests did not sing.
f) Changing locations in the cage. It was quantified by
ranking transitions between adjacent positions in each
of three 3D axes (vertical axis: floor, perch, and hang-
ing on top of cage; depth axis: hanging on near wall, not
on wall, and on distal wall; and horizontal axis: same 5
areas described above for “position relative to mirror”;
Supplementary Figure S1B). Each transversal to an adja-
cent area was counted as 1, on each of the 3 axes, and
counts were then added for all axes.
g) Fast movements. In addition to perching or sitting in dif-
ferent parts of the cage (above variable), waxbills would
also fly or hop continuously, without stopping, for peri-
ods of time. We summed the length of time the bird was
doing those continuous, fast movements.
Position relative to mirror diered between the two 5-min
periods (Wilcoxon paired-samples signed-rank test, Z = 3.02,
N=507, P=0.002), but in the opposite direction of what was
predicted for prosocial behavior: birds were on average further
away, rather than closer, to the mirror when it was exposed.
Therefore, we did not use this behavior in subsequent analy-
ses. Three other behaviors augmented in the period with the
mirror exposed: facing the mirror (Z = 11.88, P < 0.001),
number of vocalizations (Z=9.74, P<0.001), and fast move-
ments (Z=3.96, P<0.001; for the other behaviors, all |Z| ≤
0.91, all P ≥ 0.36). We used these 3 behaviors with the mir-
ror exposed in a principal component analysis (PCA), which
returned a single principal component (PC) with eigenvalue
larger than 1 (1.32, equivalent to 44% of variance). This PC
had a high negative loading of facing the mirror (−0.701) and
high positive loadings of fast movements (0.807) and also of
vocalizations (0.423) and therefore discriminates between
birds responding more actively to the mirror image (high
PC scores: moving and vocalizing more) and those respond-
ing with increased attentiveness (low PC scores: facing the
mirrormore).
4) The open-field test adapted a well-established laboratory pro-
tocol (e.g., Verbeek etal. 1994; Dingemanse etal. 2002; Drent
et al. 2003) for use in the field (for other adaptations of this
protocol to the field, see van Dongen etal. 2010). The basic
protocol consists in releasing a bird into a novel environment
with artificial “trees” and measuring the extent of exploration.
We used a tent with internal dimensions of 240 × 210 × 150 cm
and rounded ceiling, which provides an area proportionally
adequate for the size of waxbills (compare with dimensions
on standard laboratory tests of great tits, Parus major, which
are about twice as large as waxbills, e.g., 400 × 240 × 230 cm;
Dingemanse et al. 2002). The tent walls were translucent
allowing light to enter but not allowing to see the outside,
and the colors were neutral (black floor and clear gray walls;
Supplementary Figure S2). The floor was divided into quad-
rants, in the center of each we placed an artificial “tree” (a
vertical wooden rod 60 cm high, with 3 horizontal perches),
plus a fifth tree in the center of the tent (Supplementary Figure
S2). We let the bird acclimate for 1 min under a black opaque
bowl in between 2 quadrants at the entrance of the tent. The
test began by lifting the bowl to the ceiling with a string, and
behavior was filmed for 10 min with a small digital camera on
the border of the tent floor. In no case was a bird seen to be
interested in the camera. From the video recordings, we quan-
tified 6 behavioral variables that comprehensively characterize
the birds’ most noticeable activities inside the tent:
a) Extent of exploration. It was calculated as the number of
quadrants the bird landed at plus the number of trees that
were visited (counting each quadrant or tree only once).
b) Changing locations. It was calculated by counting every
transition between quadrants, trees, or hanging on the
wall/ceiling of the tent. Unlike the previous variable, this
is influenced by repeated visits to the same quadrant or
tree and reflects overall activity.
c) Fast movements. In addition to perching or hopping to
dierent places (above variables), waxbills would some-
times fly continuously, often circularly around the tent.
We summed the length of time the bird was doing those
continuous, fast movements.
d) Total time the bird spent on the floor and e) total time
the bird spent perched in the trees. These 2 variables
allow discriminating between birds that remain mostly on
the ground, those that remain mostly on the perches, and
those that remain mostly on the walls (possibly trying to
escape the room; low scores on both variables).
f) Number of vocalizations during the test. As in the previ-
ous test, waxbills only called and did not sing.
We did not use measures on the latency of exploration (e.g., time
before visiting a certain number of quadrants or trees) because
most waxbills in these tests would visit few quadrants and trees
and the above variables provided sucient discrimination
among individuals. We ran a PCA on the 6 variables, which
returned 2 PCs with eigenvalues larger than 1 (Supplementary
Table S2). PC1 explained 41% of the variance and had high
positive loadings of the 2 variables more directly related to
exploration (extent of exploration and changing locations;
both loadings > 0.82) and positive loadings of most others
(Supplementary Table S2). The only variable with a weak
loading on this PC was the number of vocalizations, which is
not exploratory behavior, and the only variable with a negative
loading on this PC was time spent on the floor, which indicates
passive, nonexploratory behavior (Supplementary Table S2).
Therefore, PC1 reflects exploration behavior in a comprehensive
way, and we used the scores on this PC for subsequent analyses.
PC2 discriminated mostly birds that remained on the floor or
on the trees (Supplementary Table S2), independently of overall
exploration, and was not used in subsequent analyses.
Laboratory behavioraldata
We could not test within-individual repeatability of behavior in the
field, as birds were released after testing. To assess within-individual
repeatability, we conducted a complementary study in the labora-
tory during the breeding seasons of 2011 and 2012. A total of 78
adult waxbills were captured with mist-nets in a site near Coimbra
(5th site in Supplementary Table S1) during March 2011 (18 males
and 17 females) and February 2012 (22 males and 21 females) and
taken to the aviary of the Laboratory of Ethology at the University
of Coimbra. The aviary had controlled natural ventilation, and
natural illumination supplemented by artificial lights on a 12:12 h
light–dark cycle. Same-sex groups of 6 or 7 individuals were kept
in large bird cages (118 × 50 × 50 cm), with ad libitum access to
Page 4 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Carvalho etal. • Personality traits are related to ecology
water and commercial mixture of seeds (Tropical Finches Prestige,
Versele-Laga). We conducted 2 rounds of behavioral tests separated
by weeks to months (see below), during the mornings. The following
autumn, when wild flocks are large and easier for these birds to rein-
tegrate them, waxbills were transferred to a room (2.72 × 1.55 × 2.27
m) with water, food, and an abundance of perches for flight training
during 7–15days and then released at the capture location.
