Ecological Applications, 16(5), 2006, pp. 1696–1705
Ó2006 by the Ecological Society of America
CHERNOBYL AS A POPULATION SINK FOR BARN SWALLOWS:
TRACKING DISPERSAL USING STABLE-ISOTOPE PROFILES
A. P. MøLLER,
K. A. HOBSON,
T. A. MOUSSEAU,
AND A. M. PEKLO
Laboratoire de Parasitologie Evolutive, CNRS UMR 7103, Universite
´Pierre et Marie Curie, Ba
ˆt. A, 7e
7 quai St. Bernard, Case 237, 75252 Paris Cedex 05, France
Canadian Wildlife Service, 115 Perimeter Road, Saskatoon, SK S7N 0X4, Canada;
Department of Biological Sciences, University of South Carolina, Columbia, South Carolina 29208 USA
Zoological Museum, National Museum of the Natural History, National Academy of Sciences,
15 Bohdan Chmelnitsky Street, 01030 Kyiv-30, Ukraine
Abstract. Stable-isotope proﬁles of feathers can reveal the location or habitat used by
individual birds during the molting period. Heterogeneity in isotope proﬁles will reﬂect
heterogeneity in molt locations, but also heterogeneity in breeding locations, because spatial
heterogeneity in molt locations will be congruent with spatial heterogeneity in breeding
locations in species with high connectivity between breeding and molting sites. We used
information on the congruence of spatial heterogeneity in molt and breeding location to study
population processes in Barn Swallows (Hirundo rustica) from a region near Chernobyl,
Ukraine, that has been radioactively contaminated since 1986; from an uncontaminated
control region near Kanev, Ukraine; and from a sample of pre-1986 museum specimens used
to investigate patterns prior to the nuclear disaster at Chernobyl, from both regions. Previous
studies have revealed severe reductions in Barn Swallow reproductive performance and adult
survival in the Chernobyl region, implying that the population is a sink and unable to sustain
itself. Female Barn Swallows are known to disperse farther from their natal site than males,
implying that female stable-isotope proﬁles should tend to be more variable than proﬁles of
males. However, if the Barn Swallows breeding at Chernobyl are not self-sustaining, we would
expect males there also to originate from a larger area than males from the control region. We
found evidence that the sample of adult Barn Swallows from the Chernobyl region was more
isotopically heterogeneous than the control sample, as evidenced from a signiﬁcant
correlation between feather d
C and d
N values in the control region, but not in the
Chernobyl region. Furthermore, we found a signiﬁcant difference in feather d
between regions and periods (before and after 1986). When we compared the variances in
C values of feathers, we found that variances in both sexes from post-1986 samples from
Chernobyl were signiﬁcantly larger than variances for feather samples from the control
region, and than variances for historical samples from both regions. These ﬁndings suggest
that stable-isotope measurements can provide information about population processes
following environmental perturbations.
Key words: Barn Swallows; Chernobyl, Ukraine; correlation of stable-isotope proﬁles; Hirundo rustica;
sinks; sources; variance in stable-isotope proﬁles.
The mean and variance of the reproductive output of
populations of free-living organisms vary considerably.
Such variance is of considerable importance because
populations that produce a larger and less variable
number of recruits to future generations are considered
less prone to extinction (Lande et al. 2003, Roff 2001).
Two variance components are important contributors to
such variability. Environmental variance similarly af-
fects all individuals in a population due to variation in
common environmental conditions, whereas demo-
graphic variance arises from individual variation in
realization of survival and reproduction. Whether a
population constitutes a net contributor to the main-
tenance of a globally viable population will depend on
both environmental and demographic variances of each
of the composite subpopulations. Pulliam (1988) origi-
nally suggested that populations with positive popula-
tion growth, r, are source populations, while
populations with negative rare sinks. Immigration
may allow individuals to immigrate into sites where
local recruitment would be insufﬁcient for persistence of
a population (Pulliam 1996). Source populations were
subsequently deﬁned as populations that have a positive
intrinsic growth rate, while sink populations were
deﬁned as being unable to sustain not only their own
populations, but also neighboring populations (Dias
1996, Hanski 1999). Successful identiﬁcation of source
and sink populations is therefore of considerable
Manuscript received 8 December 4004; revised 21 November
2005; accepted 14 December 2005; ﬁnal version received 6
February 2006. Corresponding Editor: F. C. James.
importance for conservation purposes, but also for
ecological research. Many studies have found evidence
consistent with predictions for source–sink metapopu-
lations using a number of different ecological and
genetic approaches (Dias 1996, Paradis 1995, Pulliam
1996, Pulliam and Danielson 1991, Stacey et al. 1997).
Here we propose a novel approach to the identiﬁcation
and study of source and sink populations. This method
is based on heterogeneity in stable-isotope proﬁles
across individuals in the different parts of a metapopu-
lation caused by different degrees of admixture in source
and sink populations (Hobson 2005).
Here we argue that source and sink populations under
certain assumptions can be identiﬁed using heterogene-
ity in stable-isotope proﬁles. The underlying assump-
tions are as follows: First, stable-isotope abundance
recorded in the feathers or other tissues reﬂect the
isotope contents of food, revealing information about
the habitats and the geographical locations where the
tissue was developed (reviews in Hobson 1999, 2004,
Rubenstein and Hobson 2004). In the case of metabol-
ically inactive tissues such as scales, feathers, and hair,
isotope proﬁles provide a frozen snapshot of the
environmental conditions experienced during develop-
ment of the character. Individuals with speciﬁc charac-
teristics will subsequently carry this signature until the
next molt. As the feathers of birds wear out, they are
replaced in a process of molt, normally at least once a
year; in the Barn Swallow once a year. Second, the ratio
of stable isotopes in the environment varies spatially,
showing increasing heterogeneity with increasing area
(e.g., Still et al. 2003, Urton and Hobson 2005). For
example, previous studies of Palaearctic breeding birds
that typically molt ﬂight feathers on their wintering
grounds have indicated that mixtures of wintering
populations can be discerned based on variation in
feather isotope proﬁles among individuals on the
breeding grounds (e.g., Chamberlain et al. 2001, Møller
and Hobson 2004). Third, population sinks are deﬁned
as requiring immigration of individuals to sustain
populations, while sources do not (Pulliam 1988). This
implies that individuals in population sinks have come
from a larger area than those from population sources.
