ArticlePDF Available

Population genetic and behavioural variation of the two remaining colonies of Providence petrel (Pterodroma solandri)

Authors:
  • Department of Planning Industry and Environment

Abstract and Figures

Knowledge of the dispersal capacity of species is crucial to assess their extinction risk, and to establish appropriate monitoring and management strategies. The Providence petrel (Pterodroma solandri) presently breeds only at Lord Howe Island (~32,000 breeding pairs) and Phillip Island-7 km south of Norfolk Island (~20 breeding pairs). A much larger colony previously existed on Norfolk Island (~1,000,000 breeding pairs) but was hunted to extinction in the 18th Century. Differences in time of return to nesting sites are presently observed between the two extant colonies. Information on whether the Phillip Island colony is a relict population from Norfolk Island, or a recent colonization from Lord Howe Island, is essential to assess long-term sustainability and conservation significance of this small colony. Here, we sequenced the mitochondrial cytochrome b gene and 14 nuclear introns, in addition to genotyping 10 microsatellite loci, to investigate connectivity of the two extant P. solandri populations. High gene flow between populations and recent colonization of Phillip Island (95 % HPD 56–200 ya) are inferred, which may delay or prevent the genetic differentiation of these insular populations. These results suggest high plasticity in behaviour in this species and imply limited genetic risks surrounding both the sustainability of the small Phillip Island colony, and a proposal for translocation of Lord Howe Island individuals to re-establish a colony on Norfolk Island.
This content is subject to copyright. Terms and conditions apply.
RESEARCH ARTICLE
Population genetic and behavioural variation of the two remaining
colonies of Providence petrel (Pterodroma solandri)
Anicee J. Lombal
1
Theodore J. Wenner
1
Nicholas Carlile
2
Jeremy J. Austin
3
Eric Woehler
4
David Priddel
2
Christopher P. Burridge
1
Received: 17 May 2016 / Accepted: 13 September 2016
!Springer Science+Business Media Dordrecht 2016
Abstract Knowledge of the dispersal capacity of species is
crucial to assess their extinction risk, and to establish
appropriate monitoring and management strategies. The
Providence petrel (Pterodroma solandri) presently breeds
only at Lord Howe Island (*32,000 breeding pairs) and
Phillip Island-7 km south of Norfolk Island (*20 breeding
pairs). A much larger colony previously existed on Norfolk
Island (*1,000,000 breeding pairs) but was hunted to
extinction in the 18th Century. Differences in time of return
to nesting sites are presently observed between the two
extant colonies. Information on whether the Phillip Island
colony is a relict population from Norfolk Island, or a recent
colonization from Lord Howe Island, is essential to assess
long-term sustainability and conservation significance of
this small colony. Here, we sequenced the mitochondrial
cytochrome bgene and 14 nuclear introns, in addition to
genotyping 10 microsatellite loci, to investigate connectiv-
ity of the two extant P. solandri populations. High gene flow
between populations and recent colonization of Phillip
Island (95 % HPD 56–200 ya) are inferred, which may delay
or prevent the genetic differentiation of these insular popu-
lations. These results suggest high plasticity in behaviour in
this species and imply limited genetic risks surrounding both
the sustainability of the small Phillip Island colony, and a
proposal for translocation of Lord Howe Island individuals
to re-establish a colony on Norfolk Island.
Keywords Oceanic seabird !Pterodroma solandri !Gene
flow !Behavioural variation !Conservation management
Introduction
Understanding mechanisms of population divergence has
important implications for successful conservation of
species (Avise 2000). While adaptation to different
Electronic supplementary material The online version of this
article (doi:10.1007/s10592-016-0887-5) contains supplementary
material, which is available to authorized users.
&Anicee J. Lombal
anicee.Lombal@utas.edu.au
Theodore J. Wenner
theodore.Wenner@utas.edu.au
Nicholas Carlile
nicholas.Carlile@environment.nsw.gov.au
Jeremy J. Austin
jeremy.austin@adelaide.edu.au
Eric Woehler
eric.woehler@utas.edu.au
David Priddel
david.priddel@environment.nsw.gov.au
Christopher P. Burridge
chris.Burridge@utas.edu.au
1
School of Biological Sciences, University of Tasmania,
Hobart, TAS 7001, Australia
2
Department of Environment and Conservation,
P.O. Box 1967, Hurstville, NSW 2220, Australia
3
Australian Centre for Ancient DNA, School of Biological
Sciences, University of Adelaide, Adelaide, SA 5005,
Australia
4
Institute for Marine and Antarctic Studies, University of
Tasmania, Hobart, TAS 7053, Australia
123
Conserv Genet
DOI 10.1007/s10592-016-0887-5
environments may be important for population persistence,
it may also inhibit movements amongst populations,
potentially reducing genetic variability through random
genetic drift and inbreeding (Frankham 1996; Hedrick and
Kalinowski 2000), which may decrease adaptability to
future environmental variations (Frankham et al. 2002).
Therefore, quantifying the dispersal of individuals, which
is driven by the variability in intrinsic patch quality
between different areas such as resource availability or
population density (Bowler and Benton 2005), is essential
to predict the long-term resilience and persistence of pop-
ulations, and to inform management decisions such as
supplementation and translocation (Frankham 1996).
Seabirds provide useful model systems for studying
mechanisms of population divergence given their often
philopatric behaviour and discrete breeding distributions
(Friesen et al. 2007; Friesen 2015). Most oceanic seabirds
breed in discrete colonies, and may constitute a population
structure known as metapopulations, where occasional
dispersal facilitates re-establishment or supplementation of
populations following declines (Oro 2003). Nevertheless,
the relative importance of the factors influencing dispersal
between seabird colonies remains unclear (Friesen 2015;
Welch et al. 2012). It is also crucial to identify factors
influencing dispersal between seabird colonies to predict
events such as genetic divergence or inbreeding depression
(Avise 1996; Charlesworth and Charlesworth 1987). While
physical barriers to dispersal and philopatry appear to be
the main inhibitors of gene flow among seabird colonies
(Friesen 2015; Warham 1990), other mechanisms have also
been detected, such as differences in foraging distribution
during the breeding and non-breeding seasons, differences
in ocean regimes, and differences in breeding phenology
(Burg and Croxall 2001; Friesen 2015; Wiley et al. 2012).
For example, allochronic populations of band-rumped
storm-petrel (Oceanodroma castro) appear genetically
isolated in five archipelagos throughout the Atlantic and
Pacific Oceans in the absence of physical barriers to gene
flow (Smith and Friesen 2007). Conversely, whether
genetic isolation exists among colonies that exhibit other
phenological or circadian differences, e.g., diurnal versus
nocturnal colony attendance has yet to be investigated.
The Providence petrel (Pterodroma solandri) is classi-
fied as vulnerable under both the IUCN Red List of
Threatened Animals (Criteria D2) and the New South Wales
Threatened Species Conservation Act 1995 due to its
restricted breeding range. The only significant breeding
locality is Lord Howe Island (*32,000 breeding pairs)
(Bester 2003), a small island located 600 km off the eastern
coast of Australia (Fig. 1). Providence petrels previously
bred on Norfolk Island (*1,000,000 breeding pairs),
located approximately 1100 km northeast of Lord Howe
Island (Fig. 1), before becoming extirpated following
European settlement by the late 18th century (Medway
2002a). This species was considered extinct within the
Norfolk Island group until 1986 when a small population
(*20 breeding pairs) was discovered on Phillip Island,
7 km south of Norfolk Island (Hermes et al. 1986) (Fig. 1).
There is no evidence justifying taxonomic separation
between Phillip Island and Lord Howe Island Providence
petrels. However, it has been reported that Lord Howe
Island individuals predominantly arrive at the colony dur-
ing daylight (Bester et al. 2002; Medway 2002b), while
Phillip Island individuals return to their breeding sites only
after dusk (pers. obs.). This may relate to the presence of
diurnal aerial predators—Brown Goshawks Accipiter fas-
ciatus—at the time of European settlement on Norfolk
Island (Medway 2002b), although no such predation risk
presently exists. Alternatively, differences in foraging
areas may explain time of return to colony (e.g., Dias et al.
2012). Given the possibility of selective significance, the
observed difference in behaviour between colonies may
inhibit gene flow between them.
Here we report a comprehensive study of the genetic
distinctiveness between the two remaining breeding colo-
nies of Providence petrel, to infer the dispersal patterns of
this species and the conservation status of the small Phillip
Island colony. We developed three genetic data sets, con-
sisting of DNA sequences from mitochondrial and 14
nuclear regions and genotypes from 10 microsatellite loci, to
investigate genetic connectivity and evolutionary history of
Providence petrel colonies. Our study is also relevant to the
proposed re-establishment of a colony on Norfolk Island
using individuals from Lord Howe Island, with the aim of
reducing the extinction risk of this species, and restoring the
input of marine-derived nutrient into the ecosystem.
Materials and methods
Sample collection and DNA extraction
We collected blood samples from P. solandri individuals
(n=151) from four localities on Lord Howe Island
(31"300S, 159"050E): Mount Gower (MG n=30), Far Flats
(FF n=79), George’s Bay (GB n=22) and Muttonbird
Point (MBP n=20) (Fig. 1). We sampled the one locality
on Phillip Island (29"120S, 167"950E) where the Providence
petrel has been observed to nest: Jacky Jacky (JJ n=32)
(Fig. 1). All blood samples were collected from Providence
petrels under Animal Ethics Permit number AEC 021028/02
issued by the Department of Environment, Climate Change
and Water (NSW). Genomic DNA was extracted from 183
individuals using a Qiagen DNeasy#Blood and Tissue kit
following the manufacturer’s protocol.
Conserv Genet
123
Mitochondrial and nuclear DNA sequencing
We sequenced 183 individuals (151 from Lord Howe
Island, 32 from Phillip Island) for a 872 bp fragment of the
mitochondrial cytochrome bgene using primers L14841
(Kocher et al. 1989) and H15547 (Edwards et al. 1991). We
also sequenced 40 individuals (20 from FF, Lord Howe
Island, 20 from JJ, Phillip Island) for *500 bp long frag-
ments of 14 avian nuclear introns (Backstro
¨m et al. 2008;
Patterson et al. 2011; Silva et al. 2011). Primer sequences,
optimal annealing temperatures and approximate locus
length in P. solandri are shown in the electronic supple-
mentary material, SI 1.