1) We did not conduct the handling test in the laboratory,
because this requires that birds are naive to being handled by
humans. Thus, we do not have within-individual repeatability
estimates for this test.
2) For the tonic immobility test, each bird was taken from the
cage and, after approximately 1 min, placed on a flat surface
(a 15.5- × 15.5-cm wooden surface at a height of 87 cm) in the
center of an empty room (1.36 × 1.06 × 2.27 m), lying dorsally
sideways to the observer, who remained at about 40 cm dis-
tance and timed how long the bird remained in tonic immo-
bility before overturning and flying, up to 60 s. Tests were
conducted from 26 March to 4 April (1st round) and from
25 to 27 May 2011 (2nd round), and from 16 to 24 February
(1st round) and 27 March 2012 (2nd round). Instances when
the bird flew o very rapidly were much less frequent than in
the field, perhaps because the test was conducted in a closed
room rather than in the open, and thus we used all data to
calculate within-individual repeatability.
3) The mirror test was conducted using a birdcage identical to
the one in the field tests. The cage was placed in the center of
a room (2.72 × 1.55 × 2.27 m) at 50 cm from the ground and,
as in the field, behavior was filmed for 5 min with the mirror
covered and for another 5 min with mirror exposed. Tests were
conducted from 27 April to 8 May (1st round) and from 9 to
11 September 2011 (2nd round), and from 16 to 20 March (1st
round) and from 18 to 20 April 2012 (2nd round). Behaviors
were quantified as described for the field tests. For comparabil-
ity with the field data, we ran a PCA on the same behaviors
included in the PCA for field data: facing the mirror, num-
ber of vocalizations, and fast movements. The PCA returned
a single PC with eigenvalue larger than 1 (1.23, equivalent to
41% of variance). Similar to the PC obtained with field data,
this PC discriminates birds that face the mirror more (trait
loading: 0.764) versus those that vocalize more (−0.654), but,
unlike in the field, fast movements (0.472) loaded in the same
direction as facing the mirror.
4) The open-field test was similar to the one in the field, but used
a room (2.72 × 1.55 × 2.27 m) rather than a tent. The floor of
the room was divided into quadrants with an artificial “tree” (a
vertical wooden rod 100 cm high, with 3 horizontal perches) in
the center of each, plus a fifth tree in the center of the room.
Tests were conducted from 14 to 18 April (1st round) and from
20 to 26 October 2011 (2nd round), and from 29 February to
13 March (1st round) and from 29 March to 3 April 2012 (2nd
round). Behaviors were quantified as described for the field
tests. As in the field, we let the bird acclimate for 1 min under
a black opaque bowl and began the test by lifting the bowl
to the ceiling with a string from outside the room. Tests were
filmed, and behaviors quantified similarly to the field tests, with
2 dierences: 1—fast movements were not quantified because,
unlike in the field tests, circular flights around the test room
were rare using these captive birds; 2—number of vocaliza-
tions was not used because in the field tests it was unrelated
to the exploration behaviors. As in the field, waxbills explored
little in this assay, and therefore we increased its length from
10 min in 2011 to 30 min in 2012. Because of this, we con-
verted all variables to values per minute before PCA. PC1 from
this PCA explained 44% of the variance and had high positive
loadings of the variables more directly related to exploration
(extent of exploration, 0.635; changing locations, 0.585; and
time in trees, 0.811) and a negative loading of time spent on
the floor (−0.603). This is very similar to the PC1 from the field
tests (Supplementary Table S2), and therefore we estimated
within-individual repeatability for the scores on this PC.
Demographic and climaticdata
We obtained data on demography (time since colonization,
and abundance of waxbills) and on climate and geography
(altitude, mean temperature and precipitation, and variation
in temperature and precipitation) for each field site. Time since
colonization was obtained from Reino (2005), where dates of
first observations for each 20 × 20 km2 are reported in intervals
of 4 or 5 years. We calculated time since colonization as the
dierence between 2011 and the midpoint of the time intervals
in Reino (2005). Waxbill abundance was taken from a recent
breeding bird atlas (Equipa Atlas 2008), which calculated
abundances by smoothing field observation data (percentage of
occurrence across sampling areas) with generalized linear models
parameterized to optimize the compromise between geographic
detail and noise in point counts (Wood 2005; Equipa Atlas 2008).
Abundances are reported as percentage of detections across 2 × 2
km areas within 10 × 10 km squares, in interval classes of 10%
or 15%; we used the midpoint of the probability interval for
each site. Altitude and climatic information were taken from the
WorldClim database (www.worldclim.org; Hijmans etal. 2005) at
5 arc-min resolution (approximately 9.3 × 9.3 km, which is smaller
than the distances between our study sites; minimum distance
between sites was 28.39 km). For mean temperature, we averaged
across the 12monthly means; for precipitation, we took the yearly
cumulative precipitation; for variation in temperature, we took the
SD across the monthly means; and for variation in precipitation,
we took the coecient of variation across the monthly values.
All variables are illustrated in Figure1 and values per site are in
Supplementary Table S1.
These 7 variables were related to some extent, and to avoid mul-
ticollinearity, we reduced them with a PCA that returned 2 PCs
(hereafter termed geography PCs) with eigenvalues larger than 1
and explain 83% of total variation. Geography PC1 has high positive
loadings of time since colonization, abundance, mean temperature,
and variation in precipitation and high negative loadings of alti-
tude, precipitation, and variation in temperature (absolute values of
all trait loadings > 0.5; Supplementary Table S3). High PC1 scores,
thus, indicate low-altitude sites, hot and dry, with high precipita-
tion variation but low temperature variation, which harbor older
and denser populations of waxbills. Geography PC2 is little related
to demography or altitude (absolute values of loadings < 0.5) and
instead was characterized by high positive loadings of both vari-
ables of climate variation, and also a high positive loading of mean
temperature and a high negative loading of precipitation (in conti-
nental Portugal, sites with stronger climate variation also have hot-
ter and drier yearly means) (Supplementary Table S3). PC2, thus,
reflects mostly climate variation and is not related to demography
(abundance or time since colonization).