Fourth, individuals from a single population will
generally be found in the same region while the tissue
is being grown (Berthold 2001), so that a population of
individuals that have come from a larger area will tend,
by necessity, to have a greater variance in isotope ratio.
Fifth, spatial heterogeneity in a sample of individuals
from the breeding grounds of birds also reﬂects spatial
heterogeneity from the wintering grounds where the
tissue has been grown. Sixth, sex differences in hetero-
geneity in samples from the breeding grounds are
expected to reﬂect spatial heterogeneity from the
wintering grounds becausefemaleBarnSwallows
generally disperse considerably further than males.
Seventh, in species with migration divides, where one
population takes one migratory route and a second
takes another migratory route (Berthold 2001), single
breeding populations can consist of an admixture of
birds from different wintering populations. R. Ambro-
sini, A. P. Møller, and N. Saino (unpublished manuscript)
have analyzed spatial distribution patterns of over 1000
recoveries of Barn Swallows (Hirundo rustica) from the
African winter quarters of birds breeding in the Western
Palearctic. These analyses, based on Mantel tests,
revealed a high degree of congruence in spatial patterns
in the winter quarters and the breeding grounds. Thus,
breeding birds from Western Europe winter in Western
Africa, those from Eastern Europe winter in East Africa,
those from Northern Europe winter in Southern Africa,
and those from Southern Europe winter in Northern
Africa. In other words, at large spatial scales, a
heterogeneous sample of birds from the winter quarters,
when considering the isotope composition of feathers
caused by incorporation of stable isotopes during the
annual molt in winter, also signiﬁes a heterogeneous
sample of birds in terms of previous breeding-ground
origins. Barn Swallows breeding in Europe could be
divided into western and eastern populations that
differed in migration routes and wintering areas (R.
Ambrosini, A. P. Møller, and N. Saino, unpublished
manuscript), with birds from Finland, parts of Sweden,
and eastern Europe migrating to eastern Africa along an
eastern ﬂyway, while birds from Norway, Denmark, and
western Europe use a western ﬂyway. Breeding pop-
ulations in Denmark consist of an admixture of the two
populations (Møller and Hobson 2004). Thus, our
assumption that heterogeneity in stable-isotope proﬁles
caused by heterogeneity in winter grounds also reﬂects
heterogeneity in terms of breeding origin is supported by
empirical evidence. These seven assumptions lead to the
critical test that if we compare variances of isotope
ratios of two or more samples, all things being equal, the
sample with the greater variance is a sink to a greater
extent than the one with the smaller variance.
However, there have, to the best of our knowledge,
been no attempts to use stable-isotope techniques to
investigate such admixture of populations to study
population processes, including identiﬁcation of sources
and sinks. Several studies used stable isotopes to
examine population demographics and population
processes (Hershey et al. 1993, Marra et al. 1998, Kelly
et al. 2002, Rubenstein et al. 2002, Sillett and Holmes
2002). Using isotopic variability in feathers among
individuals to infer diversity of population origins will
necessarily depend on the degree of variability in stable
isotopes across potential contributing populations
(Hobson et al. 2004, Hobson 2005). In general, we
should expect breeding populations that constitute sinks
to reveal greater variance in feather isotope proﬁles than
populations that constitute source populations, as
argued in the previous paragraphs. Males often differ
from females in dispersal distance from site of birth to
site of breeding, with female birds generally dispersing
much further than males, while mammals commonly
October 2006 1697VARIATION IN ISOTOPES IN BARN SWALLOWS
show the opposite pattern (e.g., Greenwood 1980,
Clobert et al. 2001). Such sex differences, if well
established in particular species, provide a unique
opportunity to use one sex as a natural control group
and the other sex as the ‘‘experimental’’ group when
studying sources and sinks. Although it is often difﬁcult
to measure the range of isotopic variation in diet used by
birds on their wintering grounds, where they molt
feathers, in source–sink situations, we would expect
individuals of the more-dispersing sex to show an
increase in the variance in isotope proﬁle due to
admixture of individuals of different origins (potentially
over hundreds or thousands of kilometers). In contrast,
the more philopatric sex should show less variation. This
prediction has a strong inference because individuals of
the two sexes would be expected to otherwise be similar
in many respects. Furthermore, if there are temporal
controls in terms of tissue samples from before and after
the creation of a sink population, strong inferences
could be predicted with respect to temporal changes in
variances in isotope proﬁles. Recently, Hobson et al.
(2004) demonstrated that individual isotopic outliers
among male Ovenbirds (Seirus aurocapillus) and Amer-
ican Redstarts (Setophaga ruticilla) in breeding popula-
tions in North America represented birds that had
dispersed from other (unknown) areas. These authors
suggested that isotopic variance within breeding pop-
ulations could be used to establish minimum estimates
of recruitment into populations of interest, provided
that sufﬁcient isotopic heterogeneity occurs among
possible molting areas of source populations. We
adopted a similar approach to examine the structure of
Barn Swallows from two different regions in Ukraine
that differ in sustainability.
The aims of this study were to investigate whether
stable-isotope proﬁles could be used to identify a
population sink in a wild population of Barn Swallows
from Ukraine (see Møller and Mousseau [2001, 2003] for
a description of this system). We can make three
predictions. First, positive correlation between stable
isotopes should break down when population admixture
increases. For several elements, stable-isotope relative
abundance in the food shows gradual clinal variation on
continental scales (e.g., Still et al. 2003, Bowen et al.
