Biological Journal of the Linnean Society, 2011, 102, 583–592. With 2 figures

Panmixia on a continental scale in a widely distributed
colonial waterbird
MATTHEW W. REUDINK1,2*, CHRISTOPHER J. KYLE1, JOSEPH J. NOCERA1,2,
REBEKAH A. OOMEN1, M. CLAY GREEN3 and CHRISTOPHER M. SOMERS4
1

Forensic Science Department, Trent University, Peterborough, ON K9J 7B8, Canada
Ontario Ministry of Natural Resources, Trent University, DNA Building, Peterborough, ON K9J
7B8, Canada
3
Department of Biology, Texas State University, San Marcos, TX 78666, USA
4
Department of Biology, University of Regina, Regina, SK S4S 0A2, Canada
2

Received 24 July 2010; accepted for publication 1 October 2010

bij_1608

583..592

Many highly mobile species, such as migratory birds, can move and disperse over long distances, yet exhibit high
levels of population genetic structuring. Although movement capabilities may enable dispersal, gene flow may be
restricted by behavioural constraints such as philopatry. In the present study, we examined patterns of genetic
differentiation across the range of a highly mobile, colonial waterbird. American white pelicans (Pelecanus
erythrorhynchos) breed across continental North America and are currently experiencing a range expansion,
especially on the eastern range limit. To assess patterns of genetic structuring, we sampled 333 individuals from
19 colonies across their North American range. The use of ten variable microsatellite markers revealed high levels
of allelic richness with no population differentiation. Both Bayesian and frequentist approaches to examining
genetic structuring revealed a single panmictic population. We found no evidence of genetic structuring across the
Continental Divide or between migratory and non-migratory colonies. The lack of any genetic structure across the
range indicates that, unlike other waterbirds with similar life-history characteristics, extensive gene flow and
presumably low philopatry appear to preclude genetic differentiation. The lack of population genetic structure in
American white pelicans provides an example of range-wide panmixia, a rare phenomenon in any terrestrial
species. © 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592.

ADDITIONAL KEYWORDS: American white pelican – gene flow – genetic structure – microsatellite –
panmixia – Pelecanus erythrorhynchos.

INTRODUCTION
High mobility may enable long-distance dispersal
(Ward, Skibinski & Woodwark, 1992), which enhances
gene flow and limits population differentiation.
However, movement capabilities alone may not
predict patterns of genetic structure (Friesen, Burg &
McCoy, 2007). Many pelagic seabirds, such as
common murres (Uria aalge Pontoppidan, 1763;
Morris-Pocock et al., 2008), razorbills (Alca torda

*Corresponding author. Current address: Department of
Biological Sciences, Thompson Rivers University, Kamloops,
BC V2C 5N3, Canada. E-mail: mattreudink@gmail.com

Linnaeus, 1758; Moum & Árnason, 2001), red-legged
kittiwakes (Rissa brevirostris Bruch, 1853; Patirana,
Hatch & Friesen, 2002), and marbled murrelets
(Brachyamphus marmoratus Gmelin, 1789; Friesen
et al., 2005) can disperse over thousands of kilometers, yet exhibit population genetic structuring within
ocean basins. This pattern of genetic structuring in
species capable of long-distance dispersal may be
driven by multiple mechanisms, including restricted
gene flow as a result of high natal philopatry (Friesen
et al., 2007), cryptic barriers to dispersal (Zardi et al.,
2007), and behavioural mechanisms (Friesen et al.,
2007). In addition, local adaptation to differing ecological conditions and strong selective pressures

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

583

584

M. W. REUDINK ET AL.

(Slatkin, 1987; Freeman-Gallant, 1996) may promote
geographic patterns of differentiation. For example,
Morris-Pocock et al. (2008) found that common
murres exhibit low population genetic structuring
within the Pacific Ocean but significant east–west
structuring in the Atlantic Ocean. The low levels of
differentiation in the Pacific were attributed to high
levels of contemporary gene flow, whereas patterns in
the Atlantic appear to be the result of historic population fragmentation (Morris-Pocock et al., 2008).
Species with movement capacities and life history
characteristics similar to pelagic seabirds (e.g. colonial, K-selected) also inhabit terrestrial environments
in continental interiors, although studies of patterns
of genetic structuring over large spatial scales (e.g.
continents) are limited. Goostrey et al. (1998) and
Marion & Le Gentil (2006) detected significant
genetic differentiation and patterns of genetic structure in great cormorants (Phalocrocorax carbo Linnaeus, 1758) across Europe. Yauk & Quinn (1999)
found genetic homogeneity within Great Lakes populations of herring gulls (Larus smithsonianus Coues,
1862) but significant genetic divergence between the
Great Lakes and eastern Canada. Although some
studies have detected genetic homogeneity within
part of a species range (Bates, Deyoung & Ballard,
2009), range-wide studies on continental waterbirds
remain rare.
American white pelicans (Pelecanus erythrorhynchos Gmelin, 1789; hereafter pelicans) are highly
mobile waterbirds that breed colonially at inland sites
across a large portion of North America from the
Pacific Coast to the Great Lakes (Knopf & Evans,
2004; Pekarik et al., 2009). Colonies follow a metapopulation model, ranging in size from a few individuals
to more than 30 000, and experience frequent local
extinction and recolonization events (Knopf & Evans,
2004; Anderson & King, 2005). Metapopulation
theory assumes that local extinctions result in a loss
of alleles from the extinct population, thereby facilitating divergence among groups. However, local
extinctions in pelicans may instead reflect colony
abandonment, resulting in a ‘reshuffling’ of alleles via
dispersal post-abandonment (Wade & McCauley,
1988). Such a situation would enhance gene flow
rather than promote differentiation, as is generally
predicted by metapopulation theory (Hanski, 1999).
Recent demographic history demonstrates a marked
rebound in overall pelican abundance, from the 1933
estimate of only 30 000 adults (Thompson, 1933;
Keith, 2005) to current estimates of approximately
134 000 breeding adults in North America (King &
Anderson, 2005). As a part of this resurgence, pelicans have undergone a range expansion, including
recent breeding attempts near Akimiski Island
(James Bay, Nunavut), 500 km from the previous