All fragments were PCR amplified with the Man-
goTaq
TM
DNA polymerase following the manufacturer’s
protocol (Bioline Inc.). PCR reactions were performed in
35 lL volumes using 50–100 ng DNA, and final concen-
trations of 0.5 U DNA polymerase, 0.2 mM of each dNTP,
1.5 mM MgCl
2
and 0.3 lM of each primer. The thermal
cycling profiles included an initial denaturation at 95 "C
for 1 min followed by 29 cycles of 95 "C for 30 s,
60–46 "C (decreasing the annealing temperature by 0.5 "C
per cycle) for 40 s, and an extension of 72 "C for 90 s, with
a final extension of 72 "C for 10 min followed by four
similar cycles but with a constant annealing temperature at
45 "C. Negative controls were included with each set of
PCRs.
Nucleotide sequences were determined on both strands
of PCR products using a 3730xl DNA Analyzer (Applied
Biosystem#) at Macrogen Inc., Korea. Sequences were
aligned using the MUSCLE algorithm (Edgar 2004) in
CODONCODE ALIGNER v3.7.1.1 (CodonCode Corpo-
ration). For sequences containing multiple heterozygous
positions, we used the maximum likelihood method
implemented in PHASE v2.2.1 (Stephens et al. 2001) to
reconstruct the haplotype phase of the sequences. We
conducted three independent runs of 10,000 iterations per
locus with a different seed number to verify convergence,
and discarded the first 1000 samples as burn-in. Phased
haplotypes showing a probability [0.8 were used for fur-
ther analyses.
Microsatellite genotyping
Genotypes of 183 individuals (151 from Lord Howe Island,
32 from Phillip Island) were determined at 10 polymorphic
microsatellite loci (Ptero9,Ptero7,Ptero6,Ptero4,Par-
m02,Parm03,Paequ 03,Paequ 13,Calex01,RBG29)
following Lombal et al. (2015).
Data analyses
Tests of assumptions and genetic variation
To assess levels of DNA sequence variation within colo-
nies (Lord Howe Island, Phillip Island), haplotypic diver-
sity h(Nei 1987), haplotype ratios X
H
, nucleotide diversity
p(Tajima 1983), and nucleotide diversity ratios p
R
AUSTRALIA
_
_
_
_
Fig. 1 Sampling locations for
Pterodroma solandri. Lord
Howe Island: Far Flats (FF,
n=79), George’s Bay (GB,
n=22), Muttonbird Point
(MBP, n =20), Mount Gower
(MG, n =30). Phillip Island:
Jacky Jacky (JJ, n =32). Mount
Bates was the location of the
extinct Norfolk Island colony
Conserv Genet
123
(Mardulyn et al. 2009) were calculated for mitochondrial
and nuclear intron DNA sequences with SPADS v 1.0
(Dellicour and Mardulyn 2014). To test whether patterns of
genetic variation deviated from neutral expectations, Taji-
ma’s Dtest (Tajima 1983) and Fu and Li’s D* test (Fu and
Li 1993) were performed using DNASP v 5.10 (Librado
and Rozas 2009).
Microsatellite loci were tested for departure from
Hardy–Weinberg equilibrium for each colony (Lord Howe
Island, Phillip Island) using exact tests in ARLEQUIN v
3.5.1.2 (Excoffier and Lischer 2010), where Markov chain
parameters were set at 10,000 dememorizations, and
10,000 iterations. The inbreeding coefficient F
is
(1–H
o
/H
E
)
was calculated per colony in FSTAT 2.9.2. (Goudet 1995),
then tested for significant departure from zero using 10,000
permutations of alleles among individuals. Allelic diversity
N
a
, and allelic richness R
s
, which uses a rarefaction method
to standardize uneven sample size (Petit et al. 1998), were
computed with the software HP-RARE v 1.0. (Kalinowski
2005).
Population connectivity and identification of dispersers
Estimates of pairwise population differentiation between
Lord Howe Island and Phillip Island (F
st
,G
st
,N
st
and U
st
)
were determined using SPADS v 1.0. (Dellicour and
Mardulyn 2014). The statistical significance of F
st
,G
st
,N
st
and U
st
values was assessed by recalculating them based on
10,000 random permutations of individuals among islands.
TCS networks (Clement et al. 2000) were inferred for
mitochondrial and nuclear DNA sequences using PopART
(http://popart.otago.ac.nz). AMOVA U-statistics (U
SC
U
ST
U
CT
) (Excoffier et al. 1992) were calculated for the mito-
chondrial locus (cyt b) (Group 1 =JJ, Phillip Island;
Group 2 =FF, MBP, MG, GB, Lord Howe Island) with
10,000 permutations of individuals and sampling sites. In
addition, to evaluate the extent to which sequence variation
was partitioned, a matrix of pairwise population differen-
tiation was constructed between all sampling sites (n =5).
F
st
and R
st
(Slatkin 1995) were calculated for
microsatellites, the latter assuming a generalized stepwise
mutation model (SMM), using FSTAT 2.9.2 (Goudet
1995), with significance assessed based on 10,000 permu-
tations of alleles among samples. Contingency tables of
alleles were generated, and classified (Kimura and Ohta
1978) using the log-likelihood statistic G (Goudet et al.
1996). G
st
were not calculated for these high mutation rate
markers as recommended by Whitlock (2011). AMOVA
(Excoffier et al. 1992) was performed with 10,000 per-
mutations of individuals among sampling sites (Group
1=JJ, Phillip Island; Group 2 =FF, MBP, MG, GB,
Lord Howe Island), and a pairwise population differentia-
tion matrix was constructed among all sampling sites
(n =5) using GENODIVE v 2.0b28 (Meirmans and Van
Tienderen 2004).
As low genetic divergence among populations could
reflect high historical dispersal among populations that are
now isolated, we used kinship-based methods to estimate
current gene flow between P. solandri colonies (Lord
Howe Island, Phillip Island), as recommended when there
are low frequency alleles present (Hardy and Vekemans
2002). The statistical rigour and power of this approach
using kinship coefficients (h
ij
) depends upon the overall
level of genetic variation, and not the degree of divergence
between populations (Palsboll et al. 2010). We calculated
h
ij
(Loiselle et al. 1995) for each pair of individuals in
GENODIVE (Meirmans and Van Tienderen 2004). To test
whether individuals collected in the same colony were
more closely related to each other than individuals col-
lected in different colonies, we performed a non-parametric
Permutational Multivariate Analysis of Variance (PER-
MANOVA) on h
ij
. This approach partitions the distance
matrix according to the source of variation (e.g., among vs.
within), and compares the sum of square distances among
and within these groups as implemented in PERMA-
NOVA ?1.0.6 software add-on running on PRIMER6
(Clarke and Warwick 2005). To assign h
ij
to independent
genetic clusters, we used a K-Means method to calculate
the Calinski-Harabasz pseudo F-statistics (Calin
´ski and
Harabasz 1974), which focuses on reducing the within-
group sum of squares, for K=2–183, with 10,000 itera-
tions per cluster, as implemented in the package clusterSim
in R v 3.2.1.
Bayesian clustering analysis and individual assignment
Bayesian clustering analysis, which uses MCMC simula-
tion to assign coancestry of individuals to independent
genetic clusters (K) based on individual microsatellite
genotypes without a priori assumptions of populations, was
implemented in STRUCTURE v 2.3.3 (Falush et al. 2003;
Pritchard et al. 2000). Exploratory runs showed that a burn-
in of 200,000 followed by 1,000,000 iterations achieved
stable estimates. 20 replicate runs were then performed for
all values of K=1–8, reflecting the highest expected
number of genetic cluster (n=5, Far Flats (FF), Mount
Gower (MG), George’s Bay (GB), Muttonbird Point (MP),
and Jacky Jacky (JJ)) plus three (Evanno et al. 2005). We
used the admixture model, and assumed correlated allele
frequencies, which is expected to perform better when
genetic structure is weak or when the number of loci is \20
(Hubisz et al. 2009), with Prior Mean =0.01, and Prior
SD =0.05. We implemented priors for alpha (a=1) and
lambda (k=1), specifying the degree of admixture
between populations and the distribution of allele fre-
quencies respectively, for all populations. The optimal
Conserv Genet
123
number of clusters (K) was estimated by calculating the
second order-rate of change (DK) of the likelihood function
(ln P(X/K)) with respect to each K(Evanno et al. 2005), as
implemented in the program STRUCTURE HARVESTER
(Earl 2012). The results of all runs were summarized in
CLUMPP v 1.1.1 (Jakobsson and Rosenberg 2007) using
the FullSearch algorithm, and then visualized using DIS-
TRUCT v 1.1 (Rosenberg 2004).
Migrant individuals between colonies (Lord Howe
Island, Phillip Island) were identified using exclusion
methods as implemented in GENECLASS 2.0 (Piry et al.
2004). We used the exclusion criterion L
h
/L
max
(Paetkau
et al. 2004) to compute the probability that an individual
belongs to a colony. We compared the Bayesian (Rannala
and Mountain 1997) and frequency based criteria (Paetkau
et al. 2004) to calculate the likelihood of individual origin.
We used the Paetkau et al. (2004) resampling methods
based on allele frequency (Paetkau et al. 2004), which
demonstrated low type I error rates (1 % of the number of
individuals per population that appear to be immigrants by
chance). This method generates population samples of the
same size as the reference population sample, as recom-
mended for detection of first generation migrants (Piry
et al. 2004). The marginal probability of given individual
multilocus genotype was compared to the distribution of
marginal probabilities of randomly generated multilocus
genotypes (100,000 replicates) with a type I error threshold
setting at a
0.01
and a
0.05
.