2005). So, we should expect a sample of individuals from
a population to be normally distributed in space because
individuals from a given population winter and molt in a
speciﬁc location, with both genetic and environmental
factors contributing to heterogeneity in this spatial
distribution. Such a normally distributed spatial distri-
bution of individuals should cause stable-isotope ratios
for C and N to be positively correlated. Such positive
correlation has previously been found within samples
(e.g., Møller and Hobson 2004), but has also been
demonstrated for C and N stable-isotope ratios within
food webs (Kelly 2000). However, if a sample from a
speciﬁc region shows signiﬁcant admixture of individu-
als from different breeding and, hence, wintering
populations, positive covariance between stable-isotope
ratios for C and N should no longer exist. The reason
for this claim is that sampling sites should no longer
show a normal distribution along the clinal gradients of
the different stable isotopes in the environment. Second,
males and females should differ in mean and variance of
stable-isotope proﬁles due to sex differences in natal
dispersal, again assuming that spatial heterogeneity in
breeding sites is reﬂected in spatial heterogeneity in
molting sites. Because female Barn Swallows disperse
from their natal site to their future breeding site more
than twice as far as males, with measured maximum
dispersal distances reaching 700 km (Møller 1994, Glutz
von Blotzheim and Bauer 1994), we would expect that
females from the two areas in Ukraine would be more
different in means and variances of isotope proﬁles than
males. This prediction is based on the fact that females
originate from a larger area than males, and that the
variance in distances between the origins of females is
greater than the variance in distances between the
origins of males. Since long-distance dispersal is much
more common in female Barn Swallows than in males
(Møller 1994), we should expect the variance in stable-
isotope proﬁle to be greater in females than in males.
Third, the variance in stable-isotope proﬁle should be
greater in population sinks than in population sources.
Capture–mark–recapture analyses of Barn Swallows
from the Chernobyl region has shown that adult survival
is 60%lower than from the Kanev region, and estimates
of reproductive success have shown that annual
fecundity is 20%lower than from the Kanev region
(Møller et al. 2005). Such large differences would suggest
that Barn Swallows in the Chernobyl region are
sustained by immigration from elsewhere. Such immi-
gration should be particularly obvious in males, because
males from the control region near Kanev were expected
to be relatively homogeneous in isotope composition
mainly due to local recruitment, while the males in the
Chernobyl region should have a larger proportion of
immigrants. We predicted that the patterns of stable
isotopes in Kanev would be the same before 1986 and in
2000, but that the patterns would differ between these
periods at Chernobyl. Explicitly, we would predict an
increase in the stable isotopic variance at Chernobyl,
particularly amongst males.
MATERIALS AND METHODS
The Barn Swallow is a small (;20 g) passerine that
feeds on insects, mainly Diptera and Hymenoptera
caught on the wing. Barn Swallows arrive in Europe
from the African winter quarters in April–June, and
leave in August–October (Møller 1994). Males and
females build nests inside buildings, where the female
lays a clutch, usually 4–5 eggs. More than 50%lay a
second clutch. Barn Swallows breed solitarily or in
colonies that can exceed 120 pairs. Long-tailed males
enjoy a mating advantage, as demonstrated by observa-
A. P. MØLLER ET AL.1698 Ecological Applications
Vol. 16, No. 5
tions and experiments (Møller 1994). Male Barn
Swallows in Chernobyl have dramatically increased
asymmetry in their elongated outermost tail feathers
and much paler throat coloration compared to males
predating the Chernobyl accident (Møller 1993). Long-
tailed males in particular are very pale compared to
controls (Camplani et al. 1999). Genetic studies have
reported increased mutation rates in Barn Swallows
from the radioactively contaminated region around
Chernobyl in Ukraine (Ellegren et al. 1997). Further-
more, the frequency of partial albinism is elevated
compared to other populations, and such phenotypic
deviants are associated with reduced ﬁtness of Barn
Swallows (Møller and Mousseau 2001). Of the pheno-
typic traits measured, those that differ the most between
Chernobyl and an uncontaminated control region near
Kanev, are also those that are most strongly associated
with male mating success (Møller and Mousseau 2003).
Barn Swallows from Ukraine winter in Southern Africa
(Fig. 1), where they undergo a complete molt. Barn
Swallows from Chernobyl and Kanev winter in the same
areas (Fig. 1). There is no evidence suggesting that males
and females winter in different areas (R. Ambrosini, A.
P. Møller, and N. Saino, unpublished manuscript).
We studied Barn Swallows in a region around
Chernobyl, Ukraine (Shestopalov 1996), by visiting
villages and checking collective farms for the presence
of Barn Swallows (Fig. 2). Once such a farm had been
located, we recorded radiation levels and captured adult
Barn Swallows for measurements and subsequent
banding. Capture was made with mist nets, providing
a random sample of adults from the breeding farms. Our
own ﬁeld measurements of radiation at the ground level
using a hand-held dosimeter (Inspector, SE Interna-
tional, Summertown, Tennessee, USA) revealed levels of
radiation of 0.390 60.317 mR/h (milliroentgens per
hour; mean 6SE) at 14 farms in the Chernobyl region.
As a control region, we used Kanev (;150 km southwest
of Kiev), which has a relatively low level of contami-
nation (Fig. 2). Mean levels of radiation were 0.025 6
0.002 mR/h at ﬁve farms in this control region. We used
the Chernobyl region for our study because radioactive
contamination arising from the explosion of a nuclear
reactor on 26 April 1986 released vast amounts of
radioactive material into the environment. Most of the
radioactive isotopes remain in the environment due to
their long half-life, with serious detrimental effects on
the reproductive success and adult survival of Barn
Swallows (Møller et al. 2005). We studied Barn
Swallows in the two study regions 6–12 June 2000 (A.