easternmost breeding population (Pekarik et al.,
2009). Pelicans forage over hundreds of kilometers
(Knopf & Evans, 2004) and can disperse long distances; yet little is known about patterns of natal or
breeding philopatry, and ultimately genetic population structure. Pelicans are currently grouped into
eastern and western metapopulations separated by
the Continental Divide (Anderson & King, 2005),
namely because of differential patterns of exchange
between local and distant colonies (Anderson &
Anderson, 2005), as well as frequent local extinction
and recolonization events. However, band return data
indicate some degree of movement across this potential barrier (Anderson & Anderson, 2005). Whether
this movement is sufficient to ameliorate possible
isolating effects from the Continental Divide remains
unknown. However, if patterns of recapture reflect
patterns of breeding dispersal, pelicans may far
exceed the one-to-ten-migrant-per-generation rule,
leading to genetic homogeneity (Mills & Allendorf,
1996).
The vast majority of pelican colonies are located in
the northern USA and Canada where breeding adults
follow a typical migration schedule, arriving at breeding colonies in early spring and departing in the
autumn. Populations separated by the Continental
Divide generally follow separate migratory pathways,
overwintering in coastal Western Mexico and the Gulf
Coast of Mexico, respectively (Knopf & Evans, 2004).
However, a few small, non-migratory colonies of pelicans exist in Mexico and coastal Texas, specifically
the Laguna Madre that extends from Corpus Christi,
Texas, USA, to La Pesca, Tamaulipas, Mexico. In the
Laguna Madre of Texas, the abundance of breeding
American white pelicans has remained relatively constant at 200–500 nests since the early 20th Century
(Chapman, 1988). In the Laguna Madre de Tamaulipas, unconfirmed reports of nesting pelicans have
occurred ever since the 1920s but were not documented until the 1960s (Selander et al., 1962). In
2008, aerial surveys of all breeding colonies in the
Laguna Madre de Tamaulipas yielded no confirmed
nesting pelicans, although a single foraging flock of
approximately 420 individuals and roosting flock of
82 individuals were photographed (C. Green, unpubl.
data.). The Texas Gulf Coast, including the Laguna
Madre, and Laguna Madre de Tamaulipas also
contain increasing numbers of wintering pelicans
(National Audubon Society, 2002; Knopf & Evans,
2004). The relationship between overwintering birds
and breeding colonies in the Laguna Madre of Texas
and Tamaulipas is unclear. Although the numbers
have remained relatively stable, suggesting some
natal site fidelity, limited banding data suggest the
potential for birds fledged from colonies in the north
to breed in the Laguna Madre colonies (Chapman &

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

PANMIXIA IN PELICANS
Chapman, 1990). Additionally, band recovery data
provide evidence for birds fledged from northern
colonies to overwinter in coastal Texas (Strait &
Sloan, 1975). If interbreeding between southern and
northern colonies is limited, we expect that these
differences in migratory behaviour will limit gene
flow and promote differentiation.
In the present study, we examined variation at ten
microsatellite loci in pelicans sampled from breeding
colonies spanning most of their North American
range. Our goals were to examine: (1) population
structure across the continent; (2) patterns of differentiation between putatively non-migratory southern
colonies and the rest of the migratory northern breeding range; and (3) whether the current classification
of pelicans into eastern and western metapopulations
is supported by variation in neutral genetic markers.

MATERIAL AND METHODS
FIELD METHODS
We collected tissue samples from 333 individuals at
19 colonies. Samples from most colonies were collected opportunistically from natural mortalities of
nestling pelicans by volunteers who were either
banding flightless young pelicans or surveying for
disease outbreaks in colonies. Muscle or skin tissue
(0.5–1.0 g) from recently dead nestlings was collected
using a sterile razor blade and stored in lysis buffer
(4.0 M urea, 0.2 M NaCl, 0.1 M Tris HCl, pH 8.0, 0.5%
n-laurylsarcosine, 0.1 M 1,2-cyclohexanediamine). We
took a small blood sample (< 500 mL) from the tarsalmetatarsal or brachial vein of flightless nestlings
from colonies in Saskatchewan, Canada, when they
were 4–10 weeks old. Similar methods were used to
collect blood samples from nestlings in California
colonies in the late 1980s. We extracted DNA from
feathers moulted on the breeding grounds from the
remaining individuals (Boles Island, Ontario,
Canada). The location of study colonies and number of
individuals sampled is presented in Table 1 and
Figure 1.