Estimation of divergence time
We used IMa and its model of isolation with migration
(Hey and Nielsen 2007) to simultaneously estimate
migration (m
1
,m
2
) and lineage divergence time (t) between
P. solandri colonies (Lord Howe Island, Phillip Island).
This coalescent-based model is based on several assump-
tions including neutrality, random mating in ancestral and
descendent populations, and free recombination between
loci, but none within loci (Hey and Nielsen 2004; Nielsen
and Wakeley 2001). Lack of recombination within nuclear
introns was tested using the four-gamete test as described
by Hudson and Kaplan (1985), and loci suspected to be
under selection were excluded from analyses (Supple-
mentary Material, SI 3). An IMa exploratory run was
performed to assess a range of prior distributions that
include most of the range over which the posterior density
is not trivial. Analyses were then run three times with
different seed numbers to test for convergence, with
10,000,000 sampled steps following a discarded burn-in of
200,000 steps, with a two-step linear heating scheme with
five chains. We implemented the Hasegawa-Kishino-Yano
(HKY) (Hasegawa et al. 1985) model for the mitochondrial
data, the infinite sites mutation model (IS) (Kimura 1969)
for the nuclear introns, and the Stepwise Mutation Model
(SMM) (Kimura and Ohta 1978) for microsatellites.
Mutation rates were given as priors to the analysis with
l=1.89 910
-8
and 3.6 910
-9
substitution/site/year for
cyt band nuclear introns respectively, as recommended for
other seabirds (Axelsson et al. 2004; Weir and Schluter
2008), and l=5910
-4
substitution/site/year for
microsatellites (Brown et al. 2010). To assess the estimates
of demographic parameters, we used a generation time
T=10 years, as calculated based on the following equa-
tion T =A?p/(1-p) (Sæther et al. 2004), with p the
adult survival rate (p =0.82) (Brooke 2004), and A the
age of sexual maturity (A =6 years) (Warham 1990).
Parameter trend line plots and values of effective sample
sizes (ESS) were inspected after each run.
Demographic history
Historical demographic changes in the only significant
colony of Providence petrels (Lord Howe Island) were
inferred from two complimentary coalescent modeling
approaches of microsatellite data using MSVAR v0.4
(Beaumont 1999) and MSVAR v1.3 (Storz and Beaumont
2002). This approach is more robust than classic methods
based on summary statistics to detect changes in population
size (Girod et al. 2011).
MsVar v0.4 inferred the magnitude of change in popu-
lation size (r =N
0
/N
1
, where N
0
and N
1
are current and
ancestral population sizes, respectively) assuming a SMM
model for the microsatellite loci. We initially conducted
three independent simulations varying the prior distribu-
tions to examine their effect on the posterior distribution.
We then ran the simulation three times under the expo-
nential and the linear model, with different seed numbers
for each dataset, for 4 910
9
iterations with parameter
values recorded every 1 910
5
iterations, resulting in
40,000 records. We discarded 10 % of recorded values for
each chain (i.e., burn-in), and we performed the Brooks,
Gelman and Rubin Convergence diagnostic tests (Gelman
and Rubin 1992) as implemented in the package BOA
(Smith 2007) for R version 3.2.1. (Venables and Smith
2001). We considered that chains converged well when
values lower than 1.1 were obtained. The chains were then
combined to estimate the 90 % high probability density
(HPD) of demographic parameters using the package
CODA as implemented in R (Plummer et al. 2006). The
strength of evidence for population increase versus
decrease was evaluated by calculating the Bayes factor of
each of the simulations (Girod et al. 2011; Storz et al.
2002). This ratio can be estimated by counting the number
of states in the chains in which the population has
decreased (i.e., N
0
/N
1
\1), and then dividing this by the
number of states in which the population has increased
Conserv Genet
123
(i.e., N
0
/N
1
[1) with BF 0–3 no support of contraction,
3–10 substantial support, [10 strong support (Storz and
Beaumont 2002).
Msvar 1.3 was used to quantify population sizes and
time of change. MsVar 1.3 uses probable genealogies of
allele frequency data to generate posterior probability
distributions of four natural demographic parameters,
U=N
0
,N
1
, ta, and h,where N
0
and N
1
are the current and
the ancestral effective population size respectively, ta is the
time since the demographic changes began, and h=4N
0
l,
the rate of mutation scaled by population size. This model
differs from the previous model in that all loci are used in
the same MCMC simulation, reducing density estimation
error, and that all parameters are free to vary among loci.
We inferred broad normal distribution priors and hyper-
priors (Supplementary material, SI 2), and we ran the
simulation three times under the exponential model to
evaluate recent changes in population size (log
10
(T) \10).
MCMC chain convergence, 90 % HPD of posterior dis-
tributions and Bayes factors were inferred as described for
Msvar v0.4.
Results
We sequenced 872 bp of the mtDNA cytochrome bgene in
151 and 32 individuals from Lord Howe Island and Phillip
Island, respectively, and a total of 7837 bp comprising 14
nuclear introns in 20 individuals from both colonies,
defining 2–9 (Phillip Island) and 1–17 (Lord Howe Island)
alleles (Supplementary Material SI 3). No significant dif-
ference in nucleotide diversities (p) between colonies was
detected (One-way ANOVA; H
0
=means of p
R
are equal
in Lord Howe Island and Phillip Island, where p
R
repre-
sents the nucleotide ratio; F =0.91; pvalue =0.349; see
p
R
values in Supplementary Material SI 3). Tajima’s
Dstatistics showed significant negative values in the
mitochondrial locus (cyt b,D=-1.987, p \0.05), and in
one nuclear intron (d-cryst,D=-2.030, p \0.05) for
Lord Howe Island and Phillip Island populations respec-
tively, while Fu and Li’s D* tests showed negative values
for two loci (cyt b,D*=-2.920, p \0.05, and 16214,
D*=-3.110, p \0.05) for Lord Howe Island, and in one
locus (Pema05,D*=-2.167, p \0.05) for Phillip Island
(Supplementary Material SI 3).
Ten microsatellite loci were genotyped in 151 and 32
individuals from Lord Howe Island and Phillip Island,
defining 2–51 and 4–30 alleles per locus, respectively. No
significant difference in genetic diversity between popula-
tions was detected (Kruskal–Wallis test; H
0
=means of R
s
are equal in Lord Howe Island and Phillip Island, where R
s
represents allelic richness; F =1.12; p-value =0.289; see
R
s
values in Table 1), and no significantly positive values
of F
is
were found for either Lord Howe Island or Phillip
Island (Table 1).
Population connectivity
Visual inspection of haplotype networks (Fig. 2, Support-
ing Material SI 4), observation of low F-statistics (global
F
st
=0.004, p [0.05; global G
st
=0.004, p [0.05,
Table 2) and lack of significant phylogeographic signals
(global U
st
=0.019, p [0.05, global N
st
=0.033,
p[0.05, Table 2) indicate no genetic differentiation
between Lord Howe Island and Phillip Island. AMOVA U-
statistics showed no differentiation between sampling
locations or group of sampling locations for cyt b(Group
1=JJ, Phillip Island; Group 2 =FF, MBP, MG, GB,
Lord Howe Island; U
SC
=0.0004, p [0.05; U
ST
=0.016,
p[0.05; U
CT
=0.016, p [0.05), and the F
st
pairwise
matrix showed no significant genetic structure between
pairs of P. solandri sampling sites (Table 3). F
st
,R
st
and
AMOVA F-statistics obtained with microsatellites were
not significantly different from zero between Phillip Island
and Lord Howe Island sampling locations (global
F
st
=0.006, p [0.05; global R
st
=0.004, p [0.05,
Table 1). Pairwise F
st
indicated no genetic differentiation
between P. solandri sampling sites (Table 3). These results
refute structuring of genetic variation between Lord Howe
Island and Phillip Island.
Kinship coefficients (h
ij
) ranged from -0.28 to 0.69 and
-0.27 to 0.36 within and between colonies, respectively.
The analysis of variance of h
ij
showed no significant dif-
ferences between ‘within-colonies’ and ‘among-colonies’
(pseudo-F
1,182
=0.993, P=0.424). The clustering anal-
ysis, based on Calinski-Harabasz pseudo-Fstatistics,
showed highest pseudo-Ffor K"2(Fig. 3), which does
not support Phillip Island and Lord Howe Island as
genetically distinct colonies.
Bayesian clustering analysis and individual
assignment
Evaluation of lnP(X/K), DK, and Q obtained with
STRUCTURE supported K=4, although genetic clusters
did not reflect geographical localities. Each individual
contained roughly equal coancestry from the four clusters
(Supplementary material SI 5). The frequency-based and
Bayesian assignment methods (Lord Howe Island vs.
Phillip Island colonies) implemented in GENECLASS 2
showed 3 and 7 (a
0.01
), and 12 and 23 (a
0.05
) first-gener-
ation migrants, respectively (Supplementary Material SI 6).
Conversely, the two methods showed equivalent results
with 59 % of individuals correctly assigned (108 out of
Conserv Genet
123
183) with an average probability of 54.28 % at a
0.05
, and
53.39 % at a
0.01
. This low confidence reflects the similarity
between likelihoods of genotypes across populations.
Estimation of divergence time
Implementations of the isolation-with-migration model
using microsatellites, nuclear introns and mitochondrial
loci resulted in unimodal posterior density curves of
migration parameters, which were similar across the three
Table 1 Characterization of genetic diversity and summary statistics in P. solandri for 10 microsatellites loci
Locus name Length (bp) Lord Howe Island (n =151) Phillip Island (n =32) All populations
AR
s
H
0
H
e
F
is
AR
s
H
0
H
e
F
is
F
st
R
st
F
is
Ptero09 187–235 17 14.47 0.671 0.879 0.234 13 10.93 0.688 0.886 0.227 -0.004 0.024 0.232
Ptero07 264–344 51 41.26 0.968 0.954 -0.008 30 21.25 1.000 0.972 -0.029 -0.007 0.002 -0.010
Parm03 177–181 6 5.00 0.654 0.663 0.011 4 3.75 0.469 0.637 0.268 0.004 0.024 0.055
Ptero06 141–149 2 2.00 0.033 0.185 0.821 3 2.94 0.156 0.347 0.553 0.038 -0.016 0.746
PaEquation 13 146–148 5 4.20 0.266 0.352 0.239 4 3.023 0.355 0.421 0.159 0.006 0.004 0.222
Calex01 237–255 14 13.75 0.859 0.859 -0.002 13 11.33 0.938 0.892 -0.052 0.008 -0.006 -0.011
Ptero04 150–168 11 10.45 0.821 0.826 0.006 10 8.22 0.906 0.812 -0.119 -0.003 0.021 -0.016
RBG29 124–136 9 7.51 0.653 0.805 0.215 6 5.90 0.813 0.796 -0.021 -0.014 -0.010 0.181
Parm02 192–198 6 5.00 0.415 0.411 0.003 5 4.20 0.375 0.489 0.235 0.006 -0.009 0.049
PaEquation 03 222–232 10 8.88 0.614 0.678 0.038 8 6.94 0.844 0.827 -0.020 0.019 -0.004 0.030
Allelic diversity A; allelic richness R
s
; and tests for departure from Hardy–Weinberg equilibrium. Inbreeding coefficient F
is
(1-H
o
/H
E
).
