Molecular Ecology (2009) 18, 4256–4269
doi: 10.1111/j.1365-294X.2009.04280.x
Range persistence during the last glacial maximum:
Carex macrocephala was not restricted to glacial refugia
M A T T H E W G . K I N G , * M A T T H E W E . H O R N I N G † and E R I C H . R O A L S O N ‡ *Department of Botany, University of British Columbia, Vancouver, BC, Canada V6T1Z4, †USDA Forest Service, Pacific
Northwest Research Station, Corvallis, OR 97331, USA, ‡School of Biological Sciences, Washington State University, Pullman,
WA 99164, USA
Abstract
The distribution of many species inhabiting northwestern North America has been
heavily influenced by the climatic changes during the late Pleistocene. Several studies
have suggested that species were restricted to glacial refugia north and ⁄ or south of the
continental ice sheet front. It is also hypothesized that the coast of northwestern North
America could have been a prime location for glacial refugia because of the lowering of
the eustatic sea level and the concomitant rise of the continental shelf because of tectonic
rebound. Alternatively, some coastal species distributions and demographics may have
been unaffected in the long-term by the last glacial maximum (LGM). We tested the
glacial refugium hypothesis on an obligate coastal plant species, Carex macrocephala by
sampling 600 individuals from 41 populations with 11 nuclear microsatellite loci and the
rpL16 plastid intragenic spacer region. The microsatellite data sets suggest a low level of
population differentiation with a standardized G¢ST = 0.032 and inbreeding was high
with an F = 0.969. The homogenization of the populations along the coast was supported
by a principal coordinate analysis, AMOVAs and SAMOVA analyses. Analyses using the
rpL16 data set support the results of the microsatellite analyses, with a low FST of 0.042.
Coalescent and mismatch analyses using rpL16 suggest that C. macrocephala has not
gone through a significant bottleneck within the past 100 000 years, although a much
earlier population expansion was indicated by the mismatch analysis. Carex macrocep­
hala exhibits the characteristics of metapopulation dynamics and on the basis of these
results, we concluded that it was not restricted to glacial refugia during the LGM, but
that it existed as a large metapopulation.
Keywords: coastal refuge, Cyperaceae, Pleistocene glaciation, population genetics
Received 25 July 2008; revision received 27 April 2009; accepted 5 May 2009
Introduction
Identifying phylogeographical patterns and inferring
the process of diversification following the last glacial
maximum (LGM) are important for understanding the
distribution of organismal diversity today. The north­
western portion of North America is an important geo­
graphical region that has been the focus of many
phylogeographical studies, where the specific goal has
been to identify the locations of potential glacial refugia
during the LGM. Several of these investigations have
Correspondence: Matthew G. King, Fax: 604 822 6089;
E-mail: kingdom@interchange.ubc.ca
focused on species with large biogeographical disjuncts,
i.e. noncontiguous distributions such as the tailed-frog
Ascaphus truei (Nielson et al. 2001); the Tamias amoenus
species complex (Demboski & Sullivan 2003); and Pleth­
odon vandykei (Carstens et al. 2004). Phylogeographical
studies have also been conducted on species native to
montane regions of the Cascade and Rocky Mountains
such as Sorex monticolus (Demboski & Cook 2001); chip­
munks, Tamias, (Good et al. 2003); and the Idaho giant
salamander Dicamptodon aterrimus (Carstens et al. 2005;
see Brunsfeld et al. 2001 for more examples). Potential
refugia have been suggested for areas south of the
continental ice sheet boundary for Tamias (Demboski &
Sullivan 2003), Dicamptodon (Steele & Storfer 2006) and
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4257
several other species (Brunsfeld et al. 2001), but some
studies have suggested colonization from refugia north
of the continental ice sheet maximum such as the
ermine, Mustela erminea (Fleming & Cook 2002); Canis
lupis (Weckworth et al. 2005); Martes and Sorex (Dembo­
ski et al. 1999). Phylogeographical studies along the
western coast of North America are also important for
inferring potential postglacial migration routes of
humans (Hetherington et al. 2003). These studies have
suggested that the potential refugia were located along
the Pacific coast; however, those taxa of interest are not
obligate coastal species and therefore speculation
remains as to the precise location of the possible coastal
refugia. Nonobligate coastal species may have persisted
in inland refugia and recolonized coastal areas follow­
ing the retreat of the ice sheet.
Clearly, the LGM had a drastic impact on most spe­
cies with distributions in northern latitudes in western
North America and that this is the most tested hypothe­
sis, but there are important alternative hypotheses to
consider. An alternative is that species distributions and
demographics remained unchanged during the LGM
and that there is no significant genetic signature (e.g.
indicating a bottleneck) from that time period. This is
possible in many organisms south of the continental ice
sheet that did not undergo drastic range shifts because
of change in global climate. However, it is possible that
contemporary population genetic architecture may
mask the genetic signature of the LGM. By using a
coastal-obligate terrestrial species Carex macrocephala,
the large-headed sedge, as our model organism, we
attempt here to test explicitly the coastal refugium
hypothesis and examine the population genetic dynam­
ics of a coastal sedge.
Carex macrocephala is a low-lying sedge confined to
sandy beaches and sand dunes distributed along the
Pacific coast of North America and Asia. The species
range encompasses �16° of latitude along the North
American west coast from mid-Oregon north through
Alaska, the Aleutian Islands, across the Bering Sea to
the Kamchatka Peninsula and south along the east coast
of Asia to Hokkaido Island of Japan (Fig. 1). It spreads
across unstable sand dunes by extensive rhizome
growth (Mastrogiuseppe 2002). Carex macrocephala has
monoecious individuals, although typically, the shoots
are unisexual (e.g. paradioecy; Standley 1985; Pannell &
Verdú 2006). There are no reports of self-incompatibility
in Carex and selfing rates are most likely high because
of wind-mediated cross-pollination of the multiple sex­
ual ramets per individual. Populations of C. macrocephala
Fig. 1 Map of the 41 sampling localities
of Carex macrocephala from the north­
west coast of North America. Black stars
represent localities and dark grey out­
lines surrounding localities represent
the first AMOVA designations of groups.
© 2009 Blackwell Publishing Ltd
4258 M . G . K I N G , M . E . H O R N I N G and E . H . R O A L S O N
and its sister species Carex kobomugi can be established
by floating rhizome mats (personal observation) and
seeds have very low in situ germination rates (Ishikawa
et al. 1993).
King & Roalson (2009a) used statistical phylogeo­
graphical coalescent analyses to suggest C. macrocephala
has existed within North America for at least
125 000 ± 20 000 years. As C. macrocephala has persisted
in North America since before the LGM, we can infer
that there must have been populations harbouring
genetic and allelic diversity during that time (Widmer
& Lexer 2001), c. 18 000 years ago (Barendregt & DukRodkin 2004). It is assumed that range expansions or
contractions would leave identifiable genetic patterns,
which may allow us to infer the location of those refu­
gia (Hewitt 1996, 2004; Widmer & Lexer 2001). The
locations of the populations of C. macrocephala would
have changed during the LGM because of the presence
of the continental ice sheet along the coast, the decline
in eustatic sea level and depression of the continental
shelf caused by the weight of the ice sheet. We would
assume, under a standard biogeographical framework,
that C. macrocephala was geographically restricted
because of these factors and genetic diversity would
have been reduced as a result of bottleneck events dur­
ing concomitant population size reductions (Abbott
et al. 2000).
Any current estimate of population genetic structure
sampled from across a species range is the result of both
historical and contemporary population level processes
(e.g. drift, migration). Often, recent events mask older
population genetic signals. This is particularly evident
when examining the genetic structure of northern popula­
tions that are the descendants of Pleistocene glacial refuge
populations (Riddle 1996). When researching potential
glacial refuge locations, we often assume that current
genetic diversity within those locations is higher than in
the nonrefuge populations because of founder effects (He­
witt 1996; Widmer & Lexer 2001). However, this assump­
tion can be misleading in areas colonized from multiple
refuge populations (Widmer & Lexer 2001; Abbott &
Brochmann 2003; Petit et al. 2003). Nevertheless, many
analytical tools are available to infer past genetic structur­
ing, particularly when sampling is robust.