P. Møller, T. A. Mousseau). The maximum distance
between farms in the Chernobyl region was ,20 km,
and in the Kanev region it was ,8 km. These distances
are small compared to natal dispersal distances, and we
considered it unlikely that this spatial scale would have
any effect on variance is stable-isotope proﬁles. The
wintering grounds of Ukrainian-breeding Barn Swal-
lows, where the annual molt takes place, is shown in Fig.
1, based on recoveries of banded Barn Swallows from
Ukraine (during the breeding season May–August) and
South Africa (during winter December–February).
Information on recoveries of banded birds was kindly
provided by the Ukrainian Ringing Center, EURING
(European Union for bird Ringing), and SAFRING
(South African Bird Ringing Unit).
As a second control sample from the Chernobyl
region and the region around Kanev, we used samples of
feathers from the Zoological Museum, Kiev, Ukraine.
We only used adult birds collected during the breeding
season (June–July) to avoid inclusion of specimens that
might be migrants from more northern populations.
Specimens attributed to the Chernobyl region derived
from the same locations that are now within the
contaminated areas (Fig. 2), while specimens from a
similar area southeast of Kiev surrounding Kanev were
considered to be controls. While long-term storage may
affect samples from museum specimens, any changes
FIG. 1. The map shows connections between the banding
sites of Barn Swallows breeding in Ukraine and winter locations
October 2006 1699VARIATION IN ISOTOPES IN BARN SWALLOWS
should be uniform across samples and so should not
inﬂuence our results in a heterogeneous way. The sample
of museum specimens was collected by shotgun, whereas
the recent samples of live birds were obtained by using
mist nets. However, because the museum samples were
collected with shotgun from breeding populations, we
feel justiﬁed to assume that they also represent a random
sample. Sample sizes for the two sexes, regions, and
periods are reported in Table 1.
Upon capture, we collected the two outermost tail
feathers from each adult. The feathers for each bird were
stored at room temperature in a plastic bag in complete
darkness until measurements were made. The tail
feathers of adult Barn Swallows are molted in the
African winter quarters (Cramp 1988; A. P. Møller,
unpublished data from South Africa, Namibia, and
Ghana), and so isotopic variability in feathers represents
variability in the isotopic proﬁle of the prey being
related to the food web from which they come. The
feathers from the museum specimens were the same as
those for the ﬁeld specimens.
Feathers were analyzed blindly with respect to
information on origin and sex. Feathers were ﬁrst
cleaned of surface oils by rinsing several times in a 2:1
chloroform : methanol solution followed by air drying in
a fume hood for several days. Feather vanes were then
subsampled and 1 mg was weighed into small tin cups.
These samples were then combusted in a Robo Prep
elemental analyzer interfaced with a 20:20 isotope-ratio
mass spectrometer (Europa Scientiﬁc, Manchester, UK).
gases were measured for their
stable-isotope ratios in dnotation relative to Pee Dee
Belemnite (PDB) and atmospheric AIR standards,
respectively, according to the formula presented in
Hobson (1995). Based on thousands of measurements
of internal laboratory standards (egg albumen and
baleen keratin), we estimated measurement error to be
0.1øand 0.3øfor d
C and d
We tested for correlation between feather d
N values using the different samples. Heterogeneity
FIG. 2. Location of villages with study farms in 2000 and levels of radiation around Chernobyl and Kanev, Ukraine. The
shaded region around Chernobyl demarks the area where .15 curies of
Cs per km
were deposited in 1986 (CIA 1986). The
radiation levels in the contaminated areas were 0.390 60.317 mR/h (mean 6SE) at 14 breeding farms in the Chernobyl region. In
the control region near Kanev (;150 km southeast of Chernobyl) radiation levels were, on average, an order of magnitude lower
(0.025 60.002 mR/h). Approximate locations of villages with study farms are indicated either as black diamonds (in contaminated
areas) or stars (relatively uncontaminated control areas).
A. P. MØLLER ET AL.1700 Ecological Applications
Vol. 16, No. 5
in Pearson correlation coefﬁcients was tested using the
procedure described in Sokal and Rohlf (1995). We did
not transform data since they did not differ signiﬁcantly
from normal distributions according to the Shapiro-
Wilk statistic. We tested for sex, region, and time effects
on isotope proﬁles with Welch’s ANOVA, which does
not require equality of variances. Means for groups were
compared using Fisher’s protected least mean-squares
test, adjusting means for unequal variances by weighting
by the reciprocal of the sample variances of the group
means. Variances in d
N values were
compared among samples for different periods using
Bartlett’s test for homogeneity of variances (Sokal and
Rohlf 1995). All statistical tests were made with JMP
(SAS Institute 2000).
We used sequential Bonferroni correction to assess the
tablewide Type I error rate by adjusting signiﬁcance
level to the number of tests (Holm 1979, Wright 1992).
Strict application of this method severely reduces the
power of tests (Wright 1992), but such sacriﬁcial loss of
power can be avoided by choosing an experiment-wise
error rate higher than the accepted 5%. We used 10%as
recommended by Wright (1992) and Chandler (1995).
Frequency distribution of stable-isotope values in feathers
and correlation among stable isotopes
There were clearly unimodal distributions of d
N feather values with distributions not differing
signiﬁcantly from normality. When split among sex,
region, and period categories, none of the distributions
differed signiﬁcantly from a normal distribution after
sequential Bonferroni correction for eight statistical
Correlations between d
C and d
N values for Barn
Swallows from the recent ﬁeld samples differed between
regions. There was a nonsigniﬁcant relationship in
Chernobyl (F¼0.16, df ¼1, 90, r
Fig. 3a), but a signiﬁcant positive relationship in Kanev
(F¼17.48, df ¼1, 52; r
¼0.25, P,0.0001, slope 6SE ¼
0.28 60.07; Fig. 3b).