Page 5 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Behavioral Ecology
Geneticdata
We collected a small (<70 μL) blood sample from each bird by
puncturing the brachial vein. Blood was stored in 96% etha-
nol until DNA extraction, which used a silica-based method, the
DNeasy Blood & Tissue Kit (Qiagen, The Netherlands), following
standard procedures. Eluted DNA was normalized for polymerase
chain reaction (PCR) (10–20 ng/µL) and stored in −20°C.
Birds were sexed following Lee et al. (2010), using primers to
amplify the CHD1 gene (CHD1-F and CHD1-R). The PCR pro-
tocol consisted of 0.5µL of HOT FIREPol Blend Master Mix 10.0
(Solis BioDyne, Estonia), 0.18 µL of each primer, and 0.4µL of
DNA template for a total volume of 4 µL. The conditions were as
follows: 15 min at 95°C, 35 cycles of 30 s at 95°C, 30 s at 55°C,
and 80 s at 72°C, followed by 10 min at 72°C. PCR products were
separated on a 2% agarose gel by electrophoresis. Individuals show-
ing double (ZW) and single (ZZ) bands were identified as females
and males, respectively.
We selected a subset of adult birds (sex ratio = 0.5) with high
and low scores in the mirror test (≈4 adults with highest and 4
with lowest scores per site) for sequencing exon 3 of the dopa-
mine receptor D4 (DRD4) gene. This is the largest exon of
DRD4, where, in other passerine species (P. major and Cyanistes
caeruleus), single nucleotide polymorphisms (SNPs) were found
to relate to personality traits (Fidler et al. 2007; Korsten et al.
2010; Kluen et al. 2012). For the samples from 2010, we ampli-
fied a 111-bp fragment from exon 3 using primers DRD4I3_F2
(5′-ACTCAAGCGCTGGGAAGAAGCC-3′) and DRD4I3_R2
(5′-AGGGAGGCCAGCACGGGTGTA-3′), which includes the
position homologous to the P. major SNP830. PCR protocol was
as follows: 2.5 µL of Phusion High-Fidelity PCR Master Mix
(Finnzymes, Finland), 0.25 µL of each primer, and 0.5µL of a
sample’s DNA; conditions: 1 min at 98°C, 38 cycles (20 s at 98°C,
10 s at 71.4°C, 6 s at 72°C, and 5 min at 72°C). For the samples
from 2011, we used recently published primers (DRD4I2F and
DRD4I3R; Mueller et al. 2011) that amplify a larger fragment
(640 bp) including the full exon 3 and used a PCR protocol that
consisted in 0.8µL of HOT FIREPol Blend Master Mix 10.0 (Solis
BioDyne, Estonia), 0.35µL of each primer, and 0.8µL of DNA
template. The conditions were as follows: 15 min at 95°C, 38 cycles
(30 s at 95°C, 40 s at 65°C); 75 s at 72°C; and 10 min at 72°C.
All sequencing reactions were run in a 3730xl DNA Analyzer
(Applied Biosystems) in Macrogen Europe (The Netherlands) and
results were edited and assembled in Geneious 5.5 (Biomatters,
New Zealand).
The final exon 3 sequence (GenBank accession number
KC894739) showed no polymorphism at the position homologous
to the P. major SNP830 (all birds homozygous to C) or the C. cae-
ruleus SNP905 (all birds homozygous to A), but we found 2 other
SNPs: a C/T SNP at exon 3 base 307 and, for the 2011 samples
with the complete exon 3 sequenced, a G/A SNP at exon 3 base
100. Like the SNP830 of P.major and SNP905 of C.caeruleus, these
are synonymous substitutions.
Analyses
Using the behavioral tests in the laboratory, we estimated the
within-individual repeatability from the variance components of
analyses of variance (Anovas) with each behavior as the dependent
variable (Lessels and Boag 1987). In addition to the identity of indi-
vidual birds as the random factor, the models included round num-
ber (1 or 2)as a fixed factor to account for possible order eects. In
all cases (N=78), individuals were tested twice, and significance of
the repeatability estimate is given by the eect of the random factor
in the Anovas.
Using the field data, we tested for eects of age, sex, date, and
possible confounding factors with a general linear model (GLM) for
each behavioral test. The model had sex and age (juvenile vs. adult)
as factors, their interaction, and the following covariates: Julian date
and its square (to detect linear and curvilinear seasonal eects),
time of day (test time minus sunrise time), and time since capture
(test time minus capture time). In all subsequent analyses, we used
the residuals of these GLMs, rather than the raw behaviors, in
order to control for sex and the above plastic eects on behavior.
We averaged residual behaviors per site and tested for spatial
autocorrelation in the behavioral data calculating Moran’s I at dif-
ferent distance classes (Valcu and Kempenaers 2010). We grouped
sites by distance intervals multiples of 60 km for the handling, tonic
immobility, and mirror tests (17 out of 253 pairs of sites <60 km)
and by intervals multiples of 100 km for the open-field test (because
it was conducted at fewer sites and only 3 of those were closer than
60 km; 10 out of 66 pairs of sites <100 km). Spatial autocorrela-
tion is detected by a significant positive Moran’s I at the smallest
distance class and then decreasing Moran’s I for larger distance
classes (Gittleman and Kot 1990). Moran’s I was not significant for
the smallest distance class for any behavior (all Z ≤ 1.46, critical
Z=1.96), indicating that, for our site distances, spatial autocorrela-
tion in behavior was negligible. Therefore, we used each site as an
independent data point in the tests of geographic variation.
We tested for an association with climate and demography with
a multiple linear regression of average behavior per site, on the 2
demography and climate PCs. Because waxbill densities and,
consequently, sample sizes of captured birds vary nonrandomly with
the ecology of sites (see above), we analyzed average behavior per site
(i.e., use sites as data points, rather than individuals grouped by site)
so that all sites are equally weighted in the analysis. In one site, with
extremely low waxbill abundance, we could only capture 3 adult birds
(Supplementary Table S1), and we report results without this site.
Analyses including this site (not shown) were qualitatively identical.