To evaluate the coastal glacial refuge hypothesis, we
have employed a multilocus approach using nuclear
microsatellites and a plastid intergenic spacer to
attempt to identify refugia of C. macrocephala along the
northwest coast of North America. Moreover, we char­
acterized contemporary population genetic structuring,
levels of inbreeding and gene flow. Through the use of
various population genetic techniques, we examined the
impact the LGM had on the distribution of C. macrocep­
hala along the Pacific coast of North America.
Materials and methods
Population sampling
Six hundred individuals of Carex macrocephala were col­
lected from 41 localities along the Pacific coast of North
America (Fig. 1, Table 1). Many of these localities were
identified using herbarium collections and these sites
represent nearly all of the known extant populations
along the North American coast. Leaf tissue was col­
lected from each locality in a manner to minimize the
probability of sampling the same genet. In populations
covering long stretches of beach, samples were collected
every 100–200 m over many kilometres. In populations
where more concentrated sampling was required to
collect enough samples, the extent of clonal spread
was observed by following rhizomes and collections
were spread as far apart as possible. Leaf tissue was
immediately desiccated in the field for optimal DNA
preservation.
Molecular techniques
Genomic DNA was extracted following a modified 2X
CTAB protocol using 30 mg of dried leaf material and
pulverized in liquid nitrogen using micropestels (Roal­
son et al. 2001). We used a cpDNA marker, the rpL16
intergenic spacer, previously shown to be variable
within populations of C. macrocephala along the North
American coast (King & Roalson 2009a). Amplification
protocols of the rpL16 spacer follow Shaw et al. (2005).
We used Eppendorf MasterTaq to minimize the poly­
merase chain reaction error rate as it has a higher fidel­
ity than other DNA Taq polymerases (product
literature; Eppendorf Corp.). Sequencing was conducted
at Washington State University’s Center for Integrated
Biotechnology sequencing core on an Applied Biosys­
tems 3730 automated DNA sequencer following ABI’s
protocol for Big Dye Terminators version 3.1. Forward
and reverse strand templates were sequenced, contigs
were created and edited using Sequencher version 4.6
(GeneCodes Corp). Sequences were manually aligned
in SE-AL version 2.0 (Rambaut 1996). Thirty-seven indi­
viduals representing 37 different haplotypes were
resequenced to verify the authenticity of the polymor­
phisms within the rpL16 spacer.
Nuclear microsatellite loci were previously character­
ized in King & Roalson (2009b). All loci comprise dinu­
cleotide repeats except CM16, CM35, and CM39 which
comprise trinucleotide repeats. These loci were devel­
oped from a size fractionated genomic library con­
structed from a single individual of C. macrocephala
from the coast of Washington state. Amplification of the
target loci followed the protocols outlined in King and
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4259
Table 1 Sampling localities of Carex macrocephala across the northwest coast of North America
Pop
Location
Latitude
Longitude
N
f
Allelic diversity
rpL16 diversity
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Mouth of Moose Creek, OR
Mouth of Beaver Creek, OR
Taft, OR
Siletz Bay, OR
Neskowin Beach, OR
Whalen Island South, OR
Whalen Island North, OR
Camp Magruder, OR
Rockaway Beach, OR
Del Rey Beach, OR
Camp Rilea Beach, OR
Leadbetter Point State Park 1, WA
Leadbetter Point State Park 2, WA
Midway Beach, WA
Westport Lighthouse State Park, WA
La Push Beach, WA
Crescent Beach, WA
Fort Worden State Park, WA
Deception Pass State Park, WA
Spencer Spit State Park, WA
Port Renfrew, BC
Bamfiled, BC
Long Beach 1 PRNP, BC
Long Beach 2 PRNP, BC
Mackenzie Beach PRNP, BC
Wickaninish Beach PRNP, BC
Oyster Beach, BC
San Josef Beach 1, BC
San Josef Beach 2, BC
Rose Spit, BC
Naikoon Provincial Park, BC
Tow Hill, BC
Tlell Beach, BC
Pasagshak Beach Kodiak Island, AK
Shelikof Beach Kruzof Island, AK
Shelikof Beach 2 Kruzof Island, AK
Yakutat Canon Beach, AK
Yakutat Coast Guard Beach, AK
Mouth of Kenai River, AK
Kalifornsky Beach, AK
Kasilof Beach, AK
44.36043333
44.52381667
44.92911667
44.9279
45.1
45.27326667
45.27865
45.58301667
45.626
46.04806667
46.11431667
46.52746667
46.60715
46.76908333
46.88728333
47.91581667
48.16228333
48.14805
48.3913
48.54036667
48.536
48.78685833
49.07245
49.07245
49.13321667
49.0191
49.8955
50.67446667
50.67446667
54.16965
54.1127
54.07225
53.57883333
57.45856667
57.16976667
57.16635
59.49281667
59.50995
60.57
61.52383333
60.38923333
124.0897167
124.0736167
124.01175
124.0236167
123.97
123.9502167
123.9503333
123.9648333
123.9435333
123.93095
123.9449333
124.04655
124.0431833
124.0941833
124.1234833
124.64235
123.707
122.7613
122.6474667
122.8553
124.4108333
125.173225
125.7668667
125.7668667
125.9026167
125.6730833
125.14685
128.2767667
128.2767667
131.6566
131.71455
131.79105
131.9318
152.45035
135.7559333
135.7560333
139.7271833
139.7740333
151.25
151.2694667
151.2965
10
10
7
3
10
8
7
4
5
20
9
15
10
10
8
10
10
10
10
7
29
26
7
20
20
20
29
30
20
26
30
30
30
10
10
15
15
10
10
10
20
0.937
0.928
0.924
1
0.849
0.938
1
0.915
0.964
0.893
0.945
0.851
0.892
0.904
1
0.941
0.977
0.833
0.909
0.843
0.895
0.826
0.889
0.921
0.904
0.964
0.887
0.863
0.952
0.908
0.895
0.905
0.94
0.917
0.947
0.987
0.935
0.69
0.914
0.984
0.945
0.12
0.31
0.42
0.55
0.24
0.39
0.39
0.57
0.47
0.24
0.36
0.19
0.29
0.23
0.23
0.25
0.25
0.22
0.15
0.26
0.17
0.16
0.47
0.23
0.21
0.20
0.13
0.15
0.22
0.18
0.16
0.16
0.16
0.34
0.34
0.23
0.25
0.25
0.31
0.31
0.22
0
0.8
0.9524
0.6667
0.8444
1
0.9524
0.5
1
0.9263
0.8056
0.7619
0.7778
0.8444
0.8571
0.9333
0.9333
0.8889
0.7778
0.8571
0.8547
0.7631
0.9048
0.7947
0.8474
0.8789
0.5493
0.9218
0.9105
0.9446
0.8023
0.9126
0.8897
0.5111
0.8
0.9143
0.7714
0.3778
0.7778
0.8222
0.5731
N refers to sample size, f is the microsatellite determined inbreeding coefficient of Weir & Cockerham (1984), microsatellite allelic
diversity is calculated as the number of alleles per individual per locus and the rpL16 haplotypic diversity.
Roalson (2008b) and amplified products were multi­
plexed with three other loci each with a different fluo­
rescent dye. Final dilution ratios of 1:40 were needed
and fragments were visualized on an Applied Biosys­
tems 3730 and scored using GeneMapper version 3.7
(Applied Biosystems). To minimize the number of mis­
called alleles, all individuals for all loci were scored
manually. Three localities, 10, 28 and 32, were reextracted and genotyped to assess genotyping error and
to confirm clonal identity (Bonin et al. 2004; ArnaudHaond et al. 2007).
© 2009 Blackwell Publishing Ltd
Clone identification
To ensure that we were only using unique genets in
further analyses, we first determined clonal identity
within each locality. The first method we used was to
excise all duplicate genotypes from the same location.