The correlations for the recent ﬁeld samples were as
follows: males from Chernobyl (r¼0.009, t¼0.07, df ¼
51, P¼0.95), females from Chernobyl (r¼0.15, t¼
0.92, df ¼37, P¼0.37); males from Kanev (r¼0.35, t¼
1.81, d.f. ¼23, P¼0.08), and females from Kanev (r¼
0.654, t¼4.49, df ¼27, P¼0.0001). The two correlation
coefﬁcients between d
C and d
N values for females
were signiﬁcantly different from each other (t¼3.41, P
,0.001), while the two coefﬁcients for males were not
signiﬁcantly different (t¼0.87, P¼0.39).
Differences in means and variances in stable-isotope
proﬁles between regions
Summary statistics for d
C and d
N values for the
various sex, age, and region categories are reported in
Table 1. Bartlett’s test for unequal variances revealed a
signiﬁcant difference for d
C(F¼15.07, df ¼7, P,
0.0001). However, a test for equal means for d
not reach signiﬁcance (Welch’s ANOVA, F¼2.09, df ¼
7, 26.48, P¼0.081).
In contrast to the results for d
C, there was no
signiﬁcant difference in variance for d
N (Table 1). A
posteriori tests revealed a signiﬁcant difference in means
between regions, with a higher mean value in Chernobyl
TABLE 1. Results from tests for unequal variances with Barlett’s test and for means with Welch’s
ANOVA for d
C and d
N from Barn Swallow feathers in relation to sex, region (Chernobyl
and Kanev, Ukraine), and period (pre-1986, post-1986).
the mean NSD
Pre-1986 6 2.79 2.28 3 1.03 0.70
Post-1986 39 8.62 6.36 53 7.34 4.52
Pre-1986 8 1.73 1.44 21 2.37 1.68
Post-1986 29 2.29 1.80 25 2.47 1.83
Pre-1986 6 0.94 0.72 3 0.61 0.42
Post-1986 39 1.18 0.95 53 1.43 1.15
Pre-1986 8 1.10 0.75 21 1.04 0.84
Post-1986 29 1.30 1.13 25 1.06 0.78
Note: For d
C, Bartlett’s test, F¼15.07, df ¼7, P,0.0001; Welch’s ANOVA, F¼2.09, df ¼7,
26.5, P¼0.08. For d
N, Bartlett’s test, F¼0.96, df ¼7, P¼0.84; Welch’s ANOVA, F¼4.18, df ¼
7, 24.1, P¼0.004.
October 2006 1701VARIATION IN ISOTOPES IN BARN SWALLOWS
than in Kanev (Fisher’s protected least squares differ-
ence test weighted by the sample variances of the group
means, P,0.001). Likewise there was a signiﬁcant
difference in means between periods (Fisher’s protected
least squares difference test, P,0.001), with lower
values in the museum samples as compared to the recent
samples. The difference between Chernobyl and Kanev
differed between periods, with the difference being small
in the museum samples, but considerable in the recent
ﬁeld samples (Fisher’s protected least squares difference
test, P,0.001). In contrast, there was no signiﬁcant sex
difference (Fisher’s protected least squares difference
When we compared isotopic variances for the two
isotopes for each sex between post-1986 samples from
Chernobyl and from Kanev, and between post-1986
samples from Chernobyl and pre-1986 samples from
both regions (Table 1), there were signiﬁcantly larger
variances in the recent samples from Chernobyl follow-
ing sequential Bonferroni correction (variance ratio test,
males: F.9.52, P,0.01; females: F.50.72, P,
0.001). None of the sex differences were signiﬁcant.
The main ﬁndings of this study were that (1)
correlations between the two isotopes differed signiﬁ-
cantly between regions for females, but not for males; (2)
mean values for d
N differed between regions and
periods; and (3) variances in stable-isotope proﬁles for
C differed between periods for both sexes. These
ﬁndings were consistent with our initial predictions for a
population sink when individuals of the two sexes differ
in natal dispersal distances, and this difference in
dispersal depended on the recent environmental disaster
in Chernobyl. We will brieﬂy discuss each of these
ﬁndings, and their limitations and implications.
Source–sink patterns of natural populations are a
major ﬁnding in population biology, because they have
important implications for theoretical and applied
ecology (Pulliam 1988, Dias 1996, Hanski 1999). It is
not straightforward identifying source and sink popula-
tions, and a number of different techniques have been
proposed to achieve correct identiﬁcation (Paradis 1995,
Dias 1996, Pulliam 1996, Pulliam and Danielson 1991,
Stacey et al. 1997). Here we proposed that stable-isotope
proﬁles of feathers from migratory birds may provide
information about heterogeneity in breeding popula-
tions, and such heterogeneity will reﬂect the net
emigration and immigration rates of different popula-
tions. We found evidence consistent with this scenario
based on (1) patterns of correlations between different
isotopes, (2) temporal and spatial patterns of variation
in isotope proﬁles, and (3) differences in variances in
We found evidence of weak positive correlations
between the two isotopes in three of four sex and region
categories, and one of these correlations reached
statistical signiﬁcance. There were statistically signiﬁcant
differences in the strength of these correlations for
females, but not for males, implying that the patterns of
correlation between feather d
differed between Chernobyl and Kanev for females.
The signiﬁcantly stronger correlation among females
from Kanev as compared to Chernobyl is consistent
with the prediction that females from Kanev come from
a smaller recruitment area than the females from
Chernobyl, while there is no evidence of males from
the two regions differing in the relationship between
C and d
N. How does emigration and immigration
and, hence, differences in migration and wintering areas
among individuals, affect such correlations? If a sample
of individuals with a greater dispersal distance molt their
feathers in a larger wintering area (as observed in
Palearctic Barn Swallow populations; R. Ambrosini, A.
P. Møller, and N. Saino, unpublished manuscript), then
FIG. 3. Values of d
C in relation to d
N in feathers from
adult male and female Barn Swallows trapped in 2000 from (a)
Chernobyl and (b) Kanev.
