Behavioral Ecology
doi:10.1093/beheco/arp199
Advance Access publication 29 January 2010
Interspecific variation in fear responses predicts
urbanization in birds
Anders Pape Møller
Laboratoire d’Ecologie, Systématique et Evolution, CNRS UMR 8079, Université Paris-Sud,
Bâtiment 362, F-91405 Orsay Cedex, France and Center for Advanced Study, Drammensveien 78,
NO-0271 Oslo, Norway
Urbanization and domestication share features in terms of characters that are favored by selection. These include loss of fear of
humans, reduced corticosterone levels, prolonged breeding seasons, and several others. Here, I test the hypothesis that urbanization results from differential colonization of urban areas by species with heterogeneous levels of fear in the ancestral rural
populations, followed by a reduction in variance in fear responses with a subsequent increase in diversity of fear responses as
urban populations become adapted to the urban environment. Using information on variance in flight initiation distances
(FIDs) when approached by a human, I show that rural populations of birds characterized by short mean flight distances and
large variances in flight distances differentially colonized urban areas. As a consequence of this urban invasion, urban populations lost variation in FID. The variance in FID was initially larger in rural than in urban populations but eventually became larger
in urban populations with time since urbanization. This secondary increase in variance in FID of urban populations was
associated with an increase in population density of urban populations, suggesting that as birds became adapted to urban
areas, they secondarily gained variance in behavioral flexibility. Key words: flight initiation distance, invasion, personality,
urbanization. [Behav Ecol 21:365–371 (2010)]
ehavioral syndromes are correlated categories of behavior
that often co-occur nonrandomly in the same individuals
(e.g., Dall et al. 2004; Sih et al. 2004a, 2004b; Sih and Bell
2008). Many different kinds of behavior show covariation
among contexts such as sexual and social behavior and antipredator behavior (e.g., Sih et al. 2004a, 2004b; Réale et al.
2007). Several studies have shown that risk taking is a component of behavioral syndromes because it is correlated with
exploratory and aggressive behavior (e.g., van Oers et al.
2004; Ward et al. 2004; Hollander et al. 2008; Garamszegi
et al. 2009; Wilson and Godin 2009).
Personalities are inherent consistent characteristics of individuals that determine how such individuals cope with various
environmental challenges by behavioral means. Personality
traits arrange individuals along continuous axes such as bold–shy, aggressive–docile, explorative–avoiding, and extrovert–
cautious behavior depending on the context (Wilson et al.
1994; Boissy 1995; Gosling and John 1999; Sih et al. 2004a;
Réale et al. 2007). Covariation among behavioral traits as
mediated by personality is likely to reflect trade-offs in behavior and life history (Stamps 2007; Wolf et al. 2007). Therefore,
no specific behavioral phenotype is generally favored by
selection, as changing and unpredictable environmental conditions will favor different phenotypes. As a result, heterogeneous populations with individuals of both bold and shy
personalities exist, allowing for coping with variable selective
forces (Dingemanse et al. 2007). Accordingly, the distribution
of phenotypes within a population or a species should reflect
the degree to which they experience these unpredictable
conditions.
B
Address correspondence to A.P. Møller. E-mail: anders.moller
@u-psud.fr.
Received 23 July 2009; revised 7 December 2009; accepted 12
December 2009.
The Author 2010. Published by Oxford University Press on behalf of
the International Society for Behavioral Ecology. All rights reserved.
For permissions, please e-mail: [email protected]
Urbanization is the process associated with invasion of urban
areas. Urbanized species of birds initially had large breeding
ranges and a high propensity for dispersal (Møller 2009). Species that subsequently became urbanized had rural populations with short flight distances, characteristic of the
situation before urbanization, compared with rural populations of species that did not succeed in becoming urbanized
(Møller 2009). Because urban areas generally have fewer predators than rural areas (Klausnitzer 1989; Marzluff et al. 2001;
Shochat et al. 2006; Møller 2009), we should expect such urban populations to prosper because they will not pay any costs
of predation from having short flight distances, nor costs for
antipredator behavior, but will gain benefits in terms of lack of
disturbance effects caused by humans and domestic animals.
There is extensive literature on the scenario for invasion of
urban areas by birds and other animals (review in Klausnitzer
1989). Numerous species have settled close to human beings,
often resulting in human habitations being the sole remaining breeding habitat today. These changes in association with
humans are often accompanied by changes in behavior, physiology, and life history (Klausnitzer 1989; Møller 2008b, 2009).
Urban population densities are often higher than densities in
nearby rural habitats, and this difference has traditionally
been attributed to higher resource abundance in urban habitats (Klausnitzer 1989). However, urban populations may also
be denser than rural populations if they are composed of
a greater diversity of personalities that can exploit a greater
diversity of habitats. An experimental release of both urban
and rural blackbirds Turdus merula from Poland in Kiev,
Ukraine, which did not have a native urban population,
showed that only urban birds survived (Graczyk 1974). Hence,
urban birds may be better adapted to urban environments,
even when these are novel.