The second method we used was to conduct a statistical
analysis in GENCLONE 2.0 (Arnaud-Haond and Belkhir
2007) to determine identical multilocus genotypes
(MLGs) in the original data set. Additionally, GENCLONE
was used to assess the levels of clonal diversity within
4260 M . G . K I N G , M . E . H O R N I N G and E . H . R O A L S O N
each locality. When microsatellites show significant
departures from Hardy–Weinberg equilibrium (HWE),
it can be difficult to assess the clonal diversity within
each locality and caution is needed when interpreting
the results (Arnaud-Haond et al. 2007). GENCLONE is able
to calculate the probability of each MLG (pgen) in the
sample in two ways, one where HWE is assumed, and
one where the algorithm accounts for departures in
HWE by estimating FIS. Following the GENCLONE analy­
sis, all duplicate MLGs were excised from the data set.
Population genetic analyses of microsatellite data sets
We first analysed the nuclear microsatellite data set to
identify population genetic patterns and to identify
whether the loci are in HWE and linkage equilibrium.
FSTAT (Goudet 1995) was used to determine if the loci
were in HWE and to estimate levels of linkage disequi­
librium (LD). We then used FSTAT to calculate allelic
diversity, observed heterozygosity (HO), expected het­
erozygosity (HE), gene diversity (HT) and Nei’s genetic
differentiation (GST). We also used FSTAT to obtain sum­
mary population genetic indices of Wright’s F-statistics,
F, f and Q, following Weir & Cockerham (1984),
where F is an estimate of inbreeding across all subpop­
ulations, f is an estimate of inbreeding within each
subpopulation and Q is an estimate of differentiation
across all subpopulations (Wright 1951; Nei 1973).
Values for GST are highly dependent on the level of
variability and hence we used RECODEDATA (Meirmans
2006) to obtain a data set that would estimate the maxi­
mum GST-value possible and scaled our GST-value to
this maximum GST-value, as recommended by Hedrick
(2005). We also ran a MICRO-CHECKER (van Oosterhout
et al. 2004) analysis to determine the levels of null
alleles within our data set and to determine if the pres­
ence of null alleles may significantly contribute to the
observed patterns.
Arlequin v3.11 was used to test two hypotheses of
genetic structuring using both the rpL16 and microsatel­
lite data sets (Excoffier et al. 2005) by running AMOVA
analyses using two different a priori population group­
ings of the 41 localities (Excoffier et al. 1992). The first
structural grouping was based on the littoral cells off
the coast of North America and the major currents off
the mainland coast. Littoral cells are compartmentalized
areas formed by near shore currents immediately off
the coast causing most of the sand and debris to move
within the cell. For example, the Columbia River littoral
cell circulates 165 km from Tillamook head in Oregon
to Point Grenville in Washington where the majority of
the sediment of the Columbia River is deposited (Twic­
hell & Cross 2002). If seed or rhizome dispersal occurs
via the near shore currents, we hypothesized that most
of the movement of individuals would take place
within the major current systems. These systems are
stable through time and may cause population substructuring. Therefore, we grouped localities into a midOregon (populations 1–5), northern Oregon (6–9),
Columbia River littoral cell (10–14), Washington coast
(15–16), San Juan Islands (17–20), Vancouver Island (21–
29), Queen Charlotte Islands (QCI) (30–33), southeast
Alaska (34–37) and south-central Alaska (38–41). A sec­
ond AMOVA was conducted where the populations were
further lumped together into larger groups of the south­
ern coast (1–17), San Juan and Vancouver Islands (18–
29), QCI (30–33) and Alaska (34–41).
These a priori groupings in the AMOVA are synthetic
and may not represent the true structuring along the
coast. Therefore, we conducted a spatial analysis of
molecular variance (SAMOVA) implemented in SAMOVA 1.0
(Dupanloup et al. 2002) as a means to infer possible
other groups without user-defined structure parameters.
SAMOVA attempts to reconstruct genetic groups of locali­
ties using the same algorithm as AMOVA, but SAMOVA
allows for variable groupings. It accomplishes this by
fixing the number of groups (K) or demes and by con­
ducting simulations to determine the most significant
fixation index value, FCT, for that K-value. We explored
K-values between 2 and 10 with one hundred simulated
annealing simulations for each K.
Statistically, it may be difficult to infer the level of
genetic structuring using population genetic techniques
that require loci to be in linkage equilibrium and HWE.
As a means of demonstrating population structuring
beyond panmixia, we conducted a principal coordinate
analysis (PCoA) on inter-individual genetic distances as
estimated using the nuclear microsatellite loci. The
advantage of PCoA analyses is that it does not rely on
loci that are in HWE or linkage equilibrium. PCoA
reduces the dimensionality of the data set while main­
taining the variance and covariance between the sam­
ples. We account for the variance in each level of the
principal coordinates where the first principal coordi­
nate always has the most variance and the second prin­
cipal component has the second most variance, etc.
Individuals will group together according to their posi­
tion within each of the principal coordinates. PCoA
analyses were conducted using GENALEX version 6 (Peak­
all & Smouse 2006) and the first three principal axes
were visualized using GNUPLOT (Williams & Kelley
2004).
Statistical tests of population structuring using the
rpL16 data set
To test for population structuring in the rpL16 data set,
we ran an AMOVA under the same conditions as the
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4261
microsatellite analyses. By using the rpL16 plastid mar­
ker, we examined the gene flow due to the movement
of seeds or rhizome mats, as plastids are thought to be
maternally inherited in sedges (as in most other angio­
sperms) and not transferred through the pollen. We
also used Arlequin to calculate population haplotype
diversity and FST-levels.
There may be more appropriate groupings of these
populations than our a priori hypotheses; we therefore
conducted a SAMOVA to identify significant grouping of
populations with the rpL16 data set. We used the pro­
gram SAMOVA 1.0 and ran analyses from two groups to 10
groups with one hundred simulated annealing processes
for each analysis as we did for the microsatellite data set.
If C. macrocephala had experienced a significant bottle­
neck during the LGM, we would also have expected
North American populations to show signs of a popula­
tion expansion following the stabilization of eustatic sea
level. We used a mismatch analysis implemented in Ar­
lequin using the rpL16 data set, to determine if a signifi­
cant population expansion had taken place. The
mismatch analysis provided estimates of s, and h0, and
h1, where s = 2lt generations since expansion began,
and h = Nel for uniparentally inherited haplotypes.
Estimates of s allow us to determine when the last
major population expansion would have begun. Like­
wise, the parameters of h0, and h1 give us estimates of
the initial and final effective population size.
As another method to determine whether genetic
diversity was reduced during the LGM we used the
program SWEEP-BOTT to test for a bottleneck event (Galtier
et al. 2000). SWEEP-BOTT uses a coalescent-based likeli­
hood approach to determine the time and strength of
bottleneck events for sequence data under the infinite
sites model. Although we used only a single plastid
locus to estimate the timing and strength of the possible
bottleneck event, this method is still more powerful and
robust than Tajima’s D (Galtier et al. 2000). We also sta­
tistically tested whether a nonfounder event or bottle­
neck event was more likely. For example, we tested
whether C. macrocephala went through a significant bot­
tleneck during the LGM, or whether the more likely
model is one where C. macrocephala has never experi­
enced a bottleneck. We removed two haplotypes, AN
and AT, each with frequencies of <1%, to fit the data
set to the infinite sites model. We conducted initial runs
of 10 000 iterations for the first phase and 50 000 itera­
tions for the second phase to assess the prior range of
parameters for theta, h = Nel for uniparentally inherited
loci. The parameter h is an estimate of the genetic diver­
sity within the metapopulation. To estimate the time
since the bottleneck occurred, SWEEP-BOTT also estimates
the parameter, T in units of 2NE generations, where NE
is the effective population size. We also estimated the
© 2009 Blackwell Publishing Ltd
parameter S, the strength of bottleneck. After the initial
runs, prior range for T was set from 0.3 to 1 and the
range of S from 0.01 to 0.2. Final runs of 100 000 steps
were conducted to assess the maximum likelihood of T,
S, and to conduct the likelihood ratio test.