Population structuring (F
st
and R
st
)
All p-values [0.05
Fig. 2 Haplotype network of Providence petrel (Pterodroma solan-
dri) mtDNA haplotypes based on the TCS algorithm. Haplotypes are
represented by circles, where the size of each circle is proportional to
the frequency of the corresponding haplotype. Lines on connecting
branches represent mutations. Black: Lord Howe Island individuals.
White Phillip Island individuals
Table 2 Summary statistics in P. solandri for the mitochondrial
Cytochrome b gene and 14 nuclear introns
Locus name F
st
U
st
G
st
N
st
Cyt b 0.0105 0.1050 0.0061 0.0200
d-cryst 0.0810 0.0810 0.0283 0.0810
Lipo2 0.0000 0.0000 0.0000 0.0000
Pema01 -0.0014 -0.0014 -0.0119 -0.0014
Pema05 0.0256 0.0256 0.0141 0.0256
Pema07 -0.0006 -0.0006 0.0135 0.0058
Pema10 0.0148 0.0148 0.0260 0.0148
Pema12 -0.0148 -0.0148 -0.0102 -0.0148
Pema13 -0.0129 -0.0129 -0.0208 -0.0129
Pema14 0.0203 0.0203 0.0130 0.0203
16214 -0.0002 -0.0002 -0.0117 -0.0002
20454 0.0166 0.0166 -0.0050 0.0166
22519 -0.0037 -0.0037 -0.0027 -0.0037
24206 -0.0221 -0.0221 -0.0197 -0.0221
24972 0.0170 0.0170 0.0178 0.0170
Pairwise population differentiation between Lord Howe Island and
Phillip Island colonies, F
st
,G
st
,N
st
and U
st
, where U
st
represents the
direct analog of Wright’s F
st
for nucleotide sequence diversity (Ex-
coffier et al. 1992)
All p-values [0.05
Conserv Genet
123
runs. Migration rates were of 0.32 migrants/generation
from the Phillip Island colony to the Lord Howe Island
colony (0.24–0.49 90 % HPD) (Fig. 4a), and 8.6
migrants/generation from the Lord Howe Island colony to
the Phillip Island colony (8.44–8.73 90 % HPD) (Fig. 4b).
Divergence time estimates were also convergent across all
analyses, corresponding to 88 years (56–200 90 % HPD)
(Fig. 4c).
Demographic history
Results from coalescent modelling of microsatellites using
MsVar v0.4 and Msvar 1.3 both showed a strong signal for
large population decrease in the Lord Howe Island colony
(Fig. 5a, b). Combining all simulations for all datasets,
contemporary effective population size, N
0
(30 [0–1862],
mean and 90 % HPD) was three orders of magnitude
smaller than the ancestral effective population size, N
1
(177,827 [5888–4,265,795], mean and 90 % HPD)
(Fig. 5b). All Bayes factors obtained with both methods
were "10 in favour of population decrease rather than
increase. The time when the ancestral Lord Howe Island
colony started to decrease (mean log
10
(t
f
)=2.785, expo-
nential model) suggests a recent decrease in this colony
(609 years, [20–10,000] mean and 90 % HPD) (Fig. 5c).
Discussion
We generated three genetic data sets consisting of DNA
sequences from mitochondrial and 14 nuclear regions and
genotypes from 10 microsatellite loci to investigate genetic
connectivity and demographic history of Providence petrel
(Pterodroma solandri) colonies, an oceanic seabird IUCN
uplisted as Vulnerable due to its restricted breeding range.
High gene flow between the two remaining colonies of
Providence petrel (Lord Howe Island and Phillip Island)
was evident despite individuals at the two colonies show-
ing different time of return to nesting sites. In addition,
time of divergence among colonies appears recent, sug-
gesting recent colonization of Phillip Island by individuals
from Lord Howe Island. These results suggest high plas-
ticity in behaviour rather than adaptive divergence in
Providence petrels, and imply limited genetic risks sur-
rounding the sustainability of the Phillip Island colony.
Contemporary genetic differentiation
The analyses conducted here on multiple datasets indicate
high genetic connectivity between the two remaining
populations of Providence petrel (Lord Howe Island and
Phillip Island). While low genetic divergence among
populations could also reflect high historical connectivity
between populations that are now isolated (Palsboll et al.
2010), we also investigated contemporary gene flow among
populations. We compared the variation of kinship coeffi-
cients within and between Providence petrel colonies (Lord
Howe Island and Phillip Island), and showed that individ-
uals coming from the same colony were as related genet-
ically as individuals coming from different colonies; the
best clustering of individuals was also independent of
breeding locality. These results confirmed high current
dispersal capacity of Providence petrels, which suggests
that species-wide genetic diversity is being maintained by
natural dispersal between colonies.
Time of colonization
Maximum likelihood estimates obtained from the isolation-
with-migration model showed that Providence petrel
Table 3 Pairwise differentiation matrix among P. solandri colonies
JJ FF MBP GB MG
JJ -0.022 -0.003 0.042 0.022
FF 0.005 -0.003 0.004 -0.017
MBP -0.010 0.012 -0.058 0.002
GB -0.004 0.002 0.000 --0.002
MG -0.003 0.003 0.004 -0.010 -
F
st
among Pterodroma solandri. Lord Howe Island: Far Flats FF;
George’s Bay GB; Muttonbird Point MBP; Mount Gower MG; Phillip
Island: Jacky Jacky JJ; Above diagonal: pairwise differentiation
matrix for mitochondrial DNA. Below diagonal: pairwise differenti-
ation matrix for 10 microsatellites
All p-values [0.05
1000
2000
3000
4000
5000
0 50 100 150
Clusters (K)
Pseudo -
F values (F)
Fig. 3 Calinski-Harabasz pseudo F-statistic density for kinship
coefficients (K =2–183). Highest density of pseudo-F-statistic
values determine the most likely number of clusters among P.
solandri individuals
Conserv Genet
123
colonies (Lord Howe Island and Phillip Island) became
separated between 56 and 200 years ago. This suggests that
individuals from Lord Howe Island were prospecting new
habitats on Phillip Island after the extirpation of the Nor-
folk Island colony. These results indicate limited genetic
risks surrounding the sustainability of the small Phillip
Island colony of Providence petrels. Indeed, as dispersal of
prospectors is positively related to the presence of con-
specifics (Serrano et al. 2004), we can expect additional
gene flow from Lord Howe Island to Phillip Island in the
near future. Conversely, the fact that the Phillip Island
colony was only discovered in 1986 may be explained by
the first explorations of this small island in the 1970s
(Priddel et al. 2010). We are presently analysing ancient
DNA samples from the Norfolk Island colony to assess its
historical connectivity with Lord Howe Island.
Behavioral difference in timing of colony attendance
Despite Providence petrel colonies being highly connected
genetically, for the period of courtship and early incuba-
tion, Lord Howe Island individuals predominantly arrive at
the colony during daylight, whereas Phillip Island indi-
viduals return to their breeding sites only after dusk.
Numerous studies have illustrated the importance of
behavioural plasticity as a fundamental trait of life history
strategies in seabirds living in highly dynamic and variable
environment (Falk et al. 2002; Paiva et al. 2009; Reed et al.
2009). Moreover, petrels have the capacity to use olfactory
senses to find burrows at night, and this strategy is not
exclusive to individuals showing nocturnal arrival at
colonies (Bonadonna and Bretagnolle 2002; Dell’Ariccia
and Bonadonna 2013). Individuals showing diurnal arrival
are also able to use olfaction as the basic sensory input for
homing at night, and use it if necessary (Dell’Ariccia and
Bonadonna 2013). These observations imply that all petrels
are able to return to their burrows at night, and that indi-
viduals alter their behaviour to environmental conditions
without necessarily requiring genetic adaptation. Hence, it
is likely that prospectors from Lord Howe Island have
switched their behaviour on Phillip Island.
Earlier studies suggest that avoidance of predators is
likely to be the main factor responsible for nocturnal col-
ony arrival in small Procellariiformes (Keitt et al. 2004;
McNeil et al. 1993; Warham 1990; Watanuki 1986).
However, Providence petrels from Phillip Island as well as
other seabird species possess a nocturnal arrival behaviour
even in the absence of diurnal predators (Keitt et al. 2004).
0.20 0.30 0.40 0.50
0.000 0.010 0.020
m1
Posterior probability
7.5 8.0 8.5 9.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
m2
Posterior probability
50 100 150 200
0 100 200 300 400
t
Posterior probability
ab
c
Fig. 4 Population divergence
genetic parameters. Marginal
posterior probability
distributions for the Isolation
with Migration demographic
parameters. amigrants/
generation from Lord Phillip
Island to Lord Howe Island
(m
1
). bmigrants/generation
from Lord Howe Island to
Phillip Island (m
2
). ctime of
divergence (t, years)
Conserv Genet
123
Considering establishment of Providence petrels on Phillip
Island in the 1800 s, this behaviour may also be a recent
adaptation to the presence of hawks on the island at the
time of European settlement (Medway 2002b). Another
explanation may be related to foraging, as has been
observed for a number of seabird taxa (Baduini 2002; Dias
et al. 2012). For example, Cory’s shearwaters (Calonectris
diomedea) show intraspecific variation in colony arrival
depending on the marine region and abundance of prey,
and are high flexibility in their daily routines (Dias et al.