LABORATORY METHODS
We extracted DNA from blood and tissue samples
using a Qiagen DNAeasy Tissue Extraction kit. We
amplified ten polymorphic microsatellites isolated
from American white pelicans (PeEr loci; Hickman
et al., 2008) and great white pelicans (Pelecanus onocrotalus Linnaeus, 1758; Pel loci; de Ponte Machado
et al., 2009) (Table 2). We created four polymerase
chain reaction (PCR) reaction multiplexes: locus
PeEr01, PeEr02, PeEr03, and PeEr09 (Multiplex A);
locus PeEr04 and PeEr07 (Multiplex B); locus Pel149
and Pel304 (Multiplex C); and locus PeEr06, and

585

Pel086 (Multiplex D). We amplified multiplexes A and
C in a 10 mL reaction volume under the conditions:
94 °C for 5 min followed by 30 cycles of 94 °C for 30 s,
59 °C for 60 s, and 72 °C for 60 s followed by a final
extension of 60 °C for 45 min. We amplified multiplexes B and D under the conditions: 94 °C for 5 min
followed by 30 cycles of 94 °C for 30 s, 62 °C for 60 s,
and 72 °C for 60 °C s, followed by a final extension of
60 °C for 45 min. We visualized DNA fragments using
an ABI 3700 DNA analyzer (PE Applied Biosystems
Inc.) and analyzed the fragments using GeneScan
Analysis software (PE Applied Biosystems Inc.). PCR
multiplexes A and B were visualized together, as were
PCR multiplexes C and D.

MICROSATELLITE DATA ANALYSIS
We tested for departures from Hardy–Weinberg equilibrium (HWE) for each of the ten loci from 333
individuals across 19 sampling locations using the
exact probability test with Markov chain parameters
set to 100 batches with 1000 iterations in GENEPOP,
version 4.0.1 (Raymond & Rousset, 1995). We also
used GENEPOP to examine genotypic linkage disequilibria (LD) among loci, calculate pairwise FST
values between the eastern, western, and southern
groups, and determine observed and expected heterozygosities at each sampled region. In addition, we
examined whether population genetic differences
could be explained by isolation by distance by conducting Mantel tests with 9999 randomizations
between pairwise FST and geographical distances in R,
version 2.11.0 (R Development Core Team, 2010). We
examined allelic richness (RS) and gene diversity
using FSTAT, version 2.9.3 (Goudet, 1995). For RS, we
used the rarefaction method (El Mousadik & Petit,
1996), which employs resampling of the genotype
data to produce sample sizes equal to the smallest
population. Because some populations were represented by small sample sizes, we excluded all populations (N = 5) with fewer than ten individuals. We
tested for differences in allelic richness, gene diversity, and observed heterozygosity among populations
using one-way analysis of variance (ANOVA). We
examined evidence for scoring errors, allelic dropout,
and possible null alleles using MICRO-CHECKER
(van Oosterhout et al., 2004) and examined error
rates in allele scoring by re-analyzing and re-scoring
alleles for 61 of 333 individuals.
We used two different individual Bayesian clustering methods to examine genetic population structure:
STRUCTURE, version 2.3 (Pritchard, Stephens &
Donnelly, 2000; Falush, Stephens & Pritchard, 2003)
and TESS, version 2.3.1 (Francois, Ancelet & Guillot,
2006; Chen et al., 2007). Algorithms within these
programs seek genetic structure from multi-locus

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

Region

West
West
West
West
West

East
East
East
East
East
East
East
East
East
East
East
East
East

South

Site

CA1
CA2
ID1
ID2
NV1

AB1
AB2
MB1
MN1
MT1
ND1
ON1
SD1
SD2
SK1
SK2
SK3
WI1

TX1

Clear Lake, CA
Clear Lake, CA (historic)
Blackfoot Reservation, ID
Minidoka NWR, ID
Anaho Island, NV
All western populations
Pelican Lake, AB
Portage Lake, AB
Pipestone Rocks, MB
Marsh Lake, MN
Medicine Lake, MT
Chase Lake, ND
Boles Island, ON
Bitter Lake, SD
LaCreek NWR, SD
Dore Lake, SK
Last Mountain Lake, SK
Reed Lake, SK
Cat Island, WI
All eastern populations
Padre Islands, TX
All southern populations

Location

2007

2007
2007
2009
2006–7
2007
2007
2008
2007
2007
2006
2006
2006
2006

2007
1980s
2007
2007
2007

Year

27.47995

55.78816
58.94363
51.38485
45.18772
48.44924
47.00490
49.87030
45.27199
43.12049
54.72759
51.36396
50.38707
47.01333

41.86419
41.86419
43.17921
42.66780
39.95344

East (°)

Table 1. Sample, location, and genetic diversity information for all sampled regions

14
33
10
6
34
97
12
19
9
26
22
13
7
7
35
18
24
24
1
217
19
19

-121.14487
-121.14487
-112.28032
-113.36792
-119.51202

-97.31483

-113.24776
-113.24913
-96.55060
-96.13020
-104.38248
-99.43528
-88.93913
-97.33938
-101.53921
-107.50980
-105.23975
-107.03568
-90.55917

N

North (°)
0.72
0.70
0.74
0.69
0.72
0.71
0.72
0.69
0.70
0.71
0.74
0.74
0.76
0.71
0.68
0.72
0.69
0.70
–
0.71
0.70
0.70

HO

0.65
0.70
0.68
0.77
0.72
0.71
0.70
0.73
0.64
0.74
0.67
0.72
0.75
0.65
0.68
0.70
0.68
0.65
–
0.69
0.75
0.75

HE

5.62
5.09
5.74
–
5.44
5.47
5.39
5.01
–
5.29
5.28
5.27
–
–
5.02
5.25
5.11
5.11
–
5.19
4.96
4.96

Arich

0.73
0.70
0.74
–
0.72
0.72
0.72
0.69
–
0.71
0.74
0.74
–
–
0.68
0.72
0.69
0.70
–
0.71
0.70
0.70

Gene
diversity

586
M. W. REUDINK ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

PANMIXIA IN PELICANS

587

Figure 1. Map of all sampled colonies with the North American Continental Divide (black line), approximate main
breeding range (light grey; Ridgely et al., 2007), nonbreeding range (medium grey), and year-round range (dark grey),
putative eastern (east of the continental divide), western (west of the continental divide), and southern (Texas)
populations. STRUCTURE results indicated that K = 1; however, our a priori prediction was that we would observe three
genetically distinct, eastern, western and southern subunits. To illustrate the lack of support for multiple clusters, we
have included bar charts alongside the putative regional populations representing probability of membership for each
individual in one of three genetic clusters (K = 3) as calculated by the Bayesian clustering algorithm in STRUCTURE.
Each stacked bar represents one individual with probability of membership into each cluster (bottom black bars = cluster
1; central light grey bars = cluster 2; top dark grey bars = cluster 3).