If invasion of urban areas was characterized by behavioral
flexibility that allows for individuals to cope with novel environments, we should expect urbanized populations to differ in
terms of innovative behavior. Sasvári (1985) showed that
Behavioral Ecology
366
urbanized species learn faster than nonurbanized species. In
contrast, Kark et al. (2007) did not show a difference in
innovative behavior with degree of urbanization. However,
Møller (2009) showed for a sample of 39 urbanized bird species and closely related nonurbanized species that those that
subsequently became urbanized had a higher rate of feeding
innovations that might allow such species to exploit novel
environments and do so in novel ways. Likewise, Liker and
Bokony (2009) found a nonsignificant trend for urban populations of house sparrows Passer domesticus to be more variable in innovative behavior than rural populations, suggesting
a role for such flexible behavior in successful invasion of urban areas. However, house sparrows have been associated with
humans for a long time (Summers-Smith 1963), suggesting
that this species may not constitute a typical example of adaptation to urban areas. Thus, we need information about
general adaptations to urban areas in many different taxa to
allow generalizations to be made.
The objective of this study was to investigate the dynamics of
the frequency distributions of flight distances as a way of investigating adaptation to urban environments. Garamszegi et al.
(2008, 2009) have shown that individual male collared flycatchers Ficedula albicollis with long flight initiation distances
(FIDs) also have specific behavior in terms of exploration,
aggression, and courtship display. Bird species are highly heterogeneous in terms of mean fear responses, and variances in
fear responses may differ independent of these mean values.
Bird species in rural habitats, characteristic of the situation
before urbanization, had short flight distances, if they successfully managed to invade urban areas (Møller 2009). Individual
barn swallows Hirundo rustica and collared flycatchers are
highly consistent in their FID, even following a challenge to
their immune system (Møller AP, Garamszegi LZ, unpublished data). Therefore, heterogeneity in FID will reflect heterogeneity in behavior among individuals rather than
phenotypic plasticity of individuals. Thus, not only speciesspecific mean values of FID may be subject to selection, but
also intraspecific variation can be expected to respond to ecological factors, as heterogeneity in behavior reflects the distribution of personalities among individuals. Here, I test
whether species with highly variable flight distances, suggesting heterogeneity in antipredator response and hence behavioral flexibility, were more likely to colonize urban habitats. If
a specific subset of individuals were able to colonize urban
habitats, we should expect such colonization to be followed
by a reduction in mean and variance in flight distance. This
loss of heterogeneity in flight distance could imply that such
urban populations were more homogeneous than the ancestral rural populations and that they were composed of individuals that only vary little in personality. Subsequently, when
urban areas were successfully exploited, disruptive selection
for exploitation of heterogeneous environments should again
result in an increase in the variance in flight distances. The
prediction is that means and variances in flight distance
should change over time, with a correlation between time
since urbanization and difference in mean and variance in
flight distance between ancestral rural and derived urban populations of the same species.
MATERIALS AND METHODS
Study sites
The analyses reported here are based on extensive data on FIDs
collected by myself in France and Denmark during February–
September 2006–2008. I defined urban areas as built-up areas
with continuous houses or multistorey buildings, with the only
interspersed areas being roads and city parks. Rural areas had
open farmland, forests, moors, lakes, and other habitats with
scattered houses and farms that were never continuous. This
simple operational definition is in accordance with those adopted in other studies (e.g., Klausnitzer 1989; Gliwicz et al. 1994;
Stephan 1999). I defined a species as being urbanized based
on information in Cramp and Perrins (1977–1994) and Glutz
von Blotzheim and Bauer (1985–1997).
Flight distances
FIDs were recorded using a modified technique developed by
Blumstein (2006). A full description of the procedures and 3
different crossvalidations of the data are reported in Møller
(2008a, 2008b, 2008c). In brief, when an individual bird had
been located with a pair of binoculars, I moved at a normal
walking speed toward the individual, while recording the
number of steps (which approximately equals the number
of meters, Møller et al. 2008). The distance from the observer
to the bird when it first took flight was recorded as the FID,
while the starting distance was the distance from where the
observer started walking up to the position of the bird and
the location of the bird. If the individual was positioned in the
vegetation, the height above ground was recorded to the nearest meter. While recording these FIDs, I also recorded date
and time of day, and the sex and the age of the individual if
external characteristics allowed sexing and aging with binoculars. FID was estimated as the Euclidian distance that equals
the square root of the sum of the squared horizontal distance
and the squared height above ground level (Blumstein 2006).