2012). However, unpublished logger data from Lord Howe
Island individuals suggests foraging throughout the Coral
and Tasman Seas during the breeding season (Carlile, per.
obs.), such that it is difficult to imagine differences in
foraging locations between Lord Howe and Phillip Island
individuals.
Demographic history
Coalescent modelling of microsatellites indicated a past
bottleneck in Providence petrel. This significant decrease in
population size is estimated to have occurred approxi-
mately 600 years ago. However, there is a broad uncer-
tainty surrounding this date estimate. A survey of
unconsolidated sediments on Lord Howe Island did not
indicate human occupation of this island before the Euro-
pean era, beginning in 1788 (Anderson 2003). However,
various pieces of evidence ascribed to origins in Tonga or
New-Zealand (e.g., pieces of wrecked canoes, adzes made
of local basalt and other wooden artefacts), as well as
results of analyses of genetic variation in the Pacific rat
(Rattus exulans) suggesting connectivity between Norfolk
Island and New Zealand populations (Matisoo-Smith et al.
432101
01234
Density
02468
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Log(N)
Density
02468
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Log(t)
Density
N1
N0
ab
c
Log(N0/N1)
Fig. 5 Population size change in P. solandri using coalescent
modeling of microsatellite data under MsVar v0.4 and Msvar v1.3.
aposterior density distributions of the effective population size
parameter Log(N
0
/N
1
) from MsVar v0.4 where 0 indicates population
stability, \0 decline, and [0 expansion. Dotted curves represent the
linear model and continuous curves represent the exponential model.
The vertical solid line represents the expected value of Log(N
0
/N
1
)
when the effective population size is stable. The straight horizontal
dotted line represents the distribution of priors for comparison.
bposterior density distributions of the current (N
0
,solid lines) and the
ancestral (N
1
,dotted lines) effective population size parameter
Log(N) using MsVar v1.3 under the exponential model. cposterior
density distributions of the time parameter (Log(t), solid lines) since
Providence petrels started to decline on Lord Howe Island using
MsVar v1.3 under the exponential model. The inferior dotted lines in
fig. band crepresent the prior distributions of each parameter for
comparison
Conserv Genet
123
2001), constituted proof of Norfolk Island having been
settled from New Zealand at about the thirteenth to four-
teenth century (Anderson and White 2001; McCarthy
1934). Assuming that the Lord Howe colony was con-
nected to the Norfolk Island colony (i.e., panmixia), the
commencement of the bottleneck may be explained by the
introduction Pacific rats or kiore (Rattus exulans) on Nor-
folk Island 600 year B.P., as kiore is well known for having
affected seabird species on other islands (Holdaway 1999;
Rayner et al. 2007; Tennyson and Martinson 2006; Towns
2009). Polynesians may have also directly exploited the
Norfolk population, as it has been seen elsewhere (Boes-
senkool et al. 2009; Holdaway and Jacomb 2000; Worthy
1999). Additionally or alternatively, given the arrival of
Polynesians in New Zealand 700 year B.P. (Wilmshurst
and Higham 2004), they may have also encountered Lord
Howe Island at the same period. They may not have settled,
which could explain lack of archaeological evidence, but
allowed kiore (Rattus exulans) to colonise. Kiore may then
have disappeared after the introduction of the ship rat
(Rattus rattus) in 1918 (Hindwood 1940). However, there
is no evidence for Kiore ever having occupied Lord Howe
Island.
Conservation implications
The local extirpation of Providence petrels has had a severe
impact on the terrestrial ecosystem of Norfolk Island,
particularly through the deficiency of phosphorus leading
to Norfolk Island pines (Araucaria heterophylla) being
highly affected by the root-rotting fungus Phellinus noxius
(Holdaway and Christian 2010). To reduce the extinction
risk of Providence petrels and to provide key nutrients for
the regeneration of threatened native forests and associated
species, a plan to re-establish a colony of Providence pet-
rels on Norfolk Island using Lord Howe Island individuals
has been proposed. Here we show that the small colony of
Providence petrels breeding on Phillip Island is genetically
connected to the Lord Howe Island colony. These results
indicate limited risks surrounding the proposed transloca-
tion of Lord Howe Island individuals to re-establish a
colony on Norfolk Island with respect to potential genetic
novelty of the Phillip Island colony. In addition, as
colonisation of Phillip Island has been recent, further gene
flow will likely occur from Lord Howe Island to the Nor-
folk Island group, including the new translocated colony,
reducing risks of inbreeding depression following translo-
cation. While kiore is still present on Norfolk Island, this
was not the proximate cause for Providence petrel extinc-
tion from Norfolk Island. There is no obvious threat to
other avian species on the island through this
reintroduction.
Acknowledgments We are grateful to David Binns, Norfolk Island
Parks and Wildlife for providing logistics, transportation and
accommodation during field-work on Norfolk Island. We thank
Honey McCoy for sharing his data on Phillip Island seabirds. The Sea
world Research and Rescue Foundation Inc (Grant SWR/4/2011)
supported this work. Field sample collection of Providence petrel
blood samples was conducted with Animal Ethics permission from
University of Tasmania Ethics Committee (Permit# A00011680). We
thank the anonymous reviewers for their careful reading of our paper.
References
Anderson A (2003) Investigating early settlement on Lord Howe
Island. Aust Archaeol 57:98–102
Anderson A, White JP (2001) The Prehistoric Archaeology of Norfolk
Island. Southwest Pacific, Australian Museum, Sydney
Avise JC (1996) Introduction: the scope of conservation genetics.
conservation genetics. Springer, US, pp 1–9
Avise JC (2000) Phylogeography: the history and formation of
species. Harvard University Press, Cambridge
Axelsson E, Smith NG, Sundstro
¨m H, Berlin S, Ellegren H (2004)
Male-biased mutation rate and divergence in autosomal,
Z-linked and W-linked introns of chicken and turkey. Mol Biol
Evol 21:1538–1547
Backstro
¨m N, Fagerberg S, Ellegren H (2008) Genomics of natural
bird populations: a gene-based set of reference markers evenly
spread across the avian genome. Mol Ecol 17:964–980
Baduini CL (2002) Parental provisioning patterns of wedge-tailed
shearwaters and their relation to chick body condition. Condor
104:823–831
Beaumont MA (1999) Detecting population expansion and decline
using microsatellites. Genetics 153:2013–2029
Bester AJ (2003) The Breeding, Foraging Ecology and Conservation
of the Providence Petrel, Pterodroma solandri, on Lord Howe
Island, Australia. Ph.D. Thesis, Charles Sturt University, Albury,
NSW
Bester A, Klomp NI, Priddel D, Carlile N (2002) Chick-provisioning
behaviour of the Providence Petrel, Pterodroma solandri. Emu
102:297–304
Boessenkool S, Austin JJ, Worthy TH, Scofield P, Cooper A, Seddon
PJ, Waters JM (2009) Relict or colonizer? Extinction and range
expansion of penguins in southern New Zealand. P Roy Soc
Lond B Bio 276:815–821
Bonadonna F, Bretagnolle V (2002) Smelling home: a good solution
for burrow-finding in nocturnal petrels? J Exp Biol
205:2519–2523
Bowler DE, Benton TG (2005) Causes and consequences of animal
dispersal strategies: relating individual behaviour to spatial
dynamics. Biol Rev 80:205–225
Brooke M (2004) Albatrosses and petrels across the world. Oxford
University Press, Oxford
Brown RM et al (2010) Range expansion and hybridization in Round
Island petrels (Pterodroma spp.): evidence from microsatellite
genotypes. Mol Ecol 19:3157–3170
Burg T, Croxall J (2001) Global relationships amongst black-browed
and grey-headed albatrosses: analysis of population structure
using mitochondrial DNA and microsatellites. Mol Ecol
10:2647–2660
Calin
´ski T, Harabasz J (1974) A dendrite method for cluster analysis.
Commun Stat A-Theor 3:1–27
Charlesworth D, Charlesworth B (1987) Inbreeding depression and its
evolutionary consequences. Annu Rev Ecol Syst 18:237–268
Clarke K, Warwick R (2005) Primer-6 computer program. Natural
Environment Research Council, Plymouth
Conserv Genet
123
Clement M, Posada D, Crandall KA (2000) TCS: a computer program
to estimate gene genealogies. Mol Ecol 9:1657–1659
Dell’Ariccia G, Bonadonna F (2013) Back home at night or out until
morning? Nycthemeral variations in homing of anosmic Cory’s
shearwaters in a diurnal colony. J Exp Biol 216:1430–1433
Dellicour S, Mardulyn P (2014) SPADS 1.0: a toolbox to perform
spatial analyses on DNA sequence data sets. Mol Ecol Res
14:647–651
Dias MP, Granadeiro JP, Catry P (2012) Working the day or the night
shift? Foraging schedules of Cory’s shearwaters vary according
to marine habitat. Mar Ecol-Prog Ser 467:245
Earl DA (2012) STRUCTURE HARVESTER: a website and program
for visualizing STRUCTURE output and implementing the
Evanno method. Conserv Genet Resour 4:359–361
Edgar RC (2004) MUSCLE: multiple sequence alignment with high
accuracy and high throughput. Nucleic Acids Res 32:1792–1797
Edwards SV, Arctander P, Wilson AC (1991) Mitochondrial resolu-
tion of a deep branch in the genealogical tree for perching birds.
P Roy Soc Lond B Bio 243:99–107
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of
clusters of individuals using the software STRUCTURE: a
simulation study. Mol Ecol 14:2611–2620
Excoffier L, Lischer HE (2010) Arlequin suite ver 3.5: a new series of
programs to perform population genetics analyses under Linux
and Windows. Mol Ecol Res 10:564–567
Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular
variance inferred from metric distances among DNA haplotypes:
application to human mitochondrial DNA restriction data.
Genetics 131:479–491
Falk K, Benvenuti S, Dall’Antonia L, Gilchrist G, Kampp K (2002)
Foraging behaviour of thick-billed murres breeding in different
sectors of the North Water polynya: an inter-colony comparison.