Table 2. Allelic information for the ten loci used in the
present study, isolated from American white pelicans
(PeEr01–PeEr09; Hickman et al., 2008) and great white
pelicans (Pel086–Pel304; de Ponte Machado et al., 2009)
Locus

Alleles

Size range

N

HO

HE

PeEr01
PeEr02
PeEr03
PeEr04
PeEr06
PeEr07
PeEr09
Pel086
Pel149
Pel304

8
10
11
11
12
4
13
7
19
7

139–171
241–268
227–263
157–201
186–214
223–235
136–182
150–174
261–341
186–204

332
328
325
288
330
319
321
323
316
321

0.67
0.80
0.81
0.81
0.76
0.27
0.83
0.77
0.83
0.42

0.67
0.79
0.80
0.82
0.77
0.29
0.86
0.76
0.92
0.42

genotypes without assuming predefined populations.
STRUCTURE (Pritchard et al., 2000; Falush et al.,
2003; Hubisz et al., 2009) estimates the probability of
assignment to a given number of clusters (K) and
estimates the membership of each individual (q) to

each cluster. In addition, STRUCTURE (Pritchard
et al., 2000; Falush et al., 2003; Hubisz et al., 2009)
uses a Markov chain Monte Carlo (MCMC) procedure
that maximizes HWE and minimizes LD. Optimal
K-clusters for the data set were based on a rate of
change in the log probability of the data (DK) within
each assumed K (Evanno, Regnaut & Goudet, 2005).
We performed clustering under the F-model as proposed by Falush et al. (2003), which assumes admixture and correlated allele frequencies, and also using
the inference of a for each subpopulation. Each run
was set to a burn-in of 500 000 and a MCMC data
collection chain of 1 000 000. We tested K = 1–19 replicated five times using LINUX servers.
We also used the clustering software TESS, version
2.3.1 (Francois et al., 2006; Chen et al., 2007). This
Bayesian clustering program uses a MCMC approach
to define genetic clusters under the assumptions of
HWE to reduce inbreeding coefficients. This algorithm also uses a hidden Markov random field
(HMRF) model on tessellations, allowing spatial relationships to nearest-neighbors to be defined, at the
same time as assuming that spatially proximate individuals are genetically similar. Incremental increases

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

588

M. W. REUDINK ET AL.

of the HMRF distance separating population centres
value (d) from 0 (full cline) to 1 (hard clusters)
permits the user to visualize genetic clustering that
may be overestimated or diminished using other
Bayesian clustering programs such as STRUCTURE
and BAPS, version 4.1 (Bayesian analysis of population structure; Corander & Marttinen, 2006). These
other approaches can have difficulties differentiating
genetic clusters along a continuous cline in genetic
variation, and therefore TESS analysis may provide
lower error rates than alternative methods when
lower levels of genetic structure are observed (Francois et al., 2006; Chen et al., 2007).
We ran TESS for 500 000 sweeps with an initial
burn-in of 50 000 sweeps testing the maximal number
of clusters from 1–10, with five runs per maximum
cluster. Next, using three different interaction parameter settings of 0, 0.5, and 0.99, we ran five replicates
at K = 2 with an initial burn-in of 50 000 sweeps
followed by 500 000 sweeps.

SAMPLING REGIONS, GENETIC CLUSTERS, AND
ASSIGNMENT TESTS

To determine the optimal number of genetic clusters,
we examined DK (Evanno et al., 2005), as well as the
geographic distribution of individuals with high coefficients of relationship. We did so under two thresholds (q > 0.80 and q > 0.50) from both STRUCTURE
and TESS to determine whether the majority of
individuals with high-ancestry coefficients could be
associated with a specific geographic region. Once
determined, all individuals collected within this
region were used to define groups for frequency-based
individual assignment (Paetkau et al., 1995, 2004)
implemented in GENECLASS2 (Piry et al., 2004). We
compared the results from these approaches to those
generated from Bayesian clustering analyses to
provide confidence in the spatial designation of
genetic groupings. Also, we used this comparison to
determine whether these individuals could be considered migrants among geographic areas or simply
cross-assigned by chance. We set an assignment
threshold of P < 0.01 and a default frequency for
missing alleles at 0.01 for all frequency-based assignment tests within GENECLASS2.

not observed for these loci in several groups, it is
unlikely that the loci are physically linked; thus, the
loci were retained for subsequent analyses. MICROCHECKER revealed no evidence of scoring errors
allelic dropout, nor null alleles in any loci.

GENETIC DIVERSITY
We found no differences in allelic richness (ANOVA:
F13, 139 = 0.12, P = 0.999) or gene diversity (ANOVA:
F13, 139 = 0.10, P = 1.00) across the 14 populations with
N > 10 individuals (Table 1). In addition, we found no
significant heterozygote excesses or deficiencies and
observed heterozygosities did not vary among the 18
sampled regions with N > 1 individual (ANOVA:
F17,179 = 0.30, P = 0.997).