Previous studies have shown that starting distance is strongly
positively correlated with FID (e.g., Blumstein 2006), thereby
causing a problem of collinearity. I eliminated this problem of
collinearity by searching habitats for birds with a pair of binoculars when choosing an individual for estimating FID. In
this way, I assured that most individuals were approached
from a distance of at least 30 m, thereby keeping starting
distances constant across species. FID was weakly negatively
related to starting distance in a model that included species,
age, habitat, country, and body mass as factors (partial F ¼
37.97, df ¼ 1, 4188, P , 0.0001), explaining only 1% of the
variance. None of the results presented in this paper changed
statistically when including starting distance as an additional
variable, and I thus excluded this variable from all subsequent
analyses for simplicity.
Flight distance varied significantly among species, using oneway analysis of variance with log10-transformed flight distance
as the response variable and species as a factor (F ¼ 42.68,
df ¼ 154, 4045, r2 ¼ 0.62, P , 0.0001, repeatability (R) following Becker (1984) R ¼ 0.60, standard error [SE] ¼ 0.04). I
estimated variation in flight distance as the standard deviation, and this standard deviation was subsequently log10transformed to obtain a normal frequency distribution (see
Statistical Methods below). The log10-transformed standard
deviation in flight distance was repeatable between samples
of 37 species from Denmark and France (F ¼ 5.50, df ¼ 36, 37,
r2 ¼ 0.84, P , 0.0001, R ¼ 0.69, SE ¼ 0.12). I was able to
estimate means and variance in FID for a total of 133 species
that were subsequently used in the analysis of colonization of
urban areas. In a second series of analyses, I used paired
estimates of means and variance in FID for urban and rural
populations of the same species, and such estimates were available for 48 species. Sample sizes are reported in Appendix 1.
Population densities of breeding birds
I censused breeding birds around Orsay (4842#N, 211#E),
France, using standardized point counts (e.g., Blondel et al.
1970) in late April–early May 2009 during early morning from
Møller • Møller urbanization and flight distance flexibility
sunrise until 08:00. A total of 100 points in urban areas and
100 points in nearby rural areas were visited during a period
of 5 min during which I recorded all individuals heard or
seen. Habitats were classified as described above. Points were
separated by distances of 100 m to avoid counting the same
individuals multiple times. If individuals could be heard singing at more than one point, they were only counted the first
time. Population densities were simply the mean number of
individuals per point of the different species in urban and
rural areas, respectively.
Urbanization
I defined a bird species as being urbanized based on information in Cramp and Perrins (1977–1994) and Glutz von
Blotzheim and Bauer (1985–1997). I combined this information with my personal experience with all the species considered, including information from colleagues interested in
urbanization of birds. Species in rural areas have subsequently
colonized urban areas, initially often with just few individuals.
Hence, population density is initially high in rural areas and
low in urban areas, followed by an increase in urban population density as the urban population adapts to the novel urban environment. A species being classified as urbanized had
to fulfill the following criteria: 1) Breeding populations occur
inside towns and cities. 2) Population densities in towns and
cities are higher than in rural habitats. Based on these criteria,
I made a list of 63 urbanized species of breeding birds of the
521 species recorded in the Western Palearctic, fulfilling the 2
criteria listed above (Møller 2009). This data set was subsequently reduced to 133 species for which I had information
on mean and variance in FID.
Timing of urbanization
Timing of urbanization will obviously result from colonization
followed by establishment or extinction and recolonization.
Obviously, there is no information on such processes, nor is
there empirical information about the development of urban
population sizes since colonization. In the following, I assume
that colonization of urban environments can be approximated
from observations by keen ornithologists that habitually follow
changes in composition and distribution of birds closely. Any
heterogeneity in colonization processes or increase in population size will cause noise in the data and ultimately make it
more difficult to discern any clear patterns. I estimated the year
when different species became urbanized in Denmark and
France using 2 different approaches. First, I asked 2 keen
Danish amateur ornithologists (William C. Årestrup and
E. Flensted-Jensen) living in my study area in Denmark to state
when different species of birds were first recorded breeding in
urban areas. An approximate year of urbanization was
recorded, with a value of 1950 assigned to species that were
known to breed in urban habitats before the 2 observers started
watching birds. Independently, I also independently recorded
these years for all species before asking for estimates from the 2
independent observers. These 3 data sets were highly consistent in assignment of year (all 3 Pearson r . 0.96, N ¼ 44
species).
Second, I recorded the timing of invasion of urban environments in Copenhagen, Denmark (300 km SE of the study area)
and Paris, France, from old published records (Gram 1908;
Fløystrup 1920, 1925 for Copenhagen, and an extensive literature survey or ornithological journals that also included
Cramp and Perrins 1977–1994 and Glutz von Blotzheim and
Bauer 1985–1997). If the year of urbanization was before records reported in these sources, I assigned 1858 as the year of
urbanization (because Gram’s observations date back to 1858,
367
and many species were urbanized before the start of his
study). Although urbanization is likely to have occurred much
earlier for many species such as house sparrow P. domesticus
and rock pigeon Columba livia, these estimates are conservative. This data set provided independent estimates of the time
of urbanization, but these were still strongly positively correlated with my estimates from Northern Jutland (Pearson r ¼
0.72, N ¼ 41 species). Given that my own estimates of timing
of urbanization from my study area were strongly positively
correlated with these published sources, I used my own estimates combined with those from Copenhagen for species that
became urbanized before 1950 in the analyses. This was done
because my own data derived from the same study area where
the flight distance data were collected. See Møller (2008b) for
further information.