We examined plots of normality of residuals and of residuals versus
expected values, which showed that the data met the assumptions
for multiple regression analysis: residuals closely approximated a
normal distribution were homoscedastic (i.e., there were no trends for
variance to change with predicted values), and the data did not have
nonlinear structure (i.e., residuals did not show curvilinear variation
with the predicted values from the linear model).
We tested for covariation of the dierent behavioral tests across
individuals with simple bivariate correlations. We also tested this
looking only at covariation among individuals within sites, having
controlled for behavioral dierences among sites using residuals
from an additional GLM with site as the single random factor.
For each behavioral test, we tested for an association with the
SNPs in the DRD4 exon 3.For the C/T SNP, we used GLMs with
genotype (C/C, C/T, or T/T) as factor. For the G/A SNP, we used
T-tests comparing genotype G/G with G/A only because the third
genotype was very rare in our samples (2 out of 107 birds). All sta-
tistical tests were run in SPSS 17.0, except for calculating Moran’s I
where we used the formulae in Gittleman and Kot (1990)).
RESULTS
In the laboratory behavioral tests, within-individual repeatability
was significant for the tonic immobility test (r=0.22, F77,77=1.575,
Page 6 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Carvalho etal. • Personality traits are related to ecology
P = 0.024) and for the mirror test (r = 0.19, F77,77 = 1.469,
P= 0.047), but not for the open-field test (r< 0.01, F77,77= 0.948,
P=0.592).
For the behavioral tests in the field, GLMs including sex, age
(juvenile vs. adult), date (linear and curvilinear), and possible con-
founding factors (time since capture to testing and time of testing)
showed that none of the behaviors diered between sexes (Table1)
and that age classes diered in the handling test, with juveniles
behaving more boldly (vocalizing and pecking more; F1,500=28.7,
P < 0.001; Table 1). There were seasonal eects in the mirror
test (more attentive response with increasing date, F1,480 = 4.5,
P = 0.03; Table 1) and in the tonic immobility test (negative lin-
ear eect, F1,183 = 9.0, P = 0.003, and positive curvilinear eect,
F1,183= 7.6, P= 0.006; Table1). Of the possible confounding fac-
tors, only testing time had an eect on tonic immobility (length of
tonic immobility decreased with time of day, F1,183=6.8, P=0.01;
Table 1), which indicates that behavioral tests were overall robust
to variation in experimental conditions. To control for eects of
behavioral plasticity, all subsequent analyses used behavior residuals
from the aboveGLMs.
Geography PC1, reflecting demography and mean climate, was
not related to mean behavior in any of the behavioral tests (all
|βst| < 0.35, all P > 0.12; Table 2). Geography PC2, reflecting
climate seasonality, was strongly related to behavior in the mirror
test (βst = −0.64, P =0.002, N= 22 sites; Figure 2A) and in the
open-field test (βst = −0.67, P =0.046, N = 10 sites; Figure 2B),
but did not predict behavior in the handling and tonic immobility
tests (Table 2). This suggests that waxbills in sites with more vari-
able climate react more attentively to social stimuli and explore less
novel environments. Post hoc correlations with the product of both
climate variation variables (variation in temperature and variation
in precipitation) resulted in correlation coecients nearly as high
as the previous (mirror test: rp = −0.55, P =0.01, N= 22 sites;
open-field test: rp=−0.59, P=0.07, N=10 sites), confirming that
climate variation was the main driver of these results.
The 2 behaviors showing the above geographic pattern were cor-
related among individuals (rp = 0.29, P < 0.001, N = 175), even
when looking only at covariation within sites (rp=0.27, P <0.001,
N= 175; Table3), with individuals that react attentively to social
stimuli also exploring less in a novel environment. There were no
significant correlations involving the other behavioral tests (all |rp|
< 0.13, all P > 0.09; Table 3). We found no associations between
behavior and DRD4 genotypes for either the C/T SNP (GLMs, all
F<2.9, all P > 0.06; Figure3A) or the G/A SNP (T-tests, all |t| <
1.5, all P > 0.14; Figure3B).
DISCUSSION
We conducted behavioral assays on waxbills throughout the
geographic range of a recent biological invasion that comprises
various demographic and ecological conditions. Behavior in
open-field and mirror tests was correlated among individuals,
even when looking only at within-site variation, consistent with
work on several species showing that exploration and social
behaviors are often part of a behavioral syndrome (Réale
et al. 2007). However, repeated tests in the laboratory showed
within-individual repeatability for only one of these behaviors.
Geographic dierences in these 2 behaviors were related to
climate variation, with waxbills exploring less and responding
more attentively to social stimuli in sites with variable climate,
and the reverse in sites with stable climate. This geographic
pattern appears adaptive (see below) and was independent of
age and seasonal eects, indicating that those simple forms of
behavioral plasticity do not explain geographic divergence.
Geographic divergence could be due to longer-term plastic
eects (e.g., eects of rearing environment) or evolved behavioral
dierences, although a candidate-gene approach could not
demonstrate genetic correlates of behavior.
The covariation of behaviors among waxbills is consistent with
individuals diering along an axis of proactive to reactive cop-
ing styles. This axis of coping styles reflects the degree to which
behavior is guided by stimuli from the environment and appears
as a fundamental aspect of individual dierences in behavior
and stress physiology for a wide variety of animals (reviewed in
Table1
Sex and plastic eects on behavior
Handling test Tonic immobility Mirror test Open-field test
Sex F1,500=0.719 (P=0.397) F1,183=2.290 (P=0.132) F1,480=0.077 (P=0.781) F1,173=0.003 (P=0.959)
Age 28.699 (<0.001) 0.207 (0.649) 0.643 (0.423) 0.563 (0.454)
Sex × age 0.204 (0.652) 0.154 (0.696) 0.843 (0.359) 0.015 (0.902)
Date 0.000 (0.984) 8.991 (0.003) 4.498 (0.034) 2.539 (0.113)
Date20.095 (0.758) 7.631 (0.006) 1.255 (0.263) 2.222 (0.138)
Time of test 0.180 (0.672) 6.806 (0.010) 0.090 (0.764) 1.077 (0.301)
Wait time to test 0.732 (0.393) 1.367 (0.244) 1.702 (0.193) 0.347 (0.556)
GLMs on each behavior, including sex and age (juvenile vs. adult) as factors, their interaction, Julian date and its square, time of day, and time since capture to
test as covariates. Significant eects (P<0.05) are in bold.