GENETIC CLUSTERING AND ASSIGNMENT TESTS
FST values for each population pair were low
(East–West: FST = 0.0016, P = 0.13; East–South:
FST = -0.00123, P = 0.86; South–West: FST = -0.0029,
P = 0.61) and we found no evidence of isolation by
distance (r = 0.04, P = 0.42). The most likely number
of clusters revealed by STRUCTURE was K = 1, with
higher values of K returning lower likelihoods
(Fig. 2). Although STRUCTURE indicated K = 1, we
were interested in further testing our a priori
assumption that eastern and western metapopulations are genetically distinct sub-units. Thus, we

RESULTS
MICROSATELLITES
None of the ten loci used in the present study deviated from HWE in any of the 19 sampling locations.
Genotypic disequilibrium was observed at two colonies: Anaho Island, Nevada, USA (PeEr03/Pel149),
and Last Mountain Lake, Saskatchewan, Canada
(Pel149/086). Because genotypic disequilibrium was

Figure 2. A, plot of the mean LnP(D) values ± 1 SD for
K = 1 through K = 10 as calculated by STRUCTURE 2.3.
LnP(D) is an estimate of the posterior probability for the
data for each K. B, plot of DK versus K.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

PANMIXIA IN PELICANS
examined inferred population assignments from K = 2
(averaged over five iterations). In total, 321 of the 333
individuals had higher likelihoods of belonging to
cluster 1 (q > 0.5); however, the highest likelihood for
cross-assignment to cluster 2 was 0.68. Assignments
of individuals using GENECLASS2 at the broadest
geographic region (K = 3; East, West, South) were
weak. Only 34.8% (116/333) of individuals were
assigned to their original sampling location and
< 10% of the samples (31/333) had a q > 0.8, suggesting a broad exchange of genes across the sampling
region. Only 16% (5/31) of individuals with q > 0.8
were assigned to their original sampling locations.
Increasing the number of populations did not improve
assignment rates.
Results from TESS Bayesian clustering analyses
indicated a highest likelihood of K = 1. When we
examined assignment probabilities from our analyses
of K = 2 with our interaction parameter set at 0 and
0.5, only 6/333 (0.018%) had assignment probabilities
to cluster 1 higher than 0.8 and no individuals
assigned to cluster 2 with assignment probabilities
higher than 0.8. When our interaction parameter was
increased to 0.99, 4/333 (0.012%) individuals had
assignments probabilities to cluster 1 lower than 0.8
and 3/333 (0.009%) individuals assigned to cluster 2
with assignment probabilities higher than 0.8.

DISCUSSION
Bayesian and frequentist approaches to analyzing
population assignments of 333 American white pelicans at ten microsatellite loci revealed a complete
lack of genetic structure across their North American
range. Despite the potential for genetic structuring
based on: (1) the continental divide limiting dispersal
between eastern and western individuals; (2) natal/
breeding philopatry common among many waterbirds; and (3) the occurrence of resident breeding
populations in Texas, we found no evidence for genetic
differentiation at any level. Although additional
markers, such as mitochondrial (mt)DNA, and an
expanded sampling scheme are necessary to fully
verify complete panmixia, the results obtained in the
present study strongly suggest a complete lack of
genetic structure in this widely distributed, colonial
waterbird.
A complete lack of genetic differentiation across the
range of a species is an uncommon phenomenon. In
those few species in which range-wide panmixia has
been observed [e.g. rasp coral (Pocillopora verrucosa
Ellis and Solander, 1796) Ridgway, Hoegh-Guldberg
& Ayre, 2001; European plaice (Pleuronectes platessa
Linnaeus, 1758) Hoarau et al., 2002; Dawson’s burrowing bee (Amegilla dawsoni Rayment, 1951) Beveridge & Simmons, 2006; hooded seal (Cystophora

589

cristata Erxleben, 1777) Coltman et al., 2007; banded
goby (Caffrogobius caffer Gunther, 1874) Neethling
et al., 2008], genetic homogeneity across the species
range has been generally attributed to high mobility
and dispersal. Although the high mobility of pelicans
appears to enable long-distance dispersal, movement
capabilities alone may not predict patterns of genetic
structuring (Friesen et al., 2007).
Many highly mobile species are capable of longdistance movements, yet strong natal philopatry and
barriers to dispersal (both apparent and cryptic) often
limit gene flow and maintain at least some degree of
genetic structuring (Friesen et al., 2007). Although
historic and contemporary barriers such as land or ice
are the best predictors of population genetic structuring in seabirds, many species still exhibit structuring
in the absence of physical barriers (Friesen et al.,
2007), which may be a result of behavioural differences among groups. For example, black-browed albatrosses (Thalassarche melanophrys Temminck, 1828)
with different foraging grounds differ genetically
(Burg & Croxall, 2001) and population-specific nonbreeding areas often predict the presence of phylogeographic structure (Friesen et al., 2007). Although a
small number of seabirds lack genetic differentiation
(e.g. grey-headed albatross (T. chrysostoma Forster,
1785); Burg & Croxall, 2001), detection of at least
some degree of genetic structuring in part of the
range appears to be common, regardless of whether
apparent barriers to dispersal exist (Friesen et al.,
2007).
Although the Continental Divide has been the
basis for defining separate eastern and western
metapopulations in pelicans (Anderson & King,
2005), the data obtained in the present study
suggest that, despite what may be an important
demographic barrier, gene flow is high enough
across the entire range to preclude differentiation.
Many terrestrial bird species are genetically differentiated across the North American Continental
Divide as a result of historic processes such as occupying eastern and western glacial refugia (Lovette,
Clegg & Smith, 2004; Peters, Gretes & Omland,
2005) or contemporary processes, such as the Great
Plains and Rocky Mountains acting as major barriers to dispersal. The current classification of pelicans into eastern and western metapopulations
separated by the Continental Divide (Anderson &
King, 2005) is supported by band return data from
1922–1981, which indicated significant demographic
isolation between groups (Anderson & Anderson,
2005). A conservative approach of including only
returns of birds > 4 years old revealed that 24/360
(3.7%) of nestlings banded east of 110° longitude
were subsequently recovered west of 110° longitude,
whereas 31/292 (10.6%) banded in the west were