Body mass
I extracted information on mean body mass of adult birds from
Cramp and Perrins (1977–1994). The entire data set is reported in electronic appendix 1.
Statistical analyses
Mean and variance in FID and body mass were log10transformed to achieve normal distributions.
Sampling effort varied among species (mean [SE] ¼ 31
observations [4], median 14, range 2–349, N ¼ 133 species
for which I had information on mean and variance in FID),
with no significant relationship with mean flight distance (F ¼
0.63, df ¼ 1, 131, r2 ¼ 0.00, P ¼ 0.43) or variance in flight
distance (F ¼ 0.56, df ¼ 1, 131, r2 ¼ 0.00, P ¼ 0.45). Hence,
such heterogeneity in sampling effort cannot be addressed by
including sample size as an additional predictor variable but
requires that each observation contribute to test statistics relative to its importance (Garamszegi and Møller 2009). Most
statistical approaches assume that all data points provide
equally precise information about the deterministic part of
total process variation, that is, the standard deviation of the
error term is constant over all values of the predictor variables
(Sokal and Rohlf 1995). I weighted each observation by log10transformed sample size in order to use all data in an unbiased fashion, with the log-transformation ensuring that each
datum was given relatively greater weight when increases in
sample sizes were small reflecting that the degree of precision
due to sampling effort is more important for increases in
small than in large sample sizes (Draper and Smith 1981;
Neter et al. 1996).
Variance in flight distance was significantly positively related
to mean flight distance (Figure 1; F ¼ 172.27, df ¼ 1, 131, r2 ¼
0.57, P , 0.0001, slope [SE] ¼ 1.935 [0.147]). Because of
collinearity problems when including both mean and variance
in the same model, I estimated residual variance in flight
distance by using the phylogenetically adjusted relationship
from an analysis of contrasts (see below) (F ¼ 97.98, df ¼
1, 125, r2 ¼ 0.44, P , 0.0001, slope [SE] ¼ 1.801 [0.182]).
The estimates of residual variance were not significantly related to mean flight distances (F ¼ 0.83, df ¼ 1, 131, r2 ¼ 0.01,
P ¼ 0.36, slope [SE] ¼ 0.134 [0.147]). These residual variance
estimates are termed variance in flight distance in the following analyses. Because log10-transformed mean FID and residual variance in FID were not correlated, I could include both
as predictors in the same model. In the comparative analyses,
I used log10-transformed mean flight distance, residual variance in flight distance and log10-transformed body mass as
predictors.
Analyses of comparative data based on species may provide
misleading conclusions if sister taxa are phenotypically more
Behavioral Ecology
368
Table 1
Summary statistics for flight initiation distances for urban and rural
populations of 48 species of birds
Variable
Mean
Urban mean (m)
6.94
Urban standard deviation (m)
2.30
Urban population density
0.33
Rural mean (m)
12.77
Rural standard deviation (m)
6.54
Rural population density
0.24
Urban–rural mean
20.212***
Urban–rural variance
20.660***
Urban–rural population density
0.040
SE
Range
N
3.51
2.55, 20.50
0.27
0.45, 9.74
0.06
0.00, 1.53
1.70
4.87, 66.67
1.25
0.21, 54.37
0.05
0.00, 1.27
0.026 21.04, 0.19
0.080 23.16, 1.10
0.088 20.85, 1.31
48
48
48
48
48
48
48
48
48
The difference in log10-transformed mean and variance between
urban and rural populations was tested against the null expectation
of zero, with the significance level being indicated by asterisks.
***P , 0.0001.
Fjeldså (2006) to resolve relationships between some species.
The phylogeny is shown in Appendix 2.
RESULTS
Figure 1
Frequency distributions of (A) difference in mean and (B) difference
in variance in FID between rural and urban populations of the same
species of birds.
similar than randomly chosen species. Therefore, I based the
analyses on statistically independent, standardized linear contrasts (Felsenstein 1985), which controls for similarity in phenotype among species due to common phylogenetic descent.
The contrasts were calculated using the software of Purvis and
Rambaut (1995), implemented in the computer program
CAIC. All regressions were forced through the origin
(Felsenstein 1985), because the dependent variable is not assumed to have changed, when the predictor variable has not
evolved. Standardization of contrast values was checked by
examination of absolute values of standardized contrasts versus their standard deviations (Garland et al. 1992; Garland
and Ives 2000). Plotting the resulting contrasts against the
variances of the corresponding nodes revealed that these
transformations made the variables suitable for regression
analyses. In order to reduce the consequent problem of
heterogeneity of variance, 1) outliers (contrasts with Studentized residuals . 3) were excluded from subsequent analyses
(Jones and Purvis 1997; the presented analyses included these
outliers), and 2) analyses were repeated with the independent
variable expressed in ranks (Møller and Birkhead 1994).