Table2
Relations of behavior with ecology across sites
Handling test Tonic immobility Mirror test Open-field test
PC1 (demography and mean climate) βst=0.343 (P=0.132) 0.348 (0.120) 0.122 (0.493) 0.098 (0.734)
PC2 (climate seasonality) 0.287 (0.203) 0.097 (0.657) −0.639 (0.002) −0.670 (0.046)
Degrees of freedom 2, 17 2, 19 2, 19 2, 7
Multiple regressions of average behavior per site, on the 2 main axes of variation (principal components, PCs) in demography and climate. Significant eects
(P<0.05) are in bold.
Page 7 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Behavioral Ecology
Koolhaas etal. 1999; Carere etal. 2010). Waxbills that explored
a novel environment more were more active and vocal facing a
social stimulus (mirror image), as expected of proactive individu-
als, whereas those that explored less were more attentive toward
their mirror image, as expected of reactive individuals. Repeated
assays in the laboratory confirmed that behavior is repeatable
within individuals in the mirror test, assessing social responses
(also in the tonic immobility test, assessing fear), but did not find
repeatable individual dierences in the open-field test, assessing
exploration behavior. A possible explanation is that the open-
field test is little eective evoking exploration for this species,
since waxbills almost never visited all quadrants or artificial trees
in the field or laboratory tests. This assay, with artificial trees in
an empty space, proved useful to investigate avian personalities
in a various species (e.g., Dingemanse et al. 2002; David et al.
2011), but there are cases in which researchers had to elaborate
on this design by extending it or providing additional structure
and stimuli (e.g., Marchetti and Zehtindjiev 2009; Mutzel et al.
2011; Liebl and Martin 2012), including another estrildid finch
(Krause and Naguib 2011). It is, thus, prudent to bear in mind
that designs more eective in evoking exploration in waxbills
might reveal consistent dierences among individuals. Despite
this, the commonness of proactive to reactive coping styles in
animals (Koolhaas etal. 1999; Carere etal. 2010), plus the cor-
relation between the waxbill open-field and mirror tests within
sites, suggests that exploration and social behaviors are part of a
behavioral syndrome. Boldness and fear also often covary along
the proactive to reactive axis (Sih etal. 2004; Carere etal. 2010),
but our assays for these (handling and tonic immobility tests) did
not reveal such covariation.
The correlated behaviors in the open-field and mirror tests dif-
fered geographically according to the degree of climate variation
of sites in an apparently adaptive way. Stronger seasonality implies
instability in the location of food and other ecological resources
throughout the year and greater importance of adjusting rou-
tines under those changing environments. Proactive coping styles
are associated with less flexible behavior (Carere et al. 2010), and
research on avian personalities, in particular, has shown a com-
promise between fast exploration and learning eciency, such that
fast explorers follow behavioral routines, whereas slow explorers
are more flexible and sensitive to environmental stimuli (Verbeek
etal. 1994; Exnerová etal. 2010; Guillette etal. 2011). Thus, pro-
active animals act primarily on the basis of previous experience,
which is fast but may be inaccurate, whereas reactive animals tend
to rely more on detailed information available in the environment.
The proactive coping style was suggested to be adaptive in stable
Figure2
Relations of tests for social and exploration behavior with climate variation. Relations of mean scores in the (A) mirror and (B) open-field tests across sites
with climate variation (second principal component of a principal component analyses on climate and demography).
Table3
Bivariate correlations across the dierent behavioral tests
Handling test Tonic immobility Mirror test Open-field test
Handling test rp (P, N)−0.007 (0.934, 144) 0.060 (0.214, 433) −0.011 (0.882, 171)
Tonic immobility 0.023 (0.784, 144) −0.091 (0.238, 170) −0.085 (0.525, 58)
Mirror test 0.026 (0.590, 433) −0.128 (0.097, 170) 0.267 (<0.001, 175)
Open-field test −0.020 (0.793, 171) −0.052 (0.699, 58) 0.290 (<0.001, 175)
Left from diagonal, correlations using residuals from a GLM to control for sex and aspects of behavioral plasticity, and right from the diagonal also controlling
for dierences among sites using residuals from an additional GLM with site as the single factor. Significant eects (P<0.05) are in bold.
Page 8 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Carvalho etal. • Personality traits are related to ecology
environmental conditions, whereas the reactive coping style may do
better under variable and unpredictable conditions (Koolhaas etal.
1999). This is consistent with the slower exploration of waxbills in
more seasonal sites being adaptive for their unstable environment.
Similarly, greater attentiveness to social stimuli, as opposed to more
active social responses, suggests that waxbills in more seasonal sites
rely more on cues from conspecifics. Individuals vary in their abili-
ties or experience, such that being attentive to conspecifics should
increase the eciency of problem solving in new or unstable envi-
ronments (Liker and Bókony 2009; Morand-Ferron and Quinn
2011). Supporting this adaptive interpretation, it was recently
shown that zebra finches (another gregarious estrildid) that explore
less use more social information (Rosa etal. 2012) and that tits with
more social connections in mixed flocks find novel food patches
faster in the wild (Aplin etal. 2012).
In other avian species, variation in exploratory behavior was
shown to be partly heritable (Dingemanse et al. 2002; van Oers
et al. 2005). Therefore, the geographic divergence in waxbill
behavior could be due to plasticity or due to selection. We found
evidence for plasticity in most behaviors, in the form of seasonal,
time of day, or age eects, consistent with the hypothesis that
behavior adjusts plastically to dierences in state. The dierences
in state we studied are only a subset of those that may contribute
to behavioral dierences among individuals (others include condi-
tion, social hierarchy, or reproductive value; Luttbeg and Sih 2010;
Wolf and Weissing 2010), but illustrate that behavioral plasticity
explains at least part of the behavioral dierences observed among
individuals. We can exclude some simple forms of behavioral
plasticity as explanations for the geographic dierences because
we controlled for age and several short-term plastic eects. Also,
behavior was related to climate variation rather than mean cli-
mate of each site, indicating that behavioral dierences among
sites were not adjustments to the immediate climatic conditions
experienced at the time of tests. Longer-term behavioral plasticity,
Figure3
Comparison of behavior across genotypes of the DRD4 exon 3.Sample sizes are over the horizontal axis.