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

590

M. W. REUDINK ET AL.

recovered east of 110° longitude (Anderson & Anderson, 2005). On the basis of the lack of genetic structure observed in the present study, current levels
of gene flow across the Continental Divide appear
to be high enough to preclude genetic isolation.
In addition, although demographically pelicans do
appear to fit a metapopulation model, local extinctions and recolonizations may enhance gene flow
through frequent ‘reshuffling’ of alleles, rather than
leading to differentiation (Wade & McCauley, 1988).
However, the use of additional markers, such as
mtDNA, may reveal patterns of differentiation
where nuclear markers do not (e.g. Mylecraine et al.,
2008), and discriminate between contemporary and
historic patterns of gene flow and the potential for
sex-biased population structure (Prugnolle & de
Meeus, 2002).
We predicted that southern populations would
exhibit significant genetic differentiation from the
northern populations; however, we observed no
genetic differentiation of birds nesting in the Padre
Island National Seashore, Texas. Breeding pelicans
have been recorded in Texas and northern Mexico
since at least 1907 (Oberholser & Kincaid, 1974) and
have likely been present for centuries (Chapman,
1988), suggesting that recent colonization/expansion
is unlikely to account for the lack of genetic structure.
One potential explanation for the lack of observed
structure is that the southern breeding colonies are
composed of both local breeders and migratory birds
from elsewhere in the range that remained in the
south rather than returning north to their natal
region. After fledging, pelicans from northern breeding colonies migrate south to wintering grounds in
the southern USA and coastal Mexico (Strait & Sloan,
1975; Knopf & Evans, 2004). In double-crested cormorant (Phalacrocorax auritus Lesson, 1831) breeding colonies in central Florida, little genetic
differentiation was found between resident birds and
migratory populations from the Great Lakes and
Atlantic coast that overwinter in central Florida
(Green et al., 2006). The results of the present study
suggest that mixing of resident and migratory pelicans may be occurring in over-wintering areas such as
the Laguna Madre of Texas, thereby limiting any
genetic differentiation between southern and northern colonies. Overall, the provenance of breeders at
the Texas colony remains unknown; future studies
employing banding, geolocators, satellite transmitters, and/or biogeochemical markers are needed to
better understand the relationship between resident
and migratory birds in Texas. Biogeochemical
markers (e.g. stable isotopes) may indicate a bird’s
geographical origin where microsatellites do not;
however, because pelicans do not breed until ⱖ 3
years of age, samples that can be obtained non-

invasively (e.g. feathers) will only provide an estimate
of the last moult location and not natal origin. Tissues
with longer turnover rates, such as bone, may not
only provide an indication of natal origin, but also
require the destructive sampling of breeders.
Unlike many other colonial waterbirds, pelicans
may not be constrained by philopatry and their
capability for long-distance movement likely facilitates frequent long-distance dispersal. High dispersal in American white pelicans, compared to
other colonial waterbirds, may be driven by erratic
fluctuations in colony size driven by water-level
fluctuations (Moreno-Matiella & Anderson, 2005),
disease (Rocke et al., 2005) or other disturbances
that force dispersal and lead to frequent remixing of
individuals at breeding colonies. Thus, we suggest
that the panmictic genetic structure of pelicans is a
result of high dispersal and erratic local colony
dynamics leading to movement and genetic
exchange. Because it would be unwieldy to manage
pelican populations at the true scale of their genetic
metapopulation, we suggest that classification of
American white pelicans into separate groups based
on geography may be practical for distinguishing
and managing pelican populations, as long as managers are aware that such groupings will not reflect
genetically distinct groups. Thus, rather than management plans aimed at protecting genetically distinct populations, plans should focus instead on
regional numbers and demographic patterns, especially at the level of colonies that have significant
local ecological and social value.

ACKNOWLEDGEMENTS
We thank M. Sovada, K. Tribby, S. ComeauKingfisher, A. McGregor, B. Madden, D. Mauser, D.
Withers, M. Wackenhut, J. DiMatteo, J. Laux, D.
Anderson, D. Wilson, J. Neill, M. Ball, S. Lockhart, N.
Gorman, L. Downing, and numerous field assistants
for assisting with the collection of pelican tissues. We
thank R. Cobb, Z. Holderby, M. Biel, D. Newstead,
and W. Howe for facilitating collection of samples at
the Padre Island National Seashore. J. MorrisPocock, T. Wheeldon, R. Gillette, and R. Reudink
provided helpful discussion, comments, and data
interpretation. We thank R. Reudink for creating
Figure 1. Funding for this project was provided by
the Climate Change Fund for Ontario (J.J.N.), the
Explorer’s Club (M.W.R.), MITACS Accelerate and
Elevate programs (M.W.R.), the Canada Research
Chairs Program, the Natural Sciences and Engineering Research Council of Canada, and the
Saskatchewan Ministry of the Environment’s Fish
and Wildlife Development Fund (C.M.S.).