Results
Nuclear microsatellite diversity
Individuals from locations 10, 28 and 32 showed low
levels of clone duplication and no reproducible allelic
dropout (Bonin et al. 2004). As no allelic dropout was
observed within our test runs, duplicate genotypes
within localities were assumed to be samples of the
same genet and were removed from our analyses lead­
ing to a final overall sample size of 548 individuals (Ar­
naud-Haond et al. 2007). The 548 samples were
identical to the data set found after running the GEN­
CLONE analysis to determine identical MLGs. Significant
deviations from HWE were found in all loci in all pop­
ulations (see below); we therefore used the pgenFIS-esti­
mate of clonal identity (Arnaud-Haond et al. 2007). This
did not result in substantially different estimates of pgen
from any sample. In all 548 unique MLG cases, each
pgenFIS was <0.001. Simpson diversity indices (D*) for
each population were above 0.97 and this could repre­
sent an inflated estimate of clonal diversity because of
inbreeding from selfing and biparental pollination
between close relatives.
Observed heterozygosities at all microsatellite loci dif­
fered significantly from HWE after Bonferroni correc­
tions (all P-values < 0.0001; Rice 1989). All loci were in
significant LD with at least two other loci. Multilocus
estimates using these nuclear microsatellites revealed
an insignificant Q (Q = 0.013 ± 0.015, P = 0.13) among
all populations with significant values of f
(f = 0.956 ± 0.04, P < 0.0001) and F (F = 0.969 ± 0.03,
P < 0.0001). By standardizing the overall GST-value by
the maximum GST calculated from the RECODEDATA
analysis, we obtained a standardized estimate of
G¢ST = 0.032. Observed heterozygosity (HO) levels ran­
ged from zero to 0.127 with a HO = 0.041 across all mi­
crosatellite loci and are summarized in Table 2. The
inbreeding estimate, f, within each locality ranged from
0.69 in locality 38 to an f of 1 in several localities
(Table 1), where the per locus f ranged from 0.803 to 1
(Table 2). Per locus Q levels were low and ranged from
0.001 to 0.058, while the microsatellite genetic diversity
levels ranged from 0.251 to 0.843 (Table 2). Re-analysis
using the modified data set from MICRO-CHECKER did not
substantially change these results.
Results of the AMOVA analyses show that a majority of
variance is accounted for within localities for both
4262 M . G . K I N G , M . E . H O R N I N G and E . H . R O A L S O N
Table 2 Per locus genetic summaries of microsatellite data set where Q, F, and f are the fixation indices of Weir & Cockerham
(1984); HO is observed heterozygosity, HE is expected heterozygosity, HT is gene diversity and A is the number of alleles
Locus
GST
Q
F
f
HO
HE
HT
A
CM01
CM07
CM12
CM13
CM16
CM25
CM27
CM28
CM35
CM36
CM39
Overall
0.012
0.009
0.011
0.015
0.016
0.063
0.015
0.012
0.006
0.011
0.019
0.016
0.009
0.01
0.009
0.012
0.012
0.058
0.011
0.001
0.004
0.008
0.002
0.013
0.832
0.988
0.958
0.961
1
0.996
0.824
0.954
1
0.922
0.975
0.979
0.816
0.876
0.954
0.956
1
0.99
0.803
0.95
1
0.916
0.914
0.956
0.083
0.062
0.02
0.025
0
0.003
0.127
0.03
0
0.049
0.055
0.041
0.518
0.492
0.564
0.716
0.423
0.157
0.666
0.569
0.237
0.539
0.464
0.459
0.59
0.545
0.628
0.843
0.506
0.42
0.782
0.648
0.251
0.603
0.603
0.547
7
7
8
9
6
3
10
6
4
5
8
73
models of genetic structuring. According to the group­
ings based on littoral cells, a statistically significant
level of differentiation was detected within localities,
between localities within groups and between groups,
the majority of the observed variation is accounted for
within localities (Table 3). Similar results were found
when groupings were based on geographical proximity.
The results of our SAMOVA analysis on the microsatellite
loci are congruent with the AMOVA analysis. Weakly sig­
nificant FCT-values were obtained for demes ranging
from K = 4 to K = 10, with the highest FCTTable 3 Results of AMOVA analyses using nuclear microsatel­
lites and rpL16 chloroplast spacer
Between
groups
Littoral cell
Microsatellite
5.37%
FCT = 0.028
cpDNA rpl16
7.24%
FCT = 0.037
Geographical regions
Microsatellite
3.25%
FCT = 0.018
cpDNA rpl16
4.77%
FCT = 0.03
Between
populations
within groups
Within
populations
9.09%
FSC = 0.034
12.56%
FSC = 0.05
85.54%
FST = 0.059
80.2%
FST = 0.083
11.09%
FSC = 0.036
12.52%
FSC = 0.036
85.66%
FST = 0.053
82.7%
FST = 0.065
Top row of the comparison indicates the per cent variation,
while the bottom row indicates the F statistic analogue. For
each markers system, we present the per cent variation
explained and the corresponding test statistic for two a priori
structuring schemes. The ‘Littoral cell’ structuring scheme is
based on the littoral currents immediately off the coast of
North America, while the ‘Geographical regions’ structuring
scheme further groups the localities based on geographical
proximity. All levels of variation were significant, and all F
statistics were significant.
score = 0.031, P = 0.061. Although these FCT-scores were
marginally significant, they were not geographically
meaningful, i.e. adjacent localities were not grouping
together, whereas distant populations from Alaska and
Oregon were being grouped together.
The results of our PCoA support a large metapopula­
tion along the coast. In Fig. 3, we would expect to see
clusters of individuals if any genetic clustering were
observed. The first three principal coordinates of the
PCoA account for 59.22% of the observed variation with
the first three principal axes accounting for 25.52%,
17.34% and 16.35% respectively (Fig. 3). The figure
shows no clustering of individuals beyond the single cen­
tral cluster, i.e. no evidence of genetic clustering.
Plastid diversity
Sequence GenBank Accession nos for the rpL16 haplo­
types are FJ424626–FJ424702. A maximum-likelihood
phylogeny of the rpL16 haplotypes displays the
extensive variation present within this marker in Carex
macrocephala (Fig. 2; King & Roalson 2009a). The rpL16
data set had a marginally significant FSTvalue = 0.042 ± 0.022, P = 0.068). Within locality, allelic
diversity ranged from zero to one with a mean rpL16
diversity of 0.801 (Table 1).
The results of rpL16 AMOVA were similar to that of the
AMOVA conducted using the microsatellite data set
(Table 3). We found weakly significant F statistics for
both sets of analyses with an FCT = 0.037 for the littoral
cell group and an FCT = 0.030 for the groups based on
geographical proximity. Results of the rpL16 SAMOVA
analysis are similar to those of the AMOVA analysis. We
found slightly significant FCT-values from a K = 4 to
K = 9 where the FCT-values ranged from an FCT = 0.019,
P = 0.041 to an FCT = 0.028, P = 0.022. However,
the implied groupings among the analyses are not
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4263
Fig. 2 Maximum-likelihood phylogram
of rpL16 cpDNA intergenic spacer from
King & Roalson (2009a). Numbers
above branches are Bayesian posterior
probabilities and numbers below
branches are bootstrap percentages.
Haplotypes were designated arbitrarily
based on preliminary assignment to a
neighbour-joining tree.
Fig. 3 Graph of the first three principal
coordinates of the PCoA based on interindividual genetic distance of the eleven
microsatellite loci. Principal axis (PA) 1
explains 25.52% of the variation, PA2
explains 17.34% of the variation, and
PA3 explains 16.35% of the variation.
congruent and are not geographically meaningful,
where most localities do not group with the nearest
population. It is important to note that although the
© 2009 Blackwell Publishing Ltd
algorithm is the same between SAMOVA and the AMOVA in
Arlequin, the K = 4 and K = 9 did not return a higher
FCT-value than the AMOVA analysis.