Mar Ecol-Prog Ser 231:293–302
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
Frankham R (1996) Relationship of genetic variation to population
size in wildlife. Conserv Biol 10:1500–1508
Frankham R, Briscoe DA, Ballou JD (2002) Introduction to conser-
vation genetics. Cambridge University Press, Cambridge
Friesen VL (2015) Speciation in seabirds: why are there so many
speciesand why aren’t there more? J Ornithol 156:27–39
Friesen V, Smith A, Gomez-Diaz E, Bolton M, Furness R, Gonza
´lez-
Solı
´s J, Monteiro L (2007) Sympatric speciation by allochrony in
a seabird. P Natl Acad Sci USA 104:18589–18594
Fu Y-X, Li W-H (1993) Statistical tests of neutrality of mutations.
Genetics 133:693–709
Gelman A, Rubin DB (1992) Inference from iterative simulation
using multiple sequences. Stat Sci. 7:457–472
Girod C, Vitalis R, Leblois R, Fre
´ville H (2011) Inferring population
decline and expansion from microsatellite data: a simulation-
based evaluation of the Msvar method. Genetics 188:165–179
Goudet J (1995) FSTAT (version 1.2): a computer program to
calculate F-statistics. J Hered 86:485–486
Goudet J, Raymond M, de Meeu
¨s T, Rousset F (1996) Testing
differentiation in diploid populations. Genetics 144:1933–1940
Hardy OJ, Vekemans X (2002) SPAGeDi: a versatile computer
program to analyse spatial genetic structure at the individual or
population levels. Mol Ecol Notes 2:618–620
Hasegawa M, Kishino H, T-a Yano (1985) Dating of the human-ape
splitting by a molecular clock of mitochondrial DNA. J Mol Evol
22:160–174
Hedrick PW, Kalinowski ST (2000) Inbreeding depression in
conservation biology. Annu Rev Ecol Syst 31:139–162
Hermes N, Evans O, Evans B (1986) Norfolk Island birds: a review
1985. Notornis 33:141–149
Hey J, Nielsen R (2004) Multilocus methods for estimating popula-
tion sizes, migration rates and divergence time, with applications
to the divergence of Drosophila pseudoobscura and D. persim-
ilis. Genetics 167:747–760
Hey J, Nielsen R (2007) Integration within the Felsenstein equation
for improved Markov chain Monte Carlo methods in population
genetics. P Natl Acad Sci USA 104:2785–2790
Hindwood KA (1940) The birds of Lord Howe Island. Emu 40:1–86
Holdaway RN (1999) Introduced predators and avifaunal extinction in
New Zealand. Extinctions in near time. Springer, US, pp 189–238
Holdaway R, Christian M (2010) Stopping the fourth wave:
conservation and restoration of the Norfolk Island ecosystem.
Wingspan 20:30–35
Holdaway RN, Jacomb C (2000) Rapid extinction of the moas (Aves:
Dinornithiformes): model, test, and implications. Science
287:2250–2254
Hubisz MJ, Falush D, Stephens M, Pritchard JK (2009) Inferring
weak population structure with the assistance of sample group
information. Mol Ecol Res 9:1322–1332
Hudson RR, Kaplan NL (1985) Statistical properties of the number of
recombination events in the history of a sample of DNA
sequences. Genetics 111:147–164
Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching
and permutation program for dealing with label switching and
multimodality in analysis of population structure. Bioinformatics
23:1801–1806
Kalinowski ST (2005) hp-rare 1.0: a computer program for perform-
ing rarefaction on measures of allelic richness. Mol Ecol Notes
5:187–189
Keitt BS, Tershy BR, Croll DA (2004) Nocturnal behavior reduces
predation pressure on black-vented shearwaters Puffinus
opisthomelas. Mar Ornithol 32:173–178
Kimura M (1969) The number of heterozygous nucleotide sites
maintained in a finite population due to steady flux of mutations.
Genetics 61:893
Kimura M, Ohta T (1978) Stepwise mutation model and distribution
of allelic frequencies in a finite population. P Natl Acad Sci USA
75:2868–2872
Kocher TD, Thomas WK, Meyer A, Edwards SV, Pa
¨a
¨bo S,
Villablanca FX, Wilson AC (1989) Dynamics of mitochondrial
DNA evolution in animals: amplification and sequencing with
conserved primers. P Natl Acad Sci USA 86:6196–6200
Librado P, Rozas J (2009) DnaSP v5: a software for comprehensive
analysis of DNA polymorphism data. Bioinformatics
25:1451–1452
Loiselle BA, Sork VL, Nason J, Graham C (1995) Spatial genetic
structure of a tropical understory shrub, Psychotria officinalis
(Rubiaceae). Am J Bot 82:1420–1425
Lombal AJ, Wenner TJ, Burridge CP (2015) Assessment of high-
resolution melting (HRM) profiles as predictors of microsatellite
variation: an example in Providence Petrel (Pterodroma solan-
dri). Genes Genomics 37:977–983
Mardulyn P, Mikhailov YE, Pasteels JM (2009) Testing phylogeo-
graphic hypotheses in a Euro-Siberian cold-adapted leaf beetle
with coalescent simulations. Evolution 63:2717–2729
Matisoo-Smith E, Horsburgh K, Robins JH, Anderson A (2001) Genetic
variation in archaeological Rattus exulans remains from the Emily
Bay settlement site, Norfolk Island. Rec Aust Mus 53:81–84
McCarthy FD (1934) Norfolk Island: additional evidence of a former
native occupation. J Polyn Soc 43:267–270
McNeil R, Drapeau P, Pierotti R (1993) Nocturnality in colonial
waterbirds: occurrence, special adaptations, and suspected
benefits. Current ornithology. Springer, US, pp 187–246
Medway DG (2002a) History and causes of the extirpation of the
Providence petrel (Pterodroma solandri) on Norfolk Island.
Notornis 49:246–258
Conserv Genet
123
Medway DG (2002b) Why were Providence petrels (Pterodroma
solandri) nocturnal at Norfolk Island? Notornis 39:263–289
Meirmans PG, Van Tienderen PH (2004) GENOTYPE and
GENODIVE: two programs for the analysis of genetic diversity
of asexual organisms. Mol Ecol Notes 4:792–794
Nei M (1987) Molecular evolutionary genetics. Columbia University
Press, New York
Nielsen R, Wakeley J (2001) Distinguishing migration from isolation:
a Markov chain Monte Carlo approach. Genetics 158:885–896
Oro D (2003) Managing seabird metapopulations in the Mediter-
ranean: constraints and challenges. Sci Mar 67:13–22
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. Mol Ecol
13:55–65
Paiva VH, Geraldes P, Ramı
´rez I, Meirinho A, Garthe S, Ramos JA
(2009) Foraging plasticity in a pelagic seabird species along a
marine productivity gradient. Mar Ecol-Prog Ser 398:259–274
Palsboll PJ, Zachariah Peery M, Berube M (2010) Detecting
populations in the ambiguous zone: kinship-based estimation
of population structure at low genetic divergence. Mol Ecol Res
10:797–805
Patterson S, Morris-Pocock J, Friesen V (2011) A multilocus
phylogeny of the Sulidae (Aves: Pelecaniformes). Mol Phylo-
genet Evol 58:181–191
Petit RJ, El Mousadik A, Pons O (1998) Identifying populations for
conservation on the basis of genetic markers. Conserv Biol
12:844–855
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. J Hered 95:536–539
Plummer M, Best N, Cowles K, Vines K (2006) CODA: convergence
diagnosis and output analysis for MCMC. R News 6:7–11
Priddel D, Carlile N, Evans O, Evans B, McCoy H (2010) A review of
the seabirds of Phillip Island in the Norfolk Island Group.
Notornis 57:113–127
Pritchard JK, Stephens M, Donnelly P (2000) Inference of population
structure using multilocus genotype data. Genetics 155:945–959
Rannala B, Mountain JL (1997) Detecting immigration by using
multilocus genotypes. P Natl Acad Sci USA 94:9197–9201
Rayner MJ, Hauber ME, Imber MJ, Stamp RK, Clout MN (2007)
Spatial heterogeneity of mesopredator release within an oceanic
island system. P Natl Acad Sci USA 104:20862–20865
Reed TE, Warzybok P, Wilson AJ, Bradley RW, Wanless S, Sydeman
WJ (2009) Timing is everything: flexible phenology and shifting
selection in a colonial seabird. J Anim Ecol 78:376–387
Rosenberg NA (2004) DISTRUCT: a program for the graphical
display of population structure. Mol Ecol Notes 4:137–138
Sæther BE et al (2004) Life-history variation predicts the effects of
demographic stochasticity on avian population dynamics. Am
Nat 164:793–802
Serrano D, Forero MG, Dona
´zar JA, Tella JL (2004) Dispersal and
social attraction affect colony selection and dynamics of lesser
kestrels. Ecology 85:3438–3447
Silva MC, Duarte MA, Coelho MM (2011) Anonymous nuclear loci
in the white-faced storm-petrel Pelagodroma marina and their
applicability to other Procellariiform seabirds. J Hered
102:362–365
Slatkin M (1995) A measure of population subdivision based on
microsatellite allele frequencies. Genetics 139:457–462
Smith BJ (2007) boa: an R package for MCMC output convergence
assessment and posterior inference. J Stat Softw 21:1–37
Smith A, Friesen V (2007) Differentiation of sympatric populations of
the band-rumped storm-petrel in the Galapagos Islands: an
examination of genetics, morphology, and vocalizations. Mol
Ecol 16:1593–1603
Stephens M, Smith NJ, Donnelly P (2001) A new statistical method
for haplotype reconstruction from population data. Am J Hum
Genet 68:978–989
Storz JF, Beaumont MA (2002) Testing for genetic evidence of
population expansion and contraction: an empirical analysis of
microsatellite DNA variation using a hierarchical Bayesian
model. Evolution 56:154–166
Storz JF, Beaumont MA, Alberts SC (2002) Genetic evidence for
long-term population decline in a savannah-dwelling primate:
inferences from a hierarchical Bayesian model. Mol Biol Evol
19:1981–1990
Tajima F (1983) Evolutionary relationship of DNA sequences in finite
populations. Genetics 105:437–460
Tennyson AJD, Martinson P (2006) Extinct birds of New Zealand. Te
Papa Press, Wellington
Towns DR (2009) Eradications as reverse invasions: lessons from
Pacific rat (Rattus exulans) removals on New Zealand islands.