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

PANMIXIA IN PELICANS

REFERENCES
Anderson JGT, Anderson KB. 2005. An analysis of band
returns of the American white pelican, 1922 to 1981. Waterbirds 28: 55–60.
Anderson DW, King T. 2005. Introduction: biology and conservation of the American white pelican. Waterbirds 28: 1–8.
Bates ME, Deyoung RW, Ballard BM. 2009. Genetic diversity and population structure of reddish egrets along the
Texas coast. Waterbirds 32: 430–436.
Beveridge M, Simmons LW. 2006. Panmixia: an example
from Dawson’s burrowing bee (Amegilla dawsoni) (Hymenoptera: Anthophorini). Molecular Ecology 15: 951–957.
Burg TM, Croxall JP. 2001. Global relationships amongst
black-browed and grey-headed albatrosses: analysis of population structure using mitochondrial DNA and microsatellites. Molecular Ecology 10: 2647–2660.
Chapman BR. 1988. History of the white pelican colonies in
south Texas and northern Tamaulipas. Colonial Waterbirds
11: 275–283.
Chapman BR, Chapman SS. 1990. Patagial tag causes
white pelican death. North American Bird Bander 15: 17.
Chen C, Durand E, Forbes F, Francois O. 2007. Bayesian
clustering algorithms ascertaining spatial population structure: a new computer program and a comparison study.
Molecular Ecology Notes 7: 747–756.
Coltman DW, Stenson G, Hammill MO, Haug T, Davis
CS, Fulton TL. 2007. Panmictic population structure in
the hooded seal (Cystophora cristata). Molecular Ecology 16:
1639–1648.
Corander J, Marttinen P. 2006. Bayesian identification of
admixture events using multilocus molecular markers.
Molecular Ecology 15: 2833–2843.
El Mousadik A, Petit RJ. 1996. Chloroplast DNA phylogeography of the argan tree of Morocco. Molecular Ecology 5:
547–555.
Evanno G, Regnaut S, Goudet J. 2005. Detecting the
number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14: 2611–2620.
Falush D, Stephens M, Pritchard JK. 2003. Inference of
population structure using multilocus genotype data: linked
loci and correlated allele frequencies. Genetics 164: 1567–
1587.
Francois O, Ancelet S, Guillot G. 2006. Bayesian clustering using hidden Markov random fields in spatial population genetics. Genetics 174: 805–816.
Freeman-Gallant CR. 1996. Microgeographic patterns of
genetic and morphological variation in savannah sparrows
(Passerculus sandwichensis). Evolution 50: 1631–1637.
Friesen VL, Birt TP, Piatt JF, Golightly RT, Newman
SH, Hébert PN, Gissing G. 2005. Population genetic
structure in marbled murrelets (Brachyramphus marmoratus), and the delineation of ‘distinct population segments’ for
conservation. Conservation Genetics 6: 607–613.
Friesen VL, Burg TM, McCoy KD. 2007. Mechanisms of
population differentiation in seabirds. Molecular Ecology 16:
1765–1785.
Goostrey A, Carss DN, Noble LR, Piertney SB. 1998.

591

Population introgression and differentiation in the great
cormorant Phlacrocorax carbo in Europe. Molecular Ecology
7: 329–338.
Goudet J. 1995. FSTAT (version 1.2): a computer program to
calculate F-statistics. Journal of Heredity 86: 485–486.
Green MC, Waits JL, Avery ML, Tobin ME, Leberg PL.
2006. Microsatellite variation of double-crested cormorant
populations in eastern North America. Journal of Wildlife
Management 70: 579–583.
Hanski I. 1999. Metapopulation ecology. Oxford: Oxford University Press.
Hickman CR, Peters MB, Crawford NG, Hagen C, Glenn
TC, Somers CM. 2008. Development and characterization of
microsatellite loci in the American white pelican (Pelecanus
erythrorhynchos). Molecular Ecology Resources 8: 1439–1441.
Hoarau G, Rijnsdorp AD, van der Veer HW, Stam WT,
Olsen JL. 2002. Population structure of plaice (Pleuronectes platessa L.) in northern Europe: microsatellites
revealed large-sale spatial and temporal homogeneity.
Molecular Ecology 11: 1165–1176.
Hubisz MJ, Falush D, Stephens M, Pritchard JK. 2009.
Inferring weak population structure with the assistance of
sample group information. Molecular Ecology Resources 9:
1322–1332.
Keith JO. 2005. An overview of the American white pelican.
Waterbirds 28: 9–17.
King DT, Anderson DW. 2005. Recent population status of
the American white pelican: a continental perspective.
Waterbirds 28: 48–54.
Knopf FL, Evans RM. 2004. American white pelican
(Pelecanus erythrorhynchos). In: Poole A, ed. The birds of
North America. Ithaca, NY: Cornel Lab of Ornithology.
Lovette IJ, Clegg SM, Smith TB. 2004. Limited utility of
mtDNA markers for determining connectivity among breeding and overwintering locations in three neotropical
migrant birds. Conservation Biology 18: 156–166.
Marion L, Le Gentil J. 2006. Ecological segregation and
population structuring of the Cormorant Phalacrocorax
carbo in Europe, in relation to the recent introgression of
continental and marine subspecies. Evolutionary Ecology
20: 193–216.
Mills LS, Allendorf FW. 1996. The one-migrant-pergeneration rule in conservation and management. Conservation Biology 10: 1509–1518.
Moreno-Matiella LA, Anderson DW. 2005. Water level
variation and its effects on nesting habitat configuration
and availability for the American white pelican at Clear
Lake Reservoir, California. Waterbirds 28: 73–82.
Morris-Pocock JA, Taylor SA, Birt TP, Damus M, Piatt
JF, Warheit KI, Friesen VL. 2008. Population genetic
structure in Atlantic and Pacific Ocean common murres
(Uria aalge): natural replicate tests of post-Pleistocene evolution. Molecular Ecology 17: 4859–4873.
Moum T, Árnason E. 2001. Genetic diversity and population
history of two related seabird species based on mitochondrial DNA control region sequences. Molecular Ecology 10:
2463–2478.
Mylecraine KA, Gibbs HL, Anderson CS, Shieldcastle