Page 9 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Behavioral Ecology
such as due to past experience or rearing environments, remains a
candidate explanation for the geographic divergence (Stamps and
Groothuis 2010).
Dierential dispersal, when individuals of certain behavioral
types are more likely to disperse, can also create geographic
dierences in behavior (reviewed in Cote, Clobert, et al. 2010).
During range expansions it could create dierences between
older and newly colonized sites, or between more and less
densely populated sites if individuals disperse depending on their
degree of sociality (e.g., Cote and Clobert 2007; Jokela et al.
2008; Cote, Fogarty, etal. 2010). Dierences between more and
less populated sites could also originate by frequency-dependent
selection, if the fitness of some behavioral types depends on
population density (e.g., Cote etal. 2008). But in this biological
invasion, of waxbills we did not find evidence that behavior
changed with the demography of sites, that is, from older sites
with higher population densities toward sites colonized more
recently and with smaller populations. Waxbills gather and move
in flocks (Clement et al. 1993), and perhaps the potential for
dispersal-mediated geographic dierences in behavior is buered
because flocks carry individuals with high dispersal propensities
and also other individuals that follow the former (e.g., Cote etal.
2011).
Yet another mechanism that could cause behavioral dierences
during range expansions is adaptation to novel environments at
the range edged, as suggested, for example, for neophobia and
exploration behavior in house sparrows, Passer domesticus (Martin
and Fitzgerald 2005; Liebl and Martin 2012). Waxbill behavior
did dier according to the degree of climate variation across sites,
consistent with the idea that ecological novelty (here, in the form
of annual fluctuations) is a driver of behavioral change. But this
ecological variation is orthogonal to the direction of the waxbill
invasion and thus did not result in divergence between older and
newly colonized sites. Thus, ecological dierences among sites were
more behaviorally relevant than the direction of invasion, suggest-
ing it is important to consider ecology when interpreting behav-
ioral dierences during range expansions, as the direction of some
range expansions may concur with behaviorally relevant ecological
gradients.
In an attempt to distinguish between the 2 explanations for the
geographic divergence in waxbill behavior, evolved behavioral dif-
ferences or long-term behavioral plasticity, we assessed a candi-
date gene implicated in exploration and other personality traits.
Personality traits have been linked to variation in the DRD4 gene
in birds (Fidler etal. 2007; Korsten etal. 2010; Kluen etal. 2012)
and various other vertebrates (e.g., fish: Boehmler et al. 2007;
humans: Munafo etal. 2008). In birds, laboratory assays of explo-
ration in great tits (Fidler et al. 2007; Korsten etal. 2010) and
short field assays of escape behavior in blue tits (Kluen etal. 2012)
were linked to synonymous SNPs in the third intracellular loop
of DRD4, possibly due to regulatory eects (Fidler et al. 2007).
We found 2 SNPs in the same intracellular loop also causing syn-
onymous substitutions, but these were not related to variation in
any of the behaviors assayed. Although exon 3 of DRD4 is the
only successful candidate gene in avian studies of exploration and
other personality traits to date, it does not relate to exploration
behavior in all species (Atwell etal. 2012) and even within a single
species its role can vary across populations (Korsten et al. 2010).
Thus, lack of association of this candidate gene with behav-
ioral dierences does not imply lack of genetic underpinning
and instead shows the need for new genomic tools to confidently
address questions on the genetics of personality (van Oers and
Mueller 2010).
In conclusion, our results suggest adaptive geographic dier-
ences in dierent behaviors, likely as part of a behavioral syn-
drome. Ecological changes over time can aect the fitness of
behavioral phenotypes (e.g., Réale and Festa-Bianchet 2003;
Dingemanse et al. 2004; Quinn et al. 2009), and fixed or plas-
tic adaptive geographic dierences in behavior have also been
reported (e.g., Chiba etal. 2007; Urban 2007; Atwell et al. 2012;
Magnhagen etal. 2012). Our study bridges these 2 temporal and
geographic dimensions, as in this recent biological invasion (a slow
onset approximately 45years ago, and increased expansion rates
starting about 30 years ago; Reino and Silva 1998), geographic
dierences in ecology and behavior equate to fast and repeated
changes over time. Because fluctuations in ecological condi-
tions and selective optima are common in nature (Halley 1996;
Bell 2010), this suggests that varying conditions through time (or
through space plus migration) can cause selection or long-term
plastic eects that help maintain behavioral and personality dier-
ences among individuals.
SUPPLEMENTARY MATERIAL
Supplementary material can be found at http://www.beheco.
oxfordjournals.org/
FUNDING
This work was funded by grant PTDC/BIA-BEC/098414/2008 and
fellowships SFRH/BPD/40786/2007 and SFRH/BPD/46873/2008
from the Fundação para a Ciência e a Tecnologia.
Work with waxbills was done under permits 04/2010/CAPT and 45/2011/
CAPT from the Instituto da Conservação da Natureza e da Biodiversidade.
Handling editor: Regina Macedo
REFERENCES
Aplin LM, Farine DR, Morand-Ferron J, Sheldon BC. 2012. Social net-
works predict patch discovery in a wild population of songbirds. Proc R
Soc B. 279:4199–4205.
Atwell JW, Cardoso GC, Whittaker DJ, Campbell-Nelson S, Robertson
KW, Ketterson ED. 2012. Boldness behavior and stress physiology in a
novel urban environment suggest rapid correlated evolutionary adapta-
tion. Behav Ecol. 23:960–969.
Batalha HR, Ramos JA, Cardoso GC. 2013. A successful avian invasion
occupies a marginal ecological niche. Acta Oecol. 49:92–98.
Bell G. 2010. Fluctuating selection: the perpetual renewal of adaptation in
variable environments. Phil Trans R Soc B. 365:87–97.
Biro PA, Dingemanse NJ. 2009. Sampling bias resulting from animal per-
sonality. Trends Ecol Evol. 42:66–67.
Boehmler W, Carr T, Thisse C, Thisse B, Canfield VA, Levenson R. 2007.