A. P. MØLLER ET AL.1702 Ecological Applications
Vol. 16, No. 5
stable-isotope variance will also be greater in that
sample. If these birds then immigrate into a breeding
population, we would predict that the isotopic variance
of the breeding population would increase. Previous
isotopic studies led to the expectation that d
N values within food webs are positively correlated
(Kelly 2000). A previous study of adult Barn Swallows
breeding in Northern Denmark did not reveal signiﬁcant
positive correlations between feather d
C and d
values (males: r¼0.10, t¼0.82, df ¼71, P¼0.42;
females: r¼0.05, t¼0.52, df ¼94, P¼0.60; Møller and
Hobson 2004). However, this absence of a positive
correlation depended on this Danish sample of breeding
Barn Swallows having a highly heterogeneous isotope
proﬁle, presumably because at least two and likely three
different populations with different winter quarters or
winter habitats bred in the same area. When we
reanalyzed this data set by excluding the small number
of individuals with extreme isotope proﬁles, which
caused bimodality in the distributions, we found a
signiﬁcant positive correlation for both sexes (males, r¼
0.36, t¼3.15, df ¼66, P¼0.003; females, r¼0.22, t¼
2.08, df ¼82, P¼0.04), as in the samples from Kanev in
Ukraine. Therefore, weak positive correlations within
‘‘isotopic populations’’ seem to be common, with the
exception of females from Chernobyl, which clearly
differed from females from Kanev and from the
homogeneous sample from Denmark. When extreme
isotope proﬁles were included in the samples from
Denmark, correlations also became weak and non-
signiﬁcant. This suggests that, when birds originating
from other populations become admixed in a homoge-
neous sample of individuals, the positive correlation
C and d
N disappears. Such decoupling of
the two isotopes has similarly been evoked as evidence
for different ultimate sources of nutrition to omnivores
(Hobson et al. 2000), but for an aerial insectivore like
the Barn Swallow, such a mechanism is not expected.
We found differences in feather d
N values between
regions and periods. Comparison of means from the
different regions, periods, and sexes revealed a difference
between Chernobyl and Kanev, between pre- and post-
1986 samples, and a larger difference between Chernob-
yl and Kanev in post-1986 than in pre-1986 samples
(Table 1). These ﬁndings are consistent with the
prediction of a larger recruitment area for Barn
Swallows breeding in the population sink in Chernobyl.
They are also consistent with the expectation that no
such effect should be visible prior to the nuclear disaster
in 1986. That we did not see similar differences in mean
C values is not necessarily surprising, because
isotopic distributions across landscapes can be driven
by different processes. For example, in addition to
climatic factors, food web d
N values appear to be
inﬂuenced more than d
C values by land-use practices
(Rubenstein and Hobson 2004). Such difference between
food web d
C and d
N values underlines the advantage
of using more than one isotope in resolving structure
within populations (e.g., Pain et al. 2004, Yohannes et
Variances in feather d
C values differed between
regions, periods, and sexes. Both males and, in
particular, females differed signiﬁcantly between Cher-
nobyl and Kanev in the post-1986 samples. Likewise,
variances were larger in post-1986 samples from
Chernobyl compared to pre-1986 samples from either
region. Some of our samples of museum specimens were
small, despite the Barn Swallows being one of the most
common breeding bird species in Ukraine. Obviously,
results based on such small samples should be inter-
preted with caution. Long-term studies of isotope
proﬁles in swallow food webs have not been conducted
at potential wintering sites in Africa. Such changes may
be signiﬁcant in areas of variable climate (Koch et al.
1995). In other areas, ﬁsh have shown constant carbon
isotope contents over decades (Begg and Weidman 2001,
Jamieson et al. 2004), and long-term stability in stable-
isotope ratios has been demonstrated in tree-ring data
spanning hundreds of years (see data and references in
McCarroll and Loader 2004), showing that temporal
change in isotope composition of food webs is not
ubiquitous. There were no differences in variances for
N values, which did show weak sex- and period-
speciﬁc variation between the two regions, but no
interaction effects. The ﬁndings for d
C values are
consistent with our a priori predictions about differences
in variances in isotope contents related to source–sink
dynamics of the Barn Swallows in Ukraine. Variances
may differ among samples for reasons other than a
higher degree of admixture of individuals from different
populations. For example, Pain et al. (2004) suggested
that differences in variances in feather d
among populations of migratory birds wintering in
Africa could be due to the degree of isotopic hetero-
geneity within molting areas, although they did not
provide any empirical evidence to support this sugges-
tion. More recently, E. Yohannes and K. A. Hobson
(unpublished manuscript) determined low isotopic (d
N, dD) variance among years for nine species of
Palearctic migrants molting at sites in sub-Saharan
Africa. We ﬁnd the explanation of isotopic heterogeneity
proposed by Pain et al. (2004) unlikely for the Barn
Swallow, given that such heterogeneity within different
molting areas should match the spatial and temporal
patterns of variation recorded. Without any evidence for
this more complex explanation, application of Occam’s
razor would suggest that the more simple explanation
based on the Chernobyl nuclear disaster is more likely to
be the cause of heterogeneity.
Our studies of museum and ﬁeld samples of isotopes
from adult Barn Swallows from Chernobyl and a
control region near Kanev in the Ukraine showed
evidence consistent with the Barn Swallow sites around
Chernobyl having become a sink following the disaster.
What are the requirements for isotopes providing
information that would allow identiﬁcation of popula-
October 2006 1703VARIATION IN ISOTOPES IN BARN SWALLOWS
tion sinks without providing false positives? Clearly, the
relative abundance of stable isotopes must vary at a
scale that allows identiﬁcation of heterogeneous sam-
ples. If different populations of breeding birds were fully
admixed during the period when feathers are molted and
had identical foraging ecology, there would be little or
no heterogeneity in isotope proﬁles among samples.