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592

592

M. W. REUDINK ET AL.

MC. 2008. Using 2 genetic markers to discriminate among
Canada goose populations in Ohio. Journal of Wildlife Management 72: 1220–1230.
National Audubon Society. 2002. The christmas bird
count historical results [online]. Available at: http://birds.
audubon.org/christmas-bird-count [accessed 15 July 2010].
Neethling M, Matthee CA, Bowie RCK, von der Heyden
S. 2008. Evidence for panmixia despite barriers to gene flow
in the southern African endemic, Caffrogobius caffer (Teleostei: Gobiidae). BMC Evolutionary Biology 8: 325.
Oberholser HC, Kincaid EB Jr. 1974. The bird life of Texas,
Vol. 1. Austin, TX: University of Texas Press.
van Oosterhout C, Hutchinson WF, Willis DPM, Shipley
P. 2004. MICRO-CHECKER: software for identifying and
correcting genotyping errors in microsatellite data. Molecular Ecology Notes 4: 535–538.
Paetkau D, Calvert W, Stirling I, Strobeck C. 1995.
Microsatellite analysis of population structure in Canadian
polar bears. Molecular Ecology 4: 347–354.
Paetkau D, Slade R, Burden M, Estoup A. 2004. Genetic
assignment methods for the direct, real-time estimation of
migration rate: a simulation-based exploration of accuracy
and power. Molecular Ecology 13: 55–65.
Patirana A, Hatch SA, Friesen VL. 2002. Population differentiation in the red-legged kittiwake (Rissa tridactyla) as
revealed by mitochondrial DNA. Conservation Genetics 3:
335–340.
Pekarik C, Hodder C, Weseloh DVC, Matkovich C, Shutt
L, Erdman T, Matteson S. 2009. First nesting of American white pelican on Lake Superior, Ontario, Canada.
Ontario Birds 27: 42–49.
Peters JL, Gretes W, Omland KE. 2005. Late Pleistocene
divergence between eastern and western populations of
wood ducks (Aix sponsa) inferred by the ‘isolation with
migration’ coalescent method. Molecular Ecology 14: 3407–
3418.
Piry S, Alapetite A, Cornuet J-M, Paetkau D, Baudouin
L, Estoup A. 2004. GeneClass2: a software for genetic
assignment and first-generation migrant detection. Journal
of Heredity 95: 536–539.
de Ponte Machado M, Feldheim KA, Sellas AB, Bowie
RCK. 2009. Development and characterization of microsatellite loci from the great white pelican (Pelecanus onocrotalus) and widespread application to other members of
Pelecanidae. Conservation Genetics 10: 1033–1036.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of
population structure using multilocus genotype data. Genetics 155: 945–959.

Prugnolle F, de Meeus T. 2002. Inferring sex-biased dispersal from population genetic tools: a review. Heredity 88:
161–165.
R Development Core Team. 2010. R: a language and environment for statistical computing. Vienna: R Foundation for
Statistical Computing.
Raymond M, Rousset F. 1995. GENEPOP (version 1.2):
population genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248–249.
Ridgely RS, Allnutt TF, Brooks T, McNicol DK,
Mehlman DW, Young BE, Zook JR. 2007. Digital distribution maps of the birds of the western hemisphere, Version
3.0. Arlington, VA: NatureServe.
Ridgway T, Hoegh-Guldberg O, Ayre DJ. 2001. Panmixia
in Pocillopora verrucosa from South Africa. Marine Biology
139: 175–181.
Rocke T, Coverse K, Meteyer C, McLean B. 2005. The
impact of disease in the American white pelican in North
America. Waterbirds 28: 87–94.
Selander RK, Johnston RF, Wilks BJ, Baun GD. 1962.
Vertebrates from the barrier island of Tamaulipas, Mexico.
University of Kansas Publications of the Museum of Natural
History 12: 311–345.
Slatkin M. 1987. Gene flow and the geographic structure of
natural populations. Science 236: 787–792.
Strait LE, Sloan NF. 1975. Movements and mortality of
juvenile White Pelicans from North Dakota. Wilson Bulletin
87: 54–59.
Thompson BH. 1933. History and present status of the
breeding colonies of the White Pelican (Pelecanus erythrorhynchos) in the United States. National Park Service.
Contribution of Wild Life Division, Occasional Paper 1:
156–164.
Wade MJ, McCauley DE. 1988. Extinction and recolonization: their effects on the genetic differentiation of local
populations. Evolution 42: 995–1005.
Ward RD, Skibinski DOF, Woodwark M. 1992. Protein
heterozygosity, protein structure, and taxonomic differentiation. Evolution 26: 73–159.
Yauk CL, Quinn JS. 1999. Genetic structure among breeding
herring gulls (Larus argentatus) from the Great Lakes and
Eastern Canada. Journal of Great Lakes Research 25: 856–
864.
Zardi G, McQuaid CD, Teske PR, Barker NP. 2007. Unexpected genetic structure of mussel populations in South
Africa: indigenous Perna perna and invasive Mytilus
gallopro- vincialis. Marine Ecology Progress Series 337:
135–144.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102, 583–592