D4 dopamine receptor genes of zebrafish and eects of the antipsy-
chotic clozapine on larval swimming behaviour. Genes Brain Behav.
6:155–166.
Boissy A. 1995. Fear and fearfulness in animals. Q Rev Biol. 70:165–191.
Carere C, Caramaschi D, Fawcett TW. 2010. Covariation between person-
alities and individual dierences in coping with stress: converging evi-
dence and hypotheses. Curr Zool. 56:728–740.
Chiba S, Arnott SA, Conover DO. 2007. Coevolution of foraging behavior
with intrinsic growth rate: risk taking in naturally and artificially selected
growth genotypes of Menidia menidia. Oecologia. 154:237–246.
Clement P, Harris A, Davies J. 1993. Finches and sparrows. Princeton:
Princeton University Press.
Cote J, Clobert J. 2007. Social personalities influence natal dispersal in a
lizard. Proc R Soc B. 274:383–390.
Page 10 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Carvalho etal. • Personality traits are related to ecology
Cote J, Clobert J, Brodin T, Fogarty S, Sih A. 2010. Personality-dependent
dispersal: characterization, ontogeny and consequences for spatially
structured populations. Phil Trans R Soc B. 365:4065–4076.
Cote J, Dreiss A, Clobert J. 2008. Social personality trait and fitness. Proc R
Soc B. 275:2851–2858.
Cote J, Fogarty S, Brodin T, Weinersmith K, Sih A. 2011. Personality traits
and dispersal tendency in the invasive mosquitofish: group composition
matters. Proc R Soc B. 278:1670–1678.
Cote J, Fogarty S, Weinersmith K, Brodin T, Sih A. 2010. Personality traits
and dispersal tendency in the invasive mosquitofish (Gambusia anis). Proc
R Soc B. 277:1571–1579.
Dall SRX, Houston AI, McNamara JM. 2004. The behavioural ecology of
personality: consistent individual dierences from an adaptive perspec-
tive. Ecol Lett. 7:734–739.
David M, Auclair Y, Cézilly F. 2011. Personality predicts social dominance
in female zebra finches, Taeniopygia guttata, in a feeding context. Anim
Behav. 81:219–224.
Dingemanse NJ, Both C, Drent PJ, van Oers K, van Noordwijk AJ. 2002.
Repeatability and heritability of exploratory behaviour in great tits from
the wild. Anim Behav. 64:929–937.
Dingemanse NJ, Both C, Drent PJ, Tinbergen JM. 2004. Fitness conse-
quences of avian personalities in a fluctuating environment. Proc R Soc
B. 271:847–852.
Dingemanse NJ, Wolf M. 2010. Recent models for adaptive personality dif-
ferences: a review. Phil Trans R Soc B. 365:3947–3958.
van Dongen WFD, Maldonado K, Sabat P, Vásquez RA. 2010. Geographic
variation in the repeatability of a personality trait. Behav Ecol.
21:1243–1250.
Drent PJ, van Oers K, van Noordwijk AJ. 2003. Realized heritability of
personalities in the great tit (Parus major). Proc R Soc B. 270:45–51.
Equipa Atlas. 2008. Atlas das Aves Nidificantes em Portugal (1999–2005).
Lisbon: Assírio & Alvim.
Exnerová A, Svádová KH, Fucíková E, Drent P, Stys P. 2010. Personality
matters: individual variation in reactions of naive bird predators to apo-
sematic prey. Proc R Soc B. 277:723–728.
Fidler AE, van Oers K, Drent PJ, Kuhn S, Mueller JC, Kempenaers B.
2007. Drd4 gene polymorphisms are associated with personality variation
in a passerine bird. Proc R Soc B. 274:1685–1691.
Forkman B, Boissy A, Meunier-Salaün M-C, Canali E, Jones RB. 2007. A
critical review of fear tests used on cattle, pigs, sheep, poultry and horses.
Physiol Behav. 92:340–374.
Gallup GG. 1979. Tonic immobility as a measure of fear in domestic fowl.
Anim Behav. 27:316–317.
Garamszegi LZ, Eens M, Török J. 2009. Behavioural syndromes and trap-
pability in free-living collared flycatchers, Ficedula albicollis. Anim Behav.
77:803–812.
Gittleman JL, Kot M. 1990. Adaptation: statistics and a null model for esti-
mating phylogenetic eects. Syst Zool. 39:227–241.
Gosling SD, John OP. 1999. Personality dimensions in nonhuman animals:
a cross-species review. Curr Dir Psychol Sci. 8:69–75.
Guillette LM, Reddon AR, Hoeschele M, Sturdy CB. 2011. Sometimes
slower is better: slow-exploring birds are more sensitive to changes in a
vocal discrimination task. Proc R Soc B. 278:767–773.
Hall MF. 1962. Evolutionary aspects of estrildid song. Symp Zool Soc
Lond. 8:37–55.
Halley JM. 1996. Ecology, evolution and 1/f noise. Trends Ecol Evol.
11:33–37.
Hazard D, Couty M, Richard S, Guémené D. 2008. Intensity and dura-
tion of corticosterone response to stressful situations in Japanese quail
divergently selected for tonic immobility. Gen Comp Endocrinol.
155:288–297.
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. 2005. Very high
resolution interpolated climate surfaces for global land areas. Int J
Climatol. 25:1965–1978.
Jokela M, Elovainio M, Kivimäki M, Keltikangas-Järvinen L. 2008.
Temperament and migration patterns in Finland. Psychol Sci.
19:831–837.
Jones RB. 1986. The tonic immobility reaction of the domestic fowl: a
review. World Poultry Sci J. 42:82–97.
Kluen E, Kuhn S, Kempenaers B, Brommer JE. 2012. A simple cage test
captures intrinsic dierences in aspects of personality across individuals
in a passerine bird. Anim Behav. 84:279–287.
Koolhaas JM, Korte SM, de Boer SF, van der Vegt BJ, van Reenen CG,
Hopster H, de Jong IC, Ruis MAW, Blokhuis HJ. 1999. Coping styles
in animals: current status in behavior and stress physiology. Neurosci
Biobehav Rev. 23:925–935.