Hence, it seems important that stable isotopes or other
markers should show variation at a geographical scale
similar to that at which population processes (such as
local recruitment, emigration, and immigration) that
affect local population size are acting (Hobson 2005).
This is not trivial, since we found evidence consistent
with source–sink dynamics for variances in Barn
C, but not for variances in d
We suggest that this difference between d
C and d
values may arise from differences in spatial distribution
of stable-isotope ratios in the environment where
feathers are grown.
In conclusion, we found evidence of isotope proﬁles
and, in particular, variances in isotope proﬁles differing
between areas and sexes in a way that is consistent with
the Chernobyl region having become a sink following
the nuclear disaster. In particular, we found evidence of
sex differences in isotope proﬁles that are consistent with
known sex differences in dispersal distances, but also
with expected sex differences in dispersal between source
and sink populations. These ﬁndings emphasize that
stable-isotope proﬁles can provide powerful tools that
can be helpful in analyzing population processes, when
carefully considering temporal and spatial controls and
the natural history of the study organism.
Patricia Healy assisted with preparation of stable-isotope
samples. Analysis of carbon and nitrogen isotope samples was
conducted at the Department of Soil Science, University of
Saskatchewan, by M. Stocki. We are grateful to G. Milinevski
and E. Pysanets for logistical help and access to feather samples
in the Zoological Museum, Kiev, Ukraine. The Ukrainian
Ringing Center, EURING, and SAFRING kindly provided
information on recoveries of banded birds. T. Sze
provided the map of recoveries. A. P. Møller and T. A.
Mousseau received funding from the CNRS (France), the
University of South Carolina School of the Environment, the
Samuel Freeman Charitable Trust, the National Science
Foundation, and the National Geographic Society to conduct
this research. K. A. Hobson was supported by grants from The
Canadian Wildlife Service, Prairie and Northern Region. I. L.
Brisbin, Jr., R. McGregor, L. Thomas, and an anonymous
reviewer kindly provided constructive suggestions.
Begg, G. A., and C. R. Weidman. 2001. Stable delta C-13 and
delta O-18 isotopes in otoliths of haddock Melanogrammus
aegleﬁnus from the northwest Atlantic Ocean. Marine
Ecology – Progress Series 216:223–233.
Berthold, P. 2001. Bird migration: a general survey. Second
edition. Oxford University Press, Oxford, UK.
Bowen, G. J., L. I. Wassenaar, and K. A. Hobson. 2005.
Application of stable hydrogen and oxygen isotopes to
wildlife forensic investigations at global scales. Oecologia
Camplani, C., N. Saino, and A. P. Møller. 1999. Carotenoids,
sexual signals and immune function in barn swallows from
Chernobyl. Proceedings of the Royal Society of London B
Chamberlain, C. P., S. Bensch, X. Feng, S. A
˚kesson, and T.
Andersson. 2001. Stable isotopes examined across a migra-
tory divide in Scandinavian willow warblers (Phylloscopus
trochilus trochilus and Phylloscopus trochilus acredula) reﬂect
their African winter quarters. Proceedings of the Royal
Society of London B 267:43–48.
Chandler, C. R. 1995. Practical considerations in the use of
simultaneous inference for multiple tests. Animal Behaviour
CIA [Central Intelligence Agency]. 1996. Handbook of interna-
tional economic statistics. Directorate of Intelligence, Wash-
ington, D.C., USA.
Clobert, J., J. D. Nichols, E. Danchin, A. Dhondt, editors.
2001. Dispersal. Oxford University Press, Oxford, UK.
Cramp, S., editor. 1988. The birds of the Western Palearctic.
Volume 5. Oxford University Press, Oxford, UK.
Dias, P. C. 1996. Sources and sinks in population biology.
Trends in Ecology and Evolution 11:326–330.
Ellegren, H., G. Lindgren, C. R. Primmer, and A. P. Møller.
1997. Fitness loss and germline mutations in barn swallows
breeding in Chernobyl. Nature 389:593–596.
Glutz von Blotzheim, U. N., and K. M. Bauer, editors. 1994.
Handbuch der Vo
¨gel Mitteleuropas. Volume 9. AULA-
Verlag, Wiesbaden, Germany.
Greenwood, P. J. 1980. Mating systems, philopatry and
dispersal in birds and mammals. Animal Behaviour 28:
Hanski, I. 1999. Metapopulation ecology. Oxford University
Press, Oxford, UK.
Hershey, A. E., J. Pastor, B. J. Peterson, and G. W. Kling.
1993. Stable isotopes resolve the drift paradox for Baetis
mayﬂies in an Arctic river. Ecology 74:2315–2325.
Hobson, K. A. 1995. Reconstructing avian diets using stable-
isotope analysis of egg components—patterns of isotopic
fractionation and turnover. Condor 97:752–762.
Hobson, K. A. 1999. Tracing origins and migration of wildlife
using stable isotopes: a review. Oecologia 120:314–326.
Hobson, K. A. 2004. Flying ﬁngerprints: making connections
with stable isotopes and trace elements. Pages 235–246 in P.
Marra and R. Greenberg, editors. Birds of two worlds. Johns
Hopkins University Press, Washington, D.C., USA.
Hobson, K. A. 2005. Using stable isotopes to trace long-
distance dispersal in birds and other taxa. Diversity and
Hobson, K. A., B. N. McLellan, and J. Woods. 2000. Using
C) and nitrogen (d
N) isotopes to infer
trophic relationships among black and grizzly bears in Upper
Columbia River Basin, British Columbia. Canadian Journal
of Zoology 78:1332–1339.
Hobson, K. A., L. I. Wassenaar, and E. Bayne. 2004. Using
isotopic variance to detect long-distance dispersal and
philopatry in birds: an example with Ovenbirds and
American Redstarts. Condor 106:732–743.