However, these analyses all produced statistically similar conclusions. I present results based on both species-specific data
without phylogenetic adjustment and phylogenetically independent contrasts. In order to weight regressions by sample
size in the comparative analyses, I calculated weights for each
contrast by calculating the mean sample size for the taxa immediately subtended by that node (Møller and Nielsen 2007).
The comparative analyses relied on a composite phylogeny
created by using information in Hackett et al. (2008) and
Sibley and Ahlquist (1990), supplemented with Jønsson and
Summary statistics for means and variance of FIDs are reported
in Table 1. Mean flight distances were highly consistent between urban and rural populations (F ¼ 68.11, df ¼ 1, 46, r2 ¼
0.60, P , 0.0001). Likewise variances were highly consistent
between the 2 kinds of populations (F ¼ 38.27, df ¼ 1, 46, r2 ¼
0.45, P , 0.0001). This leaves 40% and 55% of residual variance, respectively, that are analyzed subsequently.
The difference in mean FID between urban and rural populations was significantly negative, implying that flight distances were generally shorter in urban than in rural populations
(Table 1; Figure 1A). However, there was considerable variation among species, with some having longer flight distances
in urban areas and some in rural areas (Figure 1A). Likewise,
the difference in variance in FID between urban and rural
populations was significantly negative, implying that flight
distances were more variable in rural than in urban populations (Table 1; Figure 1B). There was a positive correlation
between difference in variance between rural and urban
populations and difference in means between populations
(Figure 2).
Whether a species had become urbanized or not was predicted by mean and variance in FID in rural populations
(i.e., in the ancestral population before urbanization) and
body mass in a model of species-specific data (Table 2). A
comparative analysis based on contrasts only revealed
a strongly significant effect for variance in FID (Table 2).
Therefore, species that were successful in colonizing urban
areas had larger variances in FID in rural populations than
unsuccessful colonizers.
The difference in variance in FID between urban and rural
populations was significantly negatively related to time since
urbanization (Figure 3), in a model that explained 14% of
the variance. This means that when urban birds are more
variable in flight distance than rural birds, urban populations
have been established since a long time.
The difference in variance in FID between rural and urban
populations was significantly positively related to the difference in population density (Figure 4), in a model that explained 10% of the variance. This implies that urban
populations had higher population density than rural populations when the variance in flight distance was greater in
urban than in rural populations.
Møller • Møller urbanization and flight distance flexibility
Figure 2
Difference in variance in FID between rural and urban populations
of the same species in relation to difference in mean FID between
rural and urban populations of the same species. The line is the
linear regression line. The model for species weighted by log10transformed sample size had the statistics F ¼ 96.24, df ¼ 1, 46, r2 ¼
0.68, P , 0.0001, slope (SE) ¼ 0.25 (0.03), whereas the model for
contrasts had the statistics F ¼ 86.83, df ¼ 1, 46, r2 ¼ 0.65, P , 0.0001,
slope (SE) ¼ 3.35 (0.36).
DISCUSSION
The main findings of this study of variance in behavior and urbanization was that species that initially colonized urban areas
had more variable flight distance behavior than species that
failed in such colonization. Initially, urban populations lost behavioral variability during colonization, suggesting that only
a specific subset of behaviors were represented among successful colonizers. However, behavioral flexibility and population
density increased as time passed since urbanization, consistent
with the interpretation that successful urbanization was associated with regained behavioral flexibility.
Colonization of urban areas was followed by a reduction in
both mean and variance in flight distance compared with the
rural ancestors. Indeed, the difference in variance in flight distance between urban and rural populations was strongly positively correlated with the difference in mean flight distance
between urban and rural populations. Thus, individuals from
urban populations were less fearful of humans than their rural
ancestors, as shown already (Cooke 1980; Klausnitzer 1989;
Gliwicz et al. 1994; Stephan 1999; Møller 2008b). Such reduced fearfulness among urban birds may have as a consequence that they are less disturbed by humans and their
domesticated animals (mainly cats and dogs) (Møller et al.
2008). In addition, urban birds were more homogeneous in
flight distance than rural birds. This loss of heterogeneity in
flight distance could imply that such urban populations have
369
Figure 3
Difference in variance in FID between rural and urban populations
of the same species in relation to time since urbanization. The line is
the linear regression line. The model for species weighted by
log10-transformed sample size had the statistics F ¼ 6.12, df ¼ 1, 46,
r2 ¼ 0.12, P ¼ 0.017, slope (SE) ¼ 0.0060 (0.0024), whereas the
model for contrasts had the statistics F ¼ 4.44, df ¼ 1, 46, r2 ¼ 0.09,
P ¼ 0.041, slope (SE) ¼ 0.0051 (0.0024).
become more homogeneous than the ancestral rural populations, because they are composed of individuals that only vary
little in personality. If behavioral flexibility implies a greater
diversity of personalities, then urban invaders will have fewer
personalities than their ancestors.