Biol Invasions 11:1719–1733
Venables WN, Smith DM (2001) An Introduction to R. Notes on R: a
programming environment for data analysis and graphics.
J Environ Qual
Warham J (1990) The petrels: their ecology and breeding systems.
A&C Black
Watanuki Y (1986) Moonlight avoidance behavior in Leach’s storm-
petrels as a defense against slaty-backed gulls. Auk 103:14–22
Weir J, Schluter D (2008) Calibrating the avian molecular clock. Mol
Ecol 17:2321–2328
Welch A et al (2012) Population divergence and gene flow in an
endangered and highly mobile seabird. Heredity 109:19–28
Whitlock MC (2011) GST and D do not replace FST. Mol Ecol
20:1083–1091
Wiley AE et al (2012) Foraging segregation and genetic divergence
between geographically proximate colonies of a highly mobile
seabird. Oecologia 168:119–130
Wilmshurst JM, Higham TF (2004) Using rat-gnawed seeds to
independently date the arrival of Pacific rats and humans in New
Zealand. Holocene 14:801–806
Worthy TH (1999) What was on the menu? Avian extinction in New
Zealand. New Zeal J Archaeol 19:125–160
Conserv Genet
123
... Although P. solandri was considered extirpated within the Norfolk Island group, a small population (* 20 breeding pairs) was discovered on Phillip Island, 7 km south of Norfolk Island, in 1986(Hermes et al. 1986 (Fig. 1). Genetic analysis revealed this population was recently founded by individuals from Lord Howe Island (Lombal et al. 2017), 900 km southwest of Norfolk Island (Fig. 1), representing the only substantial contemporary breeding locality of this species (* 32,000 breeding pairs; Bester 2003). Despite suggestions of genetic connectivity based on the Phillip Island population, it is not known whether the Norfolk extirpation represented loss of genetic or even cryptic species diversity (see review Ramakrishnan and Hadly 2009). ...
... A phylogeny was built using Cytochrome b sequences of modern P. solandri samples (n = 176; Lombal et al. 2017) and 22 ancient samples that were successfully sequenced. Homologous data were included from P. pycrofti (GenBank accession: MH828447), P. neglecta (GenBank accession: U74341), P. nigripennis (GenBank accession: U74343), P. cervicalis (GenBank accession: EU979553), Ardenna pacifica (GenBank accession: AF076088) and A. carneipes (GenBank accession: KY443837) given that they breed or are thought to have previously bred on Norfolk Island (Hermes et al. 1986;Holdaway and Anderson 2001). ...
... Even in instances when a genetically distinct seabird population is extirpated, genetic diversity may be transferred from declining populations to larger colonies, which present greater social attraction for juveniles prospecting for nest sites (Welch et al. 2012). While mtDNA represents one of the most sensitive markers to detect changes in genetic diversity through time, it is only a single, maternally-inherited, locus, and genetic variation of adaptive significance in the nuclear genome may have been lost, although we consider this unlikely, and the species has been able to re-establish on Phillip Island, within the Norfolk group (Lombal et al. 2017). ...
Article
Full-text available
The largest anthropogenic extinction events during the Holocene occurred on Pacific islands, where thousands of bird populations were lost. Although ancient DNA approaches have become widely used to monitor the genetic variability of species through time, few studies have been conducted to identify the potential cryptic loss of genetic and species diversity within Pacific seabird species. Here we used heterochronous sampling of mitochondrial DNA (Cytochrome b) in the genus Pterodroma from Norfolk Island to quantify potential loss of genetic and species diversity. We particularly focused on the providence petrel P. solandri whose main breeding colony (~ 1,000,000 breeding pairs) became extirpated from Norfolk Island following European settlement circa 1800. We sampled subfossil bones consistent with Pterodroma spp. from Norfolk Island, and performed genetic comparisons with other populations of P. solandri and congeneric species. The majority of subfossil Norfolk Island individuals exhibited the most common mitochondrial haplotype from Lord Howe Island P. solandri, suggesting no appreciable loss of genetic variation as a consequence of the Norfolk Island extirpation. Our findings provide an example where a large seabird population was rapidly extirpated by humans without loss of species-level genetic diversity, probably as a consequence of high connectivity with other populations. However, past connectivity was insufficient to prevent the extirpation itself, which has conservation implications for predicting the resilience of threatened seabirds. In contrast, ancient DNA analyses of smaller Pterodroma bones from Norfolk Island indicate the loss of a second species, potentially P. pycrofti, P. brevipes or another closely related, possibly undescribed taxon, from the Tasman Sea.
... Although most seabird species have the ability to travel long distances (Shaffer et al. 2006), they also usually exhibit a high level of philopatry (Coulson 2001;Warham 1990), which appears to restrict gene flow among colonies in some species (Friesen et al. 2007;Friesen 2015). However, a number of seabird species showing philopatry do not present genetic structure among colonies (Austin et al. 1994;Avise et al. 1992;Lombal et al. 2016;Pearce et al. 2004;Roeder et al. 2001), such that philopatry is not always a predictor of population genetic structure or has not been accurately quantified in those species. Seabirds from different colonies also often display discrete foraging distributions during breeding or non-breeding seasons that may limit gene flow among populations and promote local differentiation (Catard et al. 2000;Peck and Congdon 2005). ...
... Differences in foraging distribution during the breeding season have also been reported. Individuals breeding east of Australia are believed to forage in more inshore waters (<1000 km from land, Reid et al. 2012) and at a higher trophic level than individuals breeding on islands in Western Australia (Bond and Lavers 2014;Lindsey 1986;Taylor and Unit 2000). ...
... Individuals breeding east of Australia may forage in more inshore waters (<1000 km, Reid et al. 2012) and at a higher trophic level than individuals breeding at western colonies, believed to forage in offshore waters (Bond and Lavers 2014;Lombal, unpublished data). This may be explained by El Niño-Southern Oscillations (ENSO) events affecting the Cape Leeuwin Current near the western coast of Australia and lower associated prey availability, and the increase of industrial fishing in this region (Bond and Lavers 2014;Cheung et al. 2012;Lindsey 1986;Taylor and Unit 2000). These observations are compatible with distinct foraging distributions during the breeding season affecting gene flow among Flesh-footed Shearwater colonies, which is consistent with Friesen (2015) showing that 91% of seabird species that either feed inshore or have population-specific foraging areas show some evidence of restriction in gene flow among colonies. ...
Article
Full-text available
Increasing evidence suggests foraging segregation as a key mechanism promoting genetic divergence within seabird species. However, testing for a relationship between population genetic structure and foraging movements among seabird colonies can be challenging. Telemetry studies suggest that Flesh-footed Shearwater Ardenna carneipes that breed at Lord Howe Island or New Zealand, versus southwestern Australia or Saint-Paul Island in the Indian Ocean, migrate to different regions (North Pacific Ocean and northern Indian Ocean, respectively) during the non-breeding season, which may inhibit gene flow among colonies. In this study, we sequenced a 858-base pair mitochondrial region and seven nuclear DNA fragments (352–654 bp) for 148 individuals to test genetic differentiation among colonies of Flesh-footed Shearwaters. Strong genetic divergence was detected between Pacific colonies relative to those further West. Molecular analysis of fisheries’ bycatch individuals sampled in the Sea of Japan indicated that individuals from both western and eastern colonies were migrating through this area, and hence the apparent segregation of the non-breeding distribution based on telemetry is invalid and cannot contribute to the population genetic structure among colonies. The genetic divergence among colonies is better explained by philopatry and evidence of differences in foraging strategies during the breeding season, as supported by the observed genetic divergence between Lord Howe Island and New Zealand colonies. We suggest molecular analysis of fisheries’ bycatch individuals as a rigorous method to identify foraging segregation, and we recommend the eastern and western A. carneipes colonies be regarded as different Management Units.
... When the Fst could not be calculated based on mitochondrial DNA (mtDNA) sequences available in GenBank, it was reported as obtained in genetic studies (see Table 1). Lombal et al. (2018); 2, Genovart et al. (2007); 3, Austin et al. (1994); 4, Gómez-Díaz et al. (2009); 5, Burg & Croxall (2004); 6, Burg et al. (2003); 7, Quillfeldt et al. (2017); 8, Cagnon et al. (2004); 9, Techow et al. (2010); 10, Techow et al. (2010); 11, ; 12, Bicknell et al. (2012); 13, Quillfeldt et al. (2017); 14, Quillfeldt et al. (2017); 15, Ovenden et al. (1991); 16 Silva et al. (2015); 17, Young (2010); 18, Walsh & Edwards (2005); 19, Techow et al. (2009); 20, Brown et al. (2010); 21, Rayner et al. (2011); 22, Gangloff et al. (2013); 23, Welch et al. (2011); 24, Wiley et al. (2012); 25, Lombal et al. (2017); 26, Abbott & Double (2003b); 27, Burg & Croxall (2001); 28, Burg & Croxall (2001); 29, Abbott & Double (2003b); 30, Steeves et al. (2003); 31, Levin & Parker (2012); 32, Morris-Pocock et al. (2010); 33, Taylor (2011a); 34, Morris-Pocock et al. (2010); 35, Taylor et al. (2011b); 36, Barlow et al. (2011); 37, Calderón et al. (2014); 38, Mercer et al. (2013); 39, Marion & Le Gentil (2006); 40, Calderón et al. (2014); 41, Younger et al. (2015); 42, Clucas et al. (2016); 43, Boessenkool et al. (2009b); 44, Banks et al. (2006); 45, Grosser et al. (2015); 46, Ritchie et al. (2004); 47, Clucas et al. (2014); 48, Clucas et al. (2014); 49, Bouzat et al. (2009); 50, Sonsthagen et al. (2012); 51, Liebers et al. (2001); 52, Sonsthagen et al. (2012); 53, Liebers & Helbig (2002); 54, Sonsthagen et al. (2012); 55, Sonsthagen et al. (2012); 56, Pons et al. (2013); 57, Patirana et al. (2002); 58, Yeung et al. (2009); 59, Faria et al. (2010); 60, Miller et al. (2013); 61, Draheim et al. (2010); 62, Avise et al. (2000); 63, Pshenichnikova et al. (2015); 64, Pshenichnikova et al. (2017); 65, Moum & Arnason (2001); 66, Wojczulanis-Jakubas et al. (2015); 67, Birt et al. (2011b); 68, Friesen et al (1996b), 69, Wallace et al. (2014); 70, Pearce et al. (2002) 71, Birt et al. (2011a); 72, Morris-Pocock et al. (2008); 73, Tigano et al. (2015). Sample sizes implemented in the generalized linear models (GLMs) were adjusted to the number of sequences available in GenBank as used in the calculation of F-statistics where this differed from the number of sequences reported in the publication. ...