4264 M . G . K I N G , M . E . H O R N I N G and E . H . R O A L S O N
The SWEEP-BOTT analysis showed no evidence for a
genetic bottleneck, with the nonfounder event model
significantly (P < 10)8) more likely ()ln likelihood
)152.8412) than the bottleneck event model ()ln likeli­
hood )172.2840). The SWEEP-BOTT likelihood ratio test
strongly suggests that the most appropriate model is
the nonfounder event model. The nonfounder event
model assumes that there has been no measurable
decline in genetic diversity. Under the bottleneck
model, the maximum-likelihood estimates of the timing
of the bottleneck event, T = 0.63699 with the strength of
the bottleneck, S = 0.112. Similarly, the results of our
mismatch analysis suggest that C. macrocephala did not
experience a significant population expansion following
the LGM. We obtained an estimate of s = 3.25 with a
95% confidence interval ranging from a s = 2.00 to 4.58.
Our initial h estimate of, h0 = 0.04 was significantly
(P < 0.001) lower than our h1 = 11.84.
Discussion
Modern population genetic estimates
Carex macrocephala is a paradioecious, rhizomatously
spreading, clonal sedge; a single genet can send out
multiple male and female inflorescences across a rela­
tively large area. This can lead to a high rate of self-pol­
lination and as such a mechanism for inbreeding in a
sexually reproducing plant. If migration were not
homogenizing the populations along the northwest
coast, we would expect drift to have a large impact on
the allele frequencies (Abbott et al. 2000); this would
manifest in a high estimate of population differentia­
tion, unlike the observed patterns reported here. Seed
and ⁄ or rhizome dispersal was high enough that drift
does not appear to have a large impact on the subpopu­
lations at plastid or nuclear loci. If the levels of gene
flow were not able to overcome drift within each
subpopulation or if gene flow was limited to pollen
dispersal, we would expect to see population sub-struc­
turing and very low haplotype diversity in the rpL16
data set. As it is, the AMOVA, SAMOVA and FST-value for
the rpL16 marker suggest high migration across the
west coast of North America. Arafeh & Kadereit (2006)
found that the high seed dispersal rates and high clonal
growth rate within Calystegia soldanella were creating lit­
tle geographical structuring along the European coast.
The ability to disperse seeds or rhizomes via the sea
may be an important factor in limiting genetic structur­
ing. Similar results have been found in other coastal
species such as Elymus athericus where fragmentation
was caused by habitat selection rather than geography,
but seed dispersal was high (Bockelmann et al. 2003),
mangroves with high seed dispersal (Nettel & Dodd
2007) and the eelgrass Zostera marina (Olsen et al.
2004).
Our overall estimates of F = 0.979 and f = 0.956 indi­
cate very high levels of inbreeding; moreover, the stan­
dardized G¢ST of 0.032 indicates a low fixation level
across North America. These patterns suggest that lev­
els of inbreeding within each subpopulation are high
enough to cause the non-HWE and LD beyond what is
normally expected for subdivided populations (Pollak
1987; Nordborg 2000). Interestingly, the inbreeding lev­
els are higher than previously reported in clonal sedges
(Stenström et al. 2001; Tyler 2002). Given the breeding
biology of Carex macrocephala, we assume that the high
levels of homozygosity are a result of inbreeding occur­
ring within each subpopulation. Nonrandom mating
within each subpopulation leads to departures from
HWE, but deviations of this magnitude are typically the
result of high levels of selfing (Fenster et al. 2003; Bal­
loux et al. 2004). Furthermore, the levels of LD between
these loci are typical when selection is acting upon the
loci or when inbreeding levels are high (Charlesworth
2003; Glémin et al. 2006). These high estimates of
inbreeding may have led to a reduction in allelic diver­
sity within some subpopulations. Ingvarsson (2002) dis­
cussed a reduction in genetic variation through
recurrent extirpation and colonization of subpopula­
tions in a metapopulation. Given that C. macrocephala
populations often persist under metapopulation dynam­
ics, this may promote inbreeding.
These inbreeding levels also make it difficult to assess
clonal diversity in each subpopulation. Standard clonal
richness indices such as the Shannon–Wiener index, the
Simpson index and the associated evenness indices can
be informative when deviations from HWE are not
strong. Calculating these indices based on the number
of multilocus lineages can lead to an understanding of
the overall clonal diversity within the population. How­
ever, when inbreeding is occurring at a higher than
expected rate, such as when selfing or biparental polli­
nation of close relatives occurs, the clonal diversity can
be difficult (Arnaud-Haond et al. 2007). Estimates of D*
were all >0.970 and this could represent an overesti­
mate of clonal diversity (Arnaud-Haond et al. 2007).
The low pgenFIS-results show that it was problematic for
us to detect clonal diversity correctly. The probable lack
of genetic structuring along the coast as well as the high
levels of inbreeding may contribute to these issues.
The AMOVA analyses support our hypothesis that
inbreeding is the cause of reduced levels of heterozy­
gosity and that migration is distributing genetically dis­
tinct individuals across the species range in North
America. By partitioning the sampling localities into
groups according to littoral cells and currents off the
northwest coast, the per cent of variation associated
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4265
with the between group category was low: 5.5% for the
microsatellite data set and 7.2% for the rpL16 data set.
By further combining localities according to general
geographical regions, the between group variance was
reduced to 3.3% for the microsatellite data set and
4.8% for the rpL16 data set, which are both still signifi­
cant amounts of variation. Additionally, between group
FCT values ranged from 0.018 to 0.037 indicating low
fixation even across large geographic areas. The higher
levels of FST are indicative of the significant pairwise
FST-values (data not shown), which can result from high
levels of inbreeding. This suggests that migration is
high enough to homogenize the genetic variation
among these regions, but the selfing rate of C. macrocep­
hala is high enough to keep the heterozygosity low.
Although the SAMOVA results returned slightly signifi­
cant FCT-values for K’s between 4 and 10, these were
not plausible. For example, the K = 4 returned a signifi­
cant grouping of populations from Alaska and Oregon.
The loss of geographical structuring along the coast is
apparent in the PCoA as well. Given the characteristics
of the nuclear microsatellite loci, we performed the
PCoA to explore population differentiation among
C. macrocephala populations. We were precluded from
analysing our data with a genetic based clustering algo­
rithm such as Structure (Pritchard et al. 2000) or AWC­
LUST (Gao & Starmer 2007) because of the extreme levels
of inbreeding, which violates important assumptions of
the underlying models. However, our PCoA analysis
serves a similar function, with individuals cluster
according to their pairwise genetic distance. When
those genetic distances are homogenized among subpopulations and individuals are sampled from a meta­
population, there should be no subgrouping within the
PCoA. According to the first three principal axes, there
is a single large group (Fig. 3), with no evidence of
population substructuring. This strongly suggests that
migration is high enough to overcome drift in the North
American population of C. macrocephala. The significant
levels of clustering found in the AMOVA and SAMOVA
analyses may be the result of long distance dispersal of
rhizomes along the coast causing similar genotypes and
haplotypes to be found in distant populations. How­
ever, the clustering was only marginally significant and
there was no congruence between the clusters identified
using nuclear and plastid loci.
Historical population genetic estimates
According to results of the SWEEP-BOTT analysis of the
rpL16 marker, C. macrocephala did not experience a sig­
nificant bottleneck during the LGM. The likelihood ratio
test suggests that the most appropriate model for this
data set is the nonfounder event model, which assumes
© 2009 Blackwell Publishing Ltd
that there has been no significant reduction in genetic
diversity in the history of the population. If we assume
that a bottleneck event had occurred at some point in
the history of the North American lineage, we estimated
the time of the bottleneck event at c. 186 000 years ago
and the strength of the bottleneck was weak (Galtier
et al. 2000). Given the error associated with using a sin­
gle marker for estimating bottleneck events under a
coalescent model, this may coincide with the divergence
of C. macrocephala in Asia from C. macrocephala in North
America c. 125 000 years ago (King & Roalson 2009a).