Invasion of urban areas was associated with reduced fearfulness as reflected by short mean flight distances and reduced
variance in flight distance. The difference in variance in flight
distance between rural and urban populations was positively
correlated with time since urbanization, suggesting that recently urbanized species have reduced behavioral variance
in urban compared with rural populations, whereas the reverse
was the case for species that have been urbanized since long.
Temporal change in flight responses to humans resemble what
happens during domestication because domesticated animals
show little or no fear reaction and no stress response when
approached by humans or other domestic animals such as dogs
(e.g., Kohane and Parsons 1988). Such behavioral adaptation
to domestication has been experimentally induced in fruitflies (Kohane and Parsons 1986, 1987), foxes (Belyaev 1969,
1979; Trut et al. 2009), and chickens (Campler et al. 2009;
Table 2
Urbanization in relation to mean and variance in FID and body mass
(g) in birds
Variable
Species
Mean FID
Variance FID
Body mass
Contrasts
Mean FID
Variance FID
Body mass
Wald v2
df
229.23
91.06
547.93
1
1
1
,0.0001
,0.0001
,0.0001
5.02 (0.33)
21.85 (0.19)
2333 (0.14)
1.21
14.38
2.33
1
1
1
0.27
0.0002
0.13
0.27 (0.25)
20.50 (0.13)
20.27 (0.17)
P
Estimate (SE)
The models had the statistics v2 ¼ 1082.65, r2 ¼ 0.21, P , 0.0001 and
F ¼ 7.81, df ¼ 3, 126, r2 ¼ 0.06, P , 0.0001
Figure 4
Difference in population density between rural and urban
populations in relation to difference in variance in FID between rural
and urban populations of the same species. The line is the linear
regression line. The model for species weighted by log10-transformed
sample size had the statistics F ¼ 4.35, df ¼ 1, 39, r 2 ¼ 0.10, P ¼ 0.044,
slope (SE) ¼ 0.760 (0.365), whereas the model for contrasts had the
statistics F ¼ 4.51, df ¼ 1, 38, r2 ¼ 0.11, P ¼ 0.040, slope (SE) ¼ 1.532
(0.721).
370
Wirén et al. 2009). Flight behavior has a genetic basis and
responds to artificial selection, as shown by rapid change in
behavior during the domestication process. This scenario is
consistent with recent evidence for behavioral flexibility and
personalities in house sparrows (Liker and Bokony 2009).
House sparrows tended to be better able to solve novel foraging tasks when originating from urban than rural populations
(Liker and Bokony 2009). Because house sparrows have been
urbanized probably since the first cities appeared in the Middle East (Summers-Smith 1963), it is likely that they have
evolved novel behavioral flexibility after the initial invasion
of urban areas.
Species that have successfully colonized urban areas often
have much larger population densities than the ancestral
populations in rural areas, and this has been associated with
higher resource abundance (e.g., Klausnitzer 1989; Stephan
1999; Møller 2008b). High urban population densities compared with rural densities imply that urban populations have
responded to greater local resource abundance and that populations have increased in size compared with rural populations. Because climatic conditions in urban areas are more
favorable than in rural areas, we may expect less weatherinduced mortality and hence a population that is closer to
carrying capacity. Therefore, higher population densities in
urban areas imply more intense intraspecific competition, selection for a more diversified way of exploiting habitats and
resources and ultimately suites of behavior that are associated
with specialization on such habitats and resources. If an urban
population of birds consists of individuals with large variance
in FID, and hence individuals with either short or long FIDs,
such a population will be able to exploit habitats with both
high and low density of predators. In contrast, a population of
individuals only having short or long FIDs will mainly be able
to exploit habitats with few predators, or habitats with many
predators. This should result in the population with greater
variance in FID having higher population density than the
population with low variance. Here, I have shown that the
difference in population density between urban and rural
habitats is positively correlated with the difference in variance
in flight distance between urban and rural habitats. Thus,
a greater diversity of behavior within species is associated with
a greater ecological success estimated as a large population
density.
In conclusion, I have shown complex dynamics in behavioral
flexibility for flight distance of birds before, during and after
successful invasion of urban areas. These results suggest that
species with flexible behavior are disproportionately successful
colonizers, colonization is accompanied by loss of flexibility,
and that behavioral flexibility is subsequently regained as populations become adapted to the urban environment.
SUPPLEMENTARY MATERIAL
Supplementary material can be found at http://www.beheco
.oxfordjournalsorg/.
REFERENCES
Becker WA. 1984. Manual of quantitative genetics. Pullman, Washington:
Academic Enterprises.
Belyaev DK. 1969. Domestication of animals. Sci J. 5:47.
Belyaev DK. 1979. Destabilizing selection as a factor in domestication.