Korsten P, Mueller JC, Hermannstädter C, Bouwman KM, Dingemanse
NJ, Drent PJ, Liedvogel M, Matthysen E, van Oers K, van Overveld T,
etal. 2010. Association between DRD4 gene polymorphism and person-
ality variation in great tits: a test across four wild populations. Mol Ecol.
19:832–843.
Krause ET, Naguib M. 2011. Compensatory growth aects exploratory
behaviour in zebra finches, Taeniopygia guttata. Anim Behav. 81:1295–1300.
Lee JC, Tsai L, Hwa P, Chan C, Huang A, Chin S, Wang L, Lin J, Linacre
A, Hsieh H. 2010. A novel strategy for avian species and gender identifi-
cation using the CHD gene. Mol Cell Probes. 24:27–31.
Lessels CM, Boag PT. 1987. Unrepeatable repeatabilities: a common mis-
take. Auk. 104:116–121.
Liebl AL, Martin LB. 2012. Exploratory behaviour and stressor hyper-
responsiveness facilitate range expansion of an introduced songbird. Proc
R Soc B. 279:4375–4381.
Liker A, Bókony V. 2009. Larger groups are more successful in innova-
tive problem solving in house sparrows. Proc Natl Acad Sci USA.
106:7893–7898.
Luttbeg B, Sih A. 2010. Risk, resources and state-dependent adaptive
behavioural syndromes. Phil Trans R Soc B. 365:3977–3990.
Magnhagen C, Hellström G, Borcherding J, Heynen M. 2012. Boldness in
two perch populations—long-term dierences and the eect of predation
pressure. J Anim Ecol. 81:1311–1318.
Marchetti C, Zehtindjiev P. 2009. Migratory orientation of sedge warblers
(Acrocephalus schoenobaenus) in relation to eating and exploratory behaviour.
Behav Process. 82:293–300.
Martí R, Moral JC. 2003. Atlas de las Aves Reproductoras de España.
Madrid: Dirección General de Conservación de la Naturaleza—Sociedad
Española de Ornitología.
Martin LB, Fitzgerald L. 2005. A taste for novelty in invading house spar-
rows, Passer domesticus. Behav Ecol. 16:702–707.
Morand-Ferron J, Quinn J. 2011. Larger groups of passerines are
more ecient problem solvers in the wild. Proc Natl Acad Sci USA.
108:15898–15903.
Mueller JC, Pulido F, Kempenaers B. 2011. Identification of a gene associ-
ated with avian migratory behaviour. Proc R Soc B. 278:2848–2856.
Munafo MR, Yalcin B, Willis-Owen SA, Flint J. 2008. Association of the
dopamine D4 receptor (DRD4) gene and approach-related personality
traits: meta-analysis and new data. Biol Psychiatry. 63:197–206.
Mutzel A, Kempenaers B, Laucht S, Dingemanse N, Dale J. 2011.
Circulating testosterone levels do not aect exploration in house spar-
rows: observational and experimental tests. Anim Behav. 81:731–739.
van Oers K, de Jong G, van Noordwijk AJ, Kempenaers B, Drent PJ. 2005.
Contribution of genetics to the study of animal personalities: a review of
case studies. Behaviour. 142:1185–1206.
van Oers K, Mueller JC. 2010. Evolutionary genomics of animal personal-
ity. Phil Trans R Soc B. 365:3991–4000.
Quinn JL, Patrick SC, Bouwhuis S, Wilkin TA, Sheldon BC. 2009.
Heterogeneous selection on a heritable temperament trait in a variable
environment. J Anim Ecol. 78:1203–1215.
Réale D, Festa-Bianchet M. 2003. Predator-induced natural selection on
temperament in bighorn ewes. Anim Behav. 65:463–470.
Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ. 2007.
Integrating animal temperament within ecology and evolution. Biol Rev.
82:291–318.
Reino LM. 2005. Variation partitioning for range expansion of an intro-
duced species: the common waxbill Estrilda astrild in Portugal. J Ornithol.
146:377–382.
Reino LM, Silva T. 1998. The distribution and expansion of the common
waxbill (Estrilda estrild) in the Iberian Peninsula. Biol Conserv Fauna.
102:163–167.
Rosa P, Nguyen V, Dubois F. 2012. Individual dierences in sampling
behaviour predict social information use in zebra finches. Behav Ecol
Sociobiol. 66:1259–1265.
Sih A, Bell AM, Johnson JC, Ziemba RE. 2004. Behavioral syndromes: an
integrative overview. Q Rev Biol. 79:241–277.
Stamps J, Groothuis TGG. 2010. Developmental perspectives on personal-
ity: implications for ecological and evolutionary studies of individual dif-
ferences. Phil Trans R Soc B. 365:4029–4041.
Stiels D, Schidelko K, Engler JO, van den Elzen R, Rödder D. 2011.
Predicting the potential distribution of the invasive common waxbill
Estrilda astrild (Passeriformes: Estrildidae). J Ornithol. 152:769–780.
Page 11 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from
Behavioral Ecology
Sullivan MJP, Davies RG, Reino L, Franco AMA. 2012. Using dispersal
information to model the species–environment relationship of spreading
non-native species. Methods Ecol Evol. 3:870–879.
Urban MC. 2007. Risky prey behavior evolves in risky habitats. Proc Natl
Acad Sci USA. 104:14377–14382.
Valcu M, Kempenaers B. 2010. Spatial autocorrelation: an overlooked con-
cept in behavioral ecology. Behav Ecol. 21:902–905.
Verbeek MEM, Drent PJ, Wiepkema PR. 1994. Consistent individual dierences
in early exploratory behaviour of male great tits. Anim Behav. 48:1113–1121.
Wolf M, van Doorn GS, Weissing FJ. 2011. On the coevolution of
social responsiveness and behavioural consistency. Proc R Soc B.
278:440–448.
Wolf M, McNamara JM. 2012. On the evolution of personalities via fre-
quency-dependent selection. Am Nat. 179:679–692.
Wolf M, Weissing FJ. 2010. An explanatory framework for adaptive person-
ality dierences. Phil Trans R Soc B. 365:3959–3968.
Wood SN. 2005. Generalized additive models: an introduction with R.
London: Chapman & Hall/CRC.
Page 12 of 12
at Universidade do Porto on May 14, 2013http://beheco.oxfordjournals.org/Downloaded from