Our initial expectation was that a bottleneck event
would have reduced the genetic diversity along the
North American coast during the LGM and that follow­
ing the retreat of the continental ice sheet and the rising
of sea level, populations of C. macrocephala would have
expanded from those refugia persisting along the North
American coast. Our mismatch analysis suggests that a
population expansion did occur, but long before the
LGM. A s = 3.25 and 95% confidence interval 2.00–4.58
indicate an expansion between 3 and 7.5 MA, assuming
a mutation rate of 1 · 10)7 ⁄ site ⁄ year (Wolfe et al. 1987;
Willyard et al. 2007). According to the multilocus
coalescent analysis by King & Roalson (2009a), the age
of the North American lineage of C. macrocephala is
c. 125 000 years and hence this proposed expansion
probably occurred well before the species arrived in
North America. We would need to change our mutation
rate estimate to 2 · 10)5 ⁄ site ⁄ year to reach a period of
time to coincide with the LGM, a rate well outside any
established mutation rate estimate for the rpL16 spacer.
The loss of geographical structuring could be
explained by two hypotheses. The first hypothesis is
that C. macrocephala was restricted to a single refugium
during the LGM. The restriction to a single refugium
would have reduced the genetic diversity, and although
present day inbreeding levels are high, it is expected
that the bottleneck signal would still be present in the
rpL16 marker. Following the LGM, those populations
would have gone through a significant expansion along
the coast following the retreat of the ice sheet. Even
though modern day gene flow is high and inbreeding
levels are masking geographical signature in the microsatellite loci, the results of our SWEEP-BOTT and mismatch
analyses using the rpL16 spacer are quite conclusive,
showing that a significant reduction in diversity is unli­
kely to have occurred in the past 125 000 years. The
supported alternative hypothesis is that C. macrocephala
was not restricted to refugia during the LGM and there­
fore did not experience significant recent bottlenecks or
population expansions. Other studies on Carex species
show that the plastid markers do typically have lowlevels of diversity. Schönswetter et al. (2006) conducted
a phylogeographical analysis on Carex atrofusca in the
4266 M . G . K I N G , M . E . H O R N I N G and E . H . R O A L S O N
Alps and Artic and found little intra-population level
diversity using amplified fragment length polymor­
phism’s, trnD-trnT and trnGUUC spacers and the rpS16
intron. This is in direct contrast to our results where a
majority of our variance is found within populations.
North American phylogeographical hypothesis
Identifying the locations of suitable habitat for C. macro­
cephala during the Pleistocene is difficult when historical
sea levels were lower than current levels. Offshore dril­
ling records indicate that ice-free areas may have
existed around the QCI, Vancouver Island and northern
Washington coast (Hetherington et al. 2003; Porter
2004). Other evidence suggests that ice-free areas were
common along the coast and that the ice sheet only
extended to the coast in the form of glaciers along the
valleys (Barrie & Conway 1999; Fulton et al. 2004; Kauf­
man & Manley 2004; Porter 2004). Sediment deposits
along the continental shelf suggest that ice-free areas
may have contained sandy beaches and it would not be
unreasonable to assume that these beaches formed
quickly. Following the retreat of the Cordilleran Ice
Sheet, there was a rapid rise in eustatic sea level of
more than 100 m occurring over 5000 years. It then took
additional 8000 years for sea level to rise another 20 m
to its current elevation (Twichell & Cross 2002). Twic­
hell & Cross (2002) used data from core drillings off the
coast of North America to imply that the sandy beaches
along the northwest coast of North America have only
been in their current position for <3000 years (Barrie &
Conway 1999; Twichell & Cross 2002). This further
strengthens our argument that C. macrocephala would
have populated any newly formed beach systems
quickly and that the ability to migrate quickly along the
coast was essential.
Other studies on species that usually inhabit areas
north of the continental ice sheet suggest reduction in
the populations into glacial refugia either north or south
of the ice sheet margin (Magri et al. 2003; Dorken &
Barrett 2004; Anderson et al. 2006; Soltis et al. 2006;
Edh et al. 2007; Hodgins & Barrett 2007; Michalski &
Durka 2007; Naciri & Gaudeul 2007). However, a recent
study by O’Connell et al. (2008) found genetic patterns
indicative of expansion from a single refugium south of
the continental ice sheet rather than from multiple refu­
gia both north and south of the ice sheet front. Most of
these studies base their conclusions on patterns that are
consistent with expansion from glacial refugia. Studies
on species inhabiting sky island populations in the
American southwest indicate that the LGM may have
enabled population migration throughout the mountain
ranges by providing suitable habitat between the ranges
and allowing these species to have a wider distribution
(Masta 2000; Knowles 2001; DeChaine & Martin 2004).
These studies suggest that climate change as a result of
Pleistocene glaciation was responsible for geographical
expansion, followed by vicariance during the subse­
quent warming periods. The post-Pleistocene isolation
of these species has left genetic patterns observable in
several species.
It would not be unreasonable to suspect expansion of
C. macrocephala during the LGM because of the lowering
of sea level, but post-Pleistocene isolation should not be
considered when G¢ST and genetic structuring is this
low. Carex macrocephala neither expanded during the
LGM followed by isolation into distinct populations as
occurred in sky islands, nor was it restricted to a few
glacial refugia where genetic bottlenecking would have
occurred.
Conclusions
The goals of our study were to determine the impact
the LGM had on the distribution and population
genetic architecture of Carex macrocephala along the
coast of North America. The results of the AMOVA and
PCoA suggest that gene flow is genetically homogeniz­
ing the subpopulations along the coast into a single
metapopulation. Furthermore, the genetic fixation level
across North America is low and inbreeding values are
high. Although we expected a bottleneck to have
occurred during the LGM, the SWEEP-BOTT analysis sug­
gests that the most appropriate model is a nonfounder
event model and that a bottleneck event could have
only occurred long before the LGM. Palaeogeological
evidence suggests that there could have been a large
number of ice-free areas along the coast that may have
had favourable C. macrocephala habitat. These pieces of
evidence and analytical results suggest that C. macrocep­
hala was widespread during the LGM, rather than being
isolated into only a few glacial refugia. Thus, the pic­
ture for this obligate coastal species differs substantially
from that of other terrestrial plants of the Pacific North­
west that underwent substantial bottlenecks in glacial
refugia during the Pleistocene (Brunsfeld et al. 2001;
Brunsfeld & Sullivan 2006; O’Connell et al. 2008).
Acknowledgements
We would like to thank Timothy Carey and Danielle Davis for
their laboratory assistance. We also wish to acknowledge the
help of the following individuals in securing collection permits
and the shipments of the collections: Andrew Friske, Erica
Wheeler, Chris Kissinger (Ministry of Environment, British
Columbia), John McIntosh (Pacific Rim National Park), Presi­
dent Guujaaw and the Haida Nation, Robert DeVelice (Chug­
ach National Forest), Mary Stensvold (Tongass National
Forest), Jay Schleier (Oregon Department of Natural Resources)
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4267
and Deaydra Wise (Washington State Parks). We would also
like to thank Richard Gomulkeiwicz, Michael Webster, Moira
Scascitelli, Nolan Kane and two anonymous reviewers for
reviewing this manuscript. This research was supported by the
Native Plant Society of Oregon, Betty Higinbotham Trust and
Washington State University.
References
Abbott RJ, Brochmann C (2003) History and evolution of the
arctic flora: in the footsteps of Eric Hultén. Molecular Ecology,
12, 299–313.
Abbott R, Smith LC, Milne RL, Crawford RM, Wolff K, Balfour
J (2000) Molecular analysis of plant migration and refugia in
the Arctic. Science, 289, 1343–1346.
Anderson LL, Hu FS, Nelson DM, Petit RJ, Paige KN (2006)
Ice-age endurance: DNA evidence of a white spruce
refugium in Alaska. Proceedings of the National Academy of
Sciences, USA, 103, 12447–12450.
Arafeh R, Kadereit JW (2006) Long-distance seed dispersal,
clone longevity and lack of phylogeographical structure in
the European distributional range of the coastal Calystegia
soldanella (L.) R.Br. (Convolvulaceae). Journal of Biogeography,
33, 1461–1469.