Holm, S. 1979. A simple sequentially rejective multiple test
procedure. Scandinavian Journal of Statistics 6:65–70.
Jamieson, R. E., H. P. Schwarcz, and J. Brattey. 2004. Carbon
isotopic records from the otoliths of Atlantic cod (Gadus
morhua) from eastern Newfoundland. Canadian Fisheries
Kelly, J. F. 2000. Stable isotopes of carbon and nitrogen in the
study of avian and mammalian trophic ecology. Canadian
Journal of Zoology 78:1–27.
Kelly, J. F., V. Atudorei, Z. D. Sharp, and D. M. Finch. 2002.
Insights into Wilson’s warbler migration from analyses of
hydrogen stable-isotope ratios. Oecologia 130:216–221.
Koch, P. L., J. Heisinger, C. Moss, R. W. Carlson, M. L. Fogel,
and A. K. Behrensmeyer. 1995. Isotopic tracking of change in
A. P. MØLLER ET AL.1704 Ecological Applications
Vol. 16, No. 5
diet and habitat use in African elephants. Science 267:1340–
Lande, R., S. Engen, and B.-E. Sæther. 2003. Stochastic
population dynamics in ecology and conservation. Oxford
University Press, Oxford, UK.
Marra, P. P., K. A. Hobson, and R. T. Holmes. 1998. Linking
winter and summer events in a migratory bird by using
stable-carbon isotopes. Science 282:1884–1886.
McCarroll, D., and N. J. Loader. 2004. Stable isotopes in tree
rings. Quaternary Science Reviews 23:771–801.
Møller, A. P. 1993. Morphology and sexual selection in the
barn swallow Hirundo rustica in Chernobyl, Ukraine.
Proceedings of the Royal Society of London B 252:51–57.
Møller, A. P. 1994. Sexual selection and the barn swallow.
Oxford University Press, Oxford, UK.
Møller, A. P., and K. A. Hobson. 2004. Heterogeneity in stable
isotope proﬁles predicts coexistence of populations of barn
swallows Hirundo rustica differing in morphology and
reproductive performance. Proceedings of the Royal Society
of London B 271:1355–1362.
Møller, A. P., and T. A. Mousseau. 2001. Albinism and
phenotype of barn swallows Hirundo rustica from Chernobyl.
Møller, A. P., and T. A. Mousseau. 2003. Mutation and sexual
selection: a test using barn swallows from Chernobyl.
Møller, A. P., T. A. Mousseau, G. Milinevsky, A. Peklo, E.
Pysanets, and T. Sze
´p. 2005. Condition, reproduction and
survival of barn swallows from Chernobyl. Journal of
Animal Ecology 74:1102–1111.
Pain, D. J., R. E. Green, B. Giessing, A. Kozulin, A. Poluda, U.
Ottosson, M. Flade, and G. M. Hilton. 2004. Using stable
isotopes to investigate migratory connectivity of the globally
threatened aquatic warbler Acrocephalus paludicola. Oecolo-
Paradis, E. 1995. Survival, immigration and habitat quality in
the Mediterranean pine vole. Journal of Animal Ecology 64:
Pulliam, H. R. 1988. Sources, sinks, and population regulation.
American Naturalist 132:652–661.
Pulliam, H. R. 1996. Sources and sinks: empirical evidence and
population consequences. Pages 45–70 in O. E. Rhodes, Jr.,
R. K. Chester, and M. H. Smith, editors. Population
dynamics in ecological space and time. University of Chicago
Press, Chicago, Illinois, USA.
Pulliam, H. R., and B. J. Danielson. 1991. Sources, sinks and
habitat selection: a landscape perspective on population
dynamics. American Naturalist 137:50–66.
Roff, D. A. 2001. Life history evolution. Sinauer Associates,
Sunderland, Massachusetts, USA.
Rubenstein, D. R., C. P. Chamberlain, R. T. Holmes, M. P.
Ayres, J. R. Waldbauer, G. R. Graves, and N. C. Tuross.
2002. Linking breeding and wintering ranges of a migratory
songbird using stable isotopes. Science 295:1062–1065.
Rubenstein, D. R., and K. A. Hobson. 2004. From birds to
butterﬂies: animal movement patterns and stable isotopes.
Trends in Ecology and Evolution 19:256–263.
SAS Institute. 2000. JMP. SAS Institute, Cary, North Carolina,
Shestopalov, V. M. 1996. Atlas of Chernobyl exclusion zone.
Ukrainian Academy of Science, Kiev, Ukraine.
Sillett, T. S., and R. Holmes. 2002. Variation in survivorship of
a migratory songbird throughout its annual cycle. Journal of
Animal Ecology 71:296–308.
Sokal, R. R., and F. J. Rohlf. 1995. Biometry. Third edition.
Freeman, New York, New York, USA.
Stacey, P. B., V. A. Johnson, and M. L. Taper. 1997. Migration
within metapopulations: the impact upon local population
dynamics. Pages 267–292 in I. A. Hanski and M. E. Gilpin,
editors. Metapopulation biology. Academic Press, San Die-
go, California, USA.
Still, C. J., J. A. Berry, G. J. Collatz, and R. S. DeFries. 2003.
The global distribution of C
vegetation: carbon cycle
implications. Global Biogeochemical Cycles 17(1):1006.
Urton, E. J. M., and K. A. Hobson. 2005. Intrapopulation
variation in gray wolf isotope (d
N and d
implications for the ecology of individuals. Oecologia 14:
Wright,S.P.1992.AdjustedP-values for simultaneous
inference. Biometrics 48:1005–1013.
Yohannes, E., K. A. Hobson, D. Pearson, L. I. Wassenaar, and
H. Biebach. 2005. Stable isotope analyses of feathers help
identify autumn stopover sites of three long-distance
migrants in northeastern Africa. Journal of Avian Biology
October 2006 1705VARIATION IN ISOTOPES IN BARN SWALLOWS