J Hered. 70:301–308.
Blondel J, Ferry C, Frochot B. 1970. La méthode des indices ponctuels
d’abondance (I.P.A.) au des relevés d’avifaune par ‘‘stations
d’ecoute’’. Alauda. 38:55–71.
Blumstein DT. 2006. Developing an evolutionary ecology of fear: how
life history and natural history traits affect disturbance tolerance in
birds. Anim Behav. 71:389–399.
Behavioral Ecology
Boissy A. 1995. Fear and fearfulness in animals. Q Rev Biol.
70:165–191.
Campler M, Jongren M, Jensen P. 2009. Fearfulness in red junglefowl
and domesticated White Leghorn chickens. Behav Proc. 81:39–43.
Cooke AS. 1980. Observations on how close certain passerine species
will tolerate an approaching human in rural and suburban areas.
Biol Conserv. 18:85–88.
Cramp S, Perrins CM, editors. 1977–1994. The birds of the Western
Palearctic. Vols. 1–9. Oxford: Oxford University Press.
Dall SRX, Houston AI, McNamara JM. 2004. The behavioural ecology
of personality: consistent individual differences from an adaptive
perspective. Ecol Lett. 7:734–739.
Dingemanse NJ, Wright J, Kazem AJN, Thomas DK, Hickling R,
Dawnay N. 2007. Behavioural syndromes differ predictably between
12 populations of three-spined stickleback. J Anim Ecol. 76:
1128–1138.
Draper NR, Smith H. 1981. Applied regression analysis. 2nd ed. New
York: John Wiley.
Felsenstein J. 1985. Phylogenies and the comparative method. Am
Nat. 125:1–15.
Fløystrup A. 1920. Fugleliv i Kjøbenhavn: Iagttagelser fra Østre Anlæg
og Botanisk Have. Dansk Orn Foren Tidsskr. 14:1–60.
Fløystrup A. 1925. Fugleliv i Kjøbenhavn: Fortsatte iagttagelser fra
Østre Anlæg og Botanisk Have. Dansk Orn Foren Tidsskr. 19:1–18.
Garamszegi LZ, Eens M, Török J. 2008. Birds reveal their personality
when singing. PLoS ONE. 3:e2647.
Garamszegi LZ, Eens M, Török J. 2009. Behavioural syndromes and
trappability in free-living collared flycatchers, Ficedula hypoleuca.
Anim Behav. 77:803–812.
Garamszegi LZ, Møller AP. Forthcoming. 2009. Effects of sample size
and intraspecific variation in phylogenetic comparative studies:
a meta-analytic review. Biol Rev.
Garland T Jr., Harvey P, Ives AR. 1992. Procedures for the analysis of
comparative data using phylogenetically independent contrasts.
Syst Biol. 41:18–32.
Garland T Jr., Ives AR. 2000. Using the past to predict the present:
confidence intervals for regression equations in phylogenetic comparative methods. Am Nat. 155:346–364.
Gliwicz J, Goszczynski J, Luniak M. 1994. Characteristic features of
animal populations under synurbanization – the case of the Blackbirds and the striped field mouse. Mem Zool. 49:237–244.
Glutz von Blotzheim UN, Bauer KM, editors. 1985–1997. Handbuch
der Vögel Mitteleuropas. Vols. 1–14. Wiesbaden (Germany): AULAVerlag.
Gosling SD, John OP. 1999. Personality dimensions in nonhuman
animals: a cross-species review. Curr Dir Psychol Sci. 8:69–75.
Graczyk R. 1974. The experiment of settling of urbanized population
of Poznan blackbird (Turdus merula L.) to Kiev (USSR) and examination of certain elements of innate behaviour. Roczn Akad Roln
Pozn. 65:49–64.
Gram C. 1908. Fuglelivet i København og omegn for halvhundrede aar
siden. Dansk Orn Foren Tidsskr. 3:27–36.
Hackett SJ, Kimball RT, Reddy S, Bowie RCK, Braun EL, Braun MJ,
Chojnowski JL, Cox WA, Han K-L, Harshman J, et al. 2008. A phylogenomic study of birds reveals their evolutionary history. Science.
320:1763–1768.
Hollander FA, Van Overveld T, Tokka I, Matthysen E. 2008. Personality and nest defence in the great tit (Parus major). Ethology. 114:
405–412.
Jones KE, Purvis A. 1997. An optimum body size for mammals? Comparative evidence from bats. Funct Ecol. 11:751–756.
Jønsson KA, Fjeldså J. 2006. A phylogenetic supertree of oscine passerine birds (Aves: Passeri). Zool Scripta. 35:149–186.
Kark S, Iwaniuk A, Schalimtzek A, Banker E. 2007. Living in the city:
can anyone become an ‘‘urban exploiter’’? J Biogeogr. 34:638–651.
Klausnitzer B. 1989. Verstädterung von Tieren. Wittenberg Lutherstadt (Germany): Neue Brehm-Bücherei.