Arnaud-Haond S, Belkhir K (2007) Genclone: a computer
program to analyse genotypic data, test for clonality
and describe spatial clonal organization. Molecular Ecology, 7,
15–17.
Arnaud-Haond S, Duarte CM, Alberto F, Serräo EA (2007)
Standardizing methods to address clonality in population
studies. Molecular Ecology, 16, 5115–5139.
Balloux F, Amos W, Coulson T (2004) Does heterozygosity
estimate inbreeding in real populations? Molecular Ecology,
13, 3021–3031.
Barendregt RW, Duk-Rodkin A (2004) Chronology and extent
of Late Cenozoic ice sheets in North America: a
magnetostratigraphic assessment. In: Quaternary Glaciations
—Extent and Chronology; Part II: North America (eds Ehlers J,
Gibbard PL), pp. 1–8. Elsevier, Amsterdam.
Barrie JV, Conway KW (1999) Late Quaternary glaciation and
postglacial stratigraphy of the northern Pacific margin of
Canada. Quaternary Research, 51, 113–123.
Bockelmann AC, Reusch TBH, Bijlsma R, Bakker JP (2003)
Habitat differentiation vs. isolation-by-distance: the genetic
population structure of Elymus athericus in European salt
marshes. Molecular Ecology, 12, 505–515.
Bonin A, Bellemain E, Bronken Eidesen P et al. (2004) How to
track and assess genotyping errors in population genetics
studies. Molecular Ecology, 13, 3261–3273.
Brunsfeld SJ, Sullivan J (2006) A multi-compartmented glacial
refugium in the northern Rocky Mountains: evidence from
the phylogeography of Cardamine constancei (Brassicaceae).
Conservation Genetics, 6, 895–904.
Brunsfeld SJ, Sullivan J, Soltis DE, Soltis PS (2001) Comparative
phylogeography of northwestern North America: a
synthesis. In: Integrating Ecological and Evolutionary Processes
in a Spatial Context (eds Silvertown J, Antonovics J), pp. 319–
339. Blackwell Science, Oxford.
Carstens BC, Stevenson AL, Degenhardt JD, Sullivan J (2004)
Testing nested phylogenetic and phylogeographic hypotheses
© 2009 Blackwell Publishing Ltd
in the Plethodon vandykei species group. Systematic Biology, 53,
781–792.
Carstens BC, Degenhardt JD, Stevenson AL, Sullivan J (2005)
Accounting for coalescent stochasticity in testing
phylogeographical
hypotheses:
modeling
Pleistocene
population structure in the Idaho giant salamander
Dicamptodon aterrimus. Molecular Ecology, 14, 255–265.
Charlesworth D (2003) Effects of inbreeding on the genetic
diversity of populations. Philosophical Transactions of the Royal
Society of London, Series B: Biological Sciences, 358, 1051–1070.
DeChaine EG, Martin AP (2004) Historic cycles of
fragmentation and expansion in Parnassius smintheus
(Papilionidae) inferred using mitochondrial DNA. Evolution,
58, 113–127.
Demboski JR, Cook JA (2001) Phylogeography of the dusky
shrew, Sorex monticolus (Insectivora, Soricidae): insight into
deep and shallow history in northwestern North America.
Molecular Ecology, 10, 1227–1240.
Demboski JR, Sullivan J (2003) Extensive mtDNA variation
within the yellow-pine chipmunk, Tamias amoenus (Rodentia:
Sciuridae), and phylogeographic inferences for northwest
North America. Molecular Phylogenetics and Evolution, 26,
389–408.
Demboski JR, Stone KD, Cook JA (1999) Further perspectives on
the Haida Gwaii glacial refugium. Evolution, 53, 2008–2012.
Dorken ME, Barrett SCH (2004) Chloroplast haplotype
variation among monoecious and dioecious populations of
Sagittaria latifolia (Alismataceae) in eastern North America.
Molecular Ecology, 13, 2699–2707.
Dupanloup I, Schneider S, Excoffier L (2002) A simulated
annealing approach to define the genetic structure of
populations. Molecular Ecology, 11, 2571–2581.
Edh K, Widen B, Ceplitis ALF (2007) Nuclear and chloroplast
microsatellites reveal extreme population differentiation and
limited gene flow in the Aegean endemic Brassica cretica
(Brassicaceae). Molecular Ecology, 16, 4972–4983.
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.
Excoffier L, Laval G, Schneider S (2005) ARLEQUIN ver. 3.0: an
integrated software package for population genetics data
analysis. Evolutionary Bioinformatics Online, 1, 47–50.
Fenster CB, Vekemans X, Hardy OJ (2003) Quantifying gene flow
from spatial genetic structure data in a metapopulation of
Chamaecrista fasciculata (Leguminosae). Evolution, 57, 995–1007.
Fleming MA, Cook JA (2002) Phylogeography of endemic
ermine (Mustela erminea) in southeast Alaska. Molecular
Ecology, 11, 795–807.
Fulton RJ, Ryder JM, Tsang S (2004) The Quaternary glacial
record of British Columbia, Canada. In: Quaternary
Glaciations—Extent and Chronology; Part II: North America (eds
Ehlers J, Gibbard PL), pp. 39–50. Elsevier, Amsterdam.
Galtier N, Depaulis F, Barton NH (2000) Detecting bottlenecks
and selective sweeps from DNA sequence polymorphism.
Genetics, 155, 981–987.
Gao X, Starmer J (2007) Human population structure detection
via multilocus genotype clustering. BMC Genetics, 8, 34.
Glémin S, Bazin E, Charlesworth D (2006) Impact of mating
systems on patterns of sequence polymorphism in flowering
4268 M . G . K I N G , M . E . H O R N I N G and E . H . R O A L S O N
plants. Proceedings of the Royal Society of London, Series B:
Biological Sciences, 273, 3011–3019.
Good JM, Demboski JR, Nagorsen DW, Sullivan J (2003)
Phylogeography and introgressive hybridization: Chipmunks
(genus Tamias) in the northern Rocky Mountains. Evolution,
57, 1900–1916.
Goudet J (1995) FSTAT (version 1.2): a computer program to
calculate F-statistics. Journal of Heredity, 86, 485–486.
Hedrick PW (2005) A standardized genetic differentiation
measure. Evolution, 59, 1633–1638.
Hetherington R, Barrie JV, Reid RGB et al. (2003) Late
Pleistocene coastal paleogeography of the Queen Charlotte
Islands, British Columbia, Canada, and its implications for
terrestrial biogeography and early postglacial human
occupation. Canadian Journal of Earth Sciences, 40, 1755–1766.
Hewitt GM (1996) Some genetic consequences of ice ages, and
their role in divergence and speciation. Biological Journal of
the Linnean Society, 58, 247–276.
Hewitt GM (2004) Genetic consequences of climatic oscillations
in the Quaternary. Philosophical Transactions: Biological
Sciences, 359, 183–195.
Hodgins KA, Barrett SCH (2007) Population structure and
genetic diversity in tristylous Narcissus triandrus: insights
from microsatellite and chloroplast DNA variation. Molecular
Ecology, 16, 2317–2332.
Ingvarsson PK (2002) A metapopulation perspective on genetic
diversity and differentiation in partially self-fertilizing
plants. Evolution, 56, 2368–2373.
Ishikawa S-I, Furukawa A, Okuda T, Oikawa T (1993)
Germination requirements in Carex kobomugi (Sea Isle).
Journal of Plant Research, 106, 245–248.
Kaufman DS, Manley WF (2004) Pleistocene maximum and
Late Wisconsinan glacier extents across Alaska, U.S.A. In:
Quaternary Glaciations—Extent and Chronology; Part II: North
America (eds Ehlers J, Gibbard PL), pp. 9–28. Elsevier,
Amsterdam.
King MG, Roalson EH (2009a) Discordance between
phylogenetics and coalescent-based divergence modelling:
exploring phylogeographic patterns of speciation in the Carex
macrocephala species complex. Molecular Ecology, 18, 468–482.
King MG, Roalson EH (2009b) Isolation and characterization of
11 microsatellite loci from Carex macrocephala (Cyperaceae).