... Morphological and phenological differences among populations inhabiting different environments may be a cause (e.g. through mate choice) or a consequence of genetic isolation. In seabirds morphological and phenological differences are not always accompanied by genetic structure (Liebers & Helbig, 2002;Wiley et al., 2012;Lombal et al., 2017), especially at high latitudes (Moum & Arnason, 2001). For example, in the lesser black-backed gull, the divergence between Larus fuscus heuglini and L. f. fuscus is reflected in behavioural and ecological segregation, but reciprocal monophyly is lacking for mtDNA (Liebers & Helbig, 2002). ...
Article
Full-text available
Elucidating the factors underlying the origin and maintenance of genetic variation among populations is crucial for our understanding of their ecology and evolution, and also to help identify conservation priorities. While intrinsic movement has been hypothesized as the major determinant of population genetic structuring in abundant vagile species, growing evidence indicates that vagility does not always predict genetic differentiation. However, identifying the determinants of genetic structuring can be challenging, and these are largely unknown for most vagile species. Although, in principle, levels of gene flow can be inferred from neutral allele frequency divergence among populations, underlying assumptions may be unrealistic. Moreover, molecular studies have suggested that contemporary gene flow has often not overridden historical influences on population genetic structure, which indicates potential inadequacies of any interpretations that fail to consider the influence of history in shaping that structure. This exhaustive review of the theoretical and empirical literature investigates the determinants of population genetic differentiation using seabirds as a model system for vagile taxa. Seabirds provide a tractable group within which to identify the determinants of genetic differentiation, given their widespread distribution in marine habitats and an abundance of ecological and genetic studies conducted on this group. Herein we evaluate mitochondrial DNA (mtDNA) variation in 73 seabird species. Lack of mutation-drift equilibrium observed in 19% of species coincided with lower estimates of genetic differentiation, suggesting that dynamic demographic histories can often lead to erroneous interpretations of contemporary gene flow, even in vagile species. Presence of land across the species sampling range, or sampling of breeding colonies representing ice-free Pleistocene refuge zones, appear to be associated with genetic differentiation in Tropical and Southern Temperate species, respectively, indicating that long-term barriers and persistence of populations are important for their genetic structuring. Conversely, biotic factors commonly considered to influence population genetic structure, such as spatial segregation during foraging, were inconsistently associated with population genetic differentiation. In light of these results, we recommend that genetic studies should consider potential historical events when identifying determinants of genetic differentiation among populations to avoid overestimating the role of contemporary factors, even for highly vagile taxa.
... In the absence of genetic divergence, the question remains whether the observed morphological variation in the species is a result of plasticity or selective processes driving local adaptation. High behavioural plasticity has been observed in other Procellariiforms, for example, Gould's petrels and Providence petrels (Pterodroma solandri) populations are genetically similar despite ecological distinctiveness (Iglesias-Vasquez et al., 2017;Lombal et al., 2017). ...
Article
Aims Pleistocene glacial cycles have had profound effects on the distribution and genetic diversity of high latitude species, which can vary with species‐specific traits, such as vagility. Demographic responses of antarctic flying seabirds to the same events remain unassessed. We addressed this knowledge gap by studying the genetic population connectivity and demographic history of a flying seabird endemic to Antarctica, the Snow petrel. We hypothesize that their high vagility due to flight may represent an advantage over non‐flying seabirds in enduring past climate variation. Location Approximately 3,000 km of coastline in East Antarctica, covering three areas in Mac. Robertson Land, Princess Elizabeth Land and Wilkes Land. An inland location was also sampled at the Prince Charles Mountains, Mac. Robertson Land. Taxon Snow petrel (Pagodroma nivea). Methods We sampled 93 individuals and sequenced a total of 5,412 base pairs, including two mitochondrial genes, four anonymous nuclear loci and a nuclear intron. We used frequentist and Bayesian approaches to examine population genetic structuring and an Extended Bayesian Skyline Plot method to infer the demographic history of the species in the study area. In addition, evidence of exposed bedrock during glacial periods was summarized in maps of the studied area representing potential refugia for the species. Results Differentiation indexes, genetic clustering and haplotype networks suggest long‐term population connectivity for Snow petrels across the study area, with no evidence for reliction into refugia that were genetically isolated. Significantly, population expansions pre‐dated the Last Glacial Maximum (LGM), but only where there was evidence of ice‐free areas during this period. Main conclusions The high vagility of Snow petrels may have been advantageous for access to foraging areas and supported large populations despite the harsh conditions during the LGM. Our results highlight that species‐specific traits can exert a strong influence on demographic responses to the same environmental events.
Article
Full-text available
• Many seabird breeding colonies have recovered from heavy anthropogenic disturbance after conservation actions. The widely distributed red‐tailed tropicbird, Phaethon rubricauda, was used as a model species to assess potential anthropogenic impacts on the genetic diversity of breeding colonies in the Pacific Ocean. • Cytochrome c oxidase subunit I and control region sequences analyses were conducted across the range of the species in the Pacific Ocean. The study sites were at islands without human‐related disturbance (non‐impacted islands) and with human‐related disturbance (impacted islands). We hypothesized that (i) breeding colonies of the red‐tailed tropicbird on impacted islands have lower genetic diversity compared with colonies on non‐impacted islands, and (ii) breeding colonies of the red‐tailed tropicbird show significant fine and broad‐scale genetic structure across the Pacific Ocean. Bayesian skyline analyses were conducted to infer past changes in population sizes. • Genetic diversity was similar between impacted and non‐impacted islands. There was significant broad‐scale genetic structure among colonies separated by over 6,000 km, but a lack of significant fine‐scale genetic structure within Australasia and Hawai'i, although a significant level of differentiation was found within Chile with ΦST analyses. Skyline analyses showed that effective population sizes remained relatively constant through time, but experienced either a slight decrease or the end of an expansion event through the last 1,000 years. These changes may be related to the arrival of humans on Pacific islands. • Impacted islands may have received immigrants from other relatively close islands, buffering the loss of genetic diversity. However, it is also possible that colonies have retained ancestral variation or that a large effective population size coupled with a long generation time (13 years) has prevented the loss of genetic diversity in human‐impacted islands. Future research using higher‐resolution markers is needed to resolve the population genetic structure of the red‐tailed tropicbird in an ecological time‐scale.
Article
We describe a model-based clustering method for using multilocus genotype data to infer population structure and assign individuals to populations. We assume a model in which there are K populations (where K may be unknown), each of which is characterized by a set of allele frequencies at each locus. Individuals in the sample are assigned (probabilistically) to populations, or jointly to two or more populations if their genotypes indicate that they are admixed. Our model does not assume a particular mutation process, and it can be applied to most of the commonly used genetic markers, provided that they are not closely linked. Applications of our method include demonstrating the presence of population structure, assigning individuals to populations, studying hybrid zones, and identifying migrants and admixed individuals. We show that the method can produce highly accurate assignments using modest numbers of loci—e.g., seven microsatellite loci in an example using genotype data from an endangered bird species. The software used for this article is available from http://www.stats.ox.ac.uk/~pritch/home.html.
Article
Comparative analyses of avian population fluctuations have shown large interspecific differences in population variability that have been difficult to relate to variation in general ecological characteristics. Here we show that interspecific variation in demographic stochasticity, caused by random variation among individuals in their fitness contributions, can be predicted from a knowledge of the species’ position along a “slow‐fast” gradient of life‐history variation, ranging from high reproductive species with short life expectancy at one end to species that often produce a single offspring but survive well at the other end of the continuum. The demographic stochasticity decreased with adult survival rate, age at maturity, and generation time or the position of the species toward the slow end of the slow‐fast life‐history gradient. This relationship between life‐history characteristics and demographic stochasticity was related to interspecific differences in the variation among females in recruitment as well as to differences in the individual variation in survival. Because reproductive decisions in birds are often subject to strong natural selection, our results provide strong evidence for adaptive modifications of reproductive investment through life‐history evolution of the influence of stochastic variation on avian population dynamics.
Article
A survey of unconsolidated sediments overlying Pleistocene calcarenites and Tertiary basalts on Lord Howe Island was undertaken in 1996 in order to test the hypothesis that human settlement had not occurred before the European era, beginning in AD 1788. The results, largely from augering in lowland areas suitable for settlement, showed almost no sign of human occupation, and two radiocarbon dates on charcoal from sand-dune deposits are both modern. As the original argument stands, the remains of the AD 1834 colony may now be used in comparative analysis of initial island colonization.
Article
The population of Providence petrels (Pterodroma solandri) that nested on Norfolk Island at the time of 1st European settlement of that island in 1788 was probably > 1 million pairs. Available evidence indicates that Europeans harvested many more Providence petrels in the years immediately after settlement than previously believed. About 1,000,000 Providence petrels, adults and young, were harvested in the 4 breeding seasons from 1790 to 1793 alone. Despite these enormous losses, many Providence petrels were apparently still nesting on Norfolk Island in 1795 when they are last mentioned in documents from the island. However, any breeding population that may have survived there until 1814 when Norfolk Island was abandoned temporarily was probably exterminated by the combined activities of introduced cats and pigs which had become very numerous by the time the island was re-occupied in 1825.