Kohane MJ, Parsons PA. 1986. Environment-dependent fitness differences in Drosophila melanogaster: temperature, domestication and the
alcohol-dehydrogenase locus. Heredity. 57:289–304.
Kohane MJ, Parsons PA. 1987. Mating ability in laboratory-adapted
and field-derived Drosophila melanogaster: the stress of domestication.
Behav Genet. 17:541–558.
Kohane MJ, Parsons PA. 1988. Domestication: evolutionary change
under stress. Evol Biol. 23:30–48.
Møller • Møller urbanization and flight distance flexibility
Liker A, Bokony V. 2009. Larger groups are more successful in innovative problem solving in house sparrows. Proc Natl Acad Sci
USA. 106:7893–7898.
Marzluff JM, Bowman R, Donnelly RE, editors. 2001. Avian conservation and ecology in an urbanizing world. New York: Kluwer.
Møller AP. 2008a. Flight distance and blood parasites in birds. Behav
Ecol. 19:1305–1313.
Møller AP. 2008b. Flight distance of urban birds, predation and selection for urban life. Behav Ecol Sociobiol. 63:63–75.
Møller AP. 2008c. Flight distance and population trends in European
breeding birds. Behav Ecol. 19:1095–1102.
Møller AP. 2009. Successful city dwellers: a comparative study of the
ecological characteristics of urban birds in the Western Palearctic.
Oecologia. 159:849–858.
Møller AP, Birkhead TR. 1994. The evolution of plumage brightness
in birds is related to extra-pair paternity. Evolution. 48:1089–1100.
Møller AP, Nielsen JT. 2007. Malaria and risk of predation: a comparative study of birds. Ecology. 88:871–881.
Møller AP, Nielsen JT, Garamszegi LZ. 2008. Risk taking by singing
males. Behav Ecol. 19:41–53.
Neter J, Kutner MH, Nachtsheim CJ, Wasserman W. 1996. Applied
linear statistical models. Chicago (IL): Irwin.
Purvis A, Rambaut A. 1995. Comparative analysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Comp Appl Biosci. 11:247–251.
Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ. 2007.
Integrating animal temperament within ecology and evolution. Biol
Rev. 82:291–318.
Sasvári L. 1985. Keypeck conditioning with reinforcements in two
different locations in thrush, tit, and sparrow species. Behav Proc.
11:245–252.
Shochat E, Warren PC, Faeth SH, McIntyre NE. 2006. From patterns
to emerging processes in mechanistic urban ecology. Trends Ecol
Evol. 21:186–191.
371
Sibley CG, Ahlquist JE. 1990. Phylogeny and classification of birds:
a study in molecular evolution. New Haven and London: Yale
University Press.
Sih A, Bell A, Johnson JC. 2004b. Behavioral syndromes: an ecological
and evolutionary overview. Trends Ecol Evol. 19:372–378.
Sih A, Bell AM. 2008. Insights for behavioral ecology from behavioral
syndromes. Adv Study Behav. 38:227–281.
Sih A, Bell AM, Johnson JC, Ziemba RE. 2004a. Behavioral syndromes:
an integrative overview. Q Rev Biol. 79:241–277.
Sokal RR, Rohlf FJ. 1995. Biometry. 3rd ed. New York: Freeman.
Stamps JA. 2007. Growth-mortality trade-offs and ‘personality traits’ in
animals. Ecol Lett. 10:355–363.
Stephan B. 1999. Die Amsel. Wittenberg-Lutherstadt (Germany):
Neue Brehm-Bücherei.
Summers-Smith D. 1963. The house sparrow. London: Collins.
Trut L, Oskina I, Kharlamova A. 2009. Animal evolution during domestication: the domesticated fox as a model. Bioessays. 31:349–360.
van Oers K, Drent PJ, de Goede P, van Noordwijk AJ. 2004. Realized
heritability and repeatability of risk-taking behaviour in relation to
avian personality. Proc R Soc Lond B. 271:65–73.
Ward AJW, Thomas P, Hart PJB, Krause J. 2004. Correlates of boldness
in three-spined sticklebacks (Gasterosteus aculeatus). Behav Ecol
Sociobiol. 55:561–568.
Wilson ADM, Godin JGJ. 2009. Boldness and behavioral syndromes in
the bluegill sunfish, Lepomis macrochirus. Behav Ecol. 20:231–237.
Wilson DS, Clark AB, Coleman K, Dearstyne T. 1994. Shyness and
boldness in humans and other animals. Trends Ecol Evol. 9:
442–446.
Wirén A, Gunnarsson U, Andersson L, Jensen P. 2009. Domesticationrelated genetic effects on social behavior in chickens: effects of
genotype at a major growth quantitative trait locus. Poultry Sci.
88:1162–1166.
Wolf M, van Doorn GS, Leimar O, Weissing FJ. 2007. Life-history
trade-offs favour the evolution of animal personalities. Nature.
447:581–584.