Conservation Genetics, 10, 531–533. DOI: 10.1007/s10592-008­
9558-5.
Knowles LL (2001) Did the Pleistocene glaciations promote
divergence? Tests of explicit refugial models in montane
grasshopprers. Molecular Ecology, 10, 691–701.
Magri D, Fineschi S, Bellarosa R et al. (2003) The distribution
of Quercus suber chloroplast haplotypes matches the
palaeogeographical history of the western Mediterranean.
Molecular Ecology, 12, 1–13.
Masta SE (2000) Phylogeography of the jumping spider
Habronattus pugillis (Araneae: Salticidae): recent vicariance of
sky island populations? Evolution, 54, 1699–1711.
Mastrogiuseppe J (2002) Flora of North America: north of
Mexico. In: Volume 23 Magnoliophyta: Commelinidae (in part):
Cyperaceae (ed. Flora of North America Editorial Committee).
pp. 307–309. Oxford University Press, New York.
Meirmans PG (2006) Using the AMOVA framework to
estimate a standardized genetic differentiation measure.
Evolution, 60, 2399–2402.
Michalski SG, Durka W (2007) High selfing and high
inbreeding depression in peripheral populations of Juncus
atratus. Molecular Ecology, 16, 4715–4727.
Naciri Y, Gaudeul M (2007) Phylogeography of the endangered
Eryngium alpinum L. (Apiaceae) in the European Alps.
Molecular Ecology, 16, 2721–2733.
Nei M (1973) Analysis of gene diversity in subdivided
populations. Proceedings of the National Academy of Sciences,
USA, 70, 3321–3323.
Nettel A, Dodd RS (2007) Drifting propagules and receding
swamps: genetic footprints of mangrove recolonization and
dispersal along tropical coasts. Evolution, 61, 958–971.
Nielson M, Lohman K, Sullivan J (2001) Phylogeography of the
tailed frog (Ascaphus truei): implications for the biogeography
of the Pacific Northwest. Evolution, 55, 147–160.
Nordborg M (2000) Linkage disequilibrium, genetrees and
selfing: an ancestral recombination graph with partial selffertilization. Genetics, 154, 923–929.
O’Connell LM, Ritland K, Thompson SL (2008) Patterns of
post-glacial colonization by western redcedar (Thuja plicata,
Cupressaceae) as revealed by microsatellite markers. Botany,
86, 194–203.
Olsen JL, Stam WT, Coyer JA et al. (2004) North Atlantic
phylogeography and large-scale population differentiation of
the seagrass Zostera marina L. Molecular Ecology, 13, 1923–
1941.
van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P
(2004) MICRO-CHECKER: software for identifying and correcting
genotyping errors in microsatellite data. Molecular Ecology
Notes, 4, 535–538.
Pannell JR, Verdú M (2006) The evolution of gender
specialization from dimorphic hermaphroditism: paths from
heterodichogamy to gynodioecy and androdioecy. Evolution,
60, 660–673.
Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in
Excel. Population genetic software for teaching and research.
Molecular Ecology Notes, 6, 288–295.
Petit RJ, Aguinagalde I, de Beaulieu J-L et al. (2003) Glacial
refugia: hotspots but not melting pots of genetic diversity.
Science, 300, 1563–1565.
Pollak E (1987) On the theory of partially inbreeding finite
populations. I. Partial selfing. Genetics, 117, 353–360.
Porter SC (2004) Glaciation of western Washington, U.S.A. In:
Quaternary Glaciations—Extent and Chronology; Part II: North
America (eds Ehlers J, Gibbard PL), pp. 289–294. Elsevier,
Amsterdam.
Pritchard JK, Stephens M, Donnelly P (2000) Inference of
population structure using multilocus genotype data.
Genetics, 155, 945–959.
Rambaut A (1996) SE-AL: Sequence alignment editor. Available
from http://tree.bio.ed.ac.uk/software/seal/.
Rice WR (1989) Analyzing tables of statistical tests. Evolution,
43, 223–225.
Riddle BR (1996) The molecular phylogeographic bridge
between deep and shallow history in continental biotas.
Trends in Ecology and Evolution, 11, 207–211.
Roalson EH, Columbus JT, Friar EA (2001) Phylogenetic
relationships in Cariceae (Cyperaceae) based on ITS (nrDNA)
and trnT-L-F (cpDNA) region sequences: assessment of
subgeneric and sectional relationships in Carex with emphasis
on section Acrocystis. Systematic Botany, 26, 318–341.
© 2009 Blackwell Publishing Ltd
PERSISTENCE OF C. MACROCEPHALA DURING THE LGM 4269
Schönswetter P, Popp M, Brochmann C (2006) Central Asian
origin of and strong genetic differentiation among
populations of the rare and disjunct Carex atrofusca
(Cyperaceae) in the Alps. Journal of Biogeography, 33, 948–956.
Shaw J, Lickey EB, Beck JT et al. (2005) The tortoise and the
hare II: relative utility of 21 noncoding chloroplast DNA
sequences for phylogenetic analysis. American Journal of
Botany, 92, 142–166.
Soltis DE, Morris AB, McLachlan JS, Manos PS, Soltis PS (2006)
Comparative phylogeography of unglaciated eastern North
America. Molecular Ecology, 15, 4261–4293.
Standley LA (1985) Paradioecy and gender ratios in Carex
macrocephala Cyperaceae. American Midland Naturalist, 113,
283–286.
Steele CA, Storfer A (2006) Coalescent-based hypothesis testing
supports multiple Pleistocene refugia in the Pacific
Northwest for the Pacific giant salamander (Dicamptodon
tenebrosus). Molecular Ecology, 15, 2477–2487.
Stenström A, Jonsson BO, Jónsdóttir IS, Fagerström T, Augner
M (2001) Genetic variation and clonal diversity in four clonal
sedges (Carex) along the Arctic coast of Eurasia. Molecular
Ecology, 10, 497–513.
Twichell DC, Cross VA (2002) Holocene evolution of the
southern Washington and northern Oregon shelf and coast:
Geologic discussion and GIS data release. USGS Open-File
Report 01-076. Available from http://pubs.usgs.gov/of/
of01-076/index.htm.
Tyler T (2002) Geographical distribution of allozyme variation
in relation to post-glacial history in Carex digitata, a
widespread European woodland sedge. Journal of
Biogeography, 29, 919–930.
Weckworth BV, Talbot S, Sage GK, Person DK, Cook JA (2005)
A signal for independent coastal and continental histories
© 2009 Blackwell Publishing Ltd
among North American wolves. Molecular Ecology, 14, 917–
931.
Weir BS, Cockerham CC (1984) Estimating F-statistics for the
analysis of population structure. Evolution, 38, 1358–1370.
Widmer A, Lexer C (2001) Glacial refugia: sanctuaries for
allelic richness, but not for gene diversity. Trends in Ecology
and Evolution, 16, 267–268.
Williams T, Kelley C (2004) GNUPLOT 4.2.2. Available from
http://www.gnuplot.info/.
Willyard A, Syring J, Gernandt DS, Liston A, Cronn R (2007)
Fossil calibration of molecular divergence infers a moderate
mutation rate and recent radiations for Pinus. Molecular
Biology and Evolution, 24, 90–101.
Wolfe KH, Li W-H, Sharp PM (1987) Rates of nucleotide
substitution vary greatly among plant mitochondrial,
chloroplast, and nuclear DNAs. Proceedings of the National
Academy of Sciences, USA, 84, 9054–9058.
Wright S (1951) The genetical structure of populations. Annals
of Eugenics, 15, 323–354.
MGK is currently studying diversification of the annual sun­
flowers using genomic tools. He is involved in projects investi­
gating ecological speciation, intragenic recombination, and the
sunflower genome project. EHR research centers on several
families of flowering plants including the Cyperaceae and Ges­
neriaceae. MEH is a geneticist with the USDA Forest Service
who studies plant adaptation and develops seed transfer
guidelines for native plant species that are important for land­
scape restoration activities (i.e., post-fire and roadside revege­
tation).