Content uploaded by Cornelya FC Klütsch
Author content
All content in this area was uploaded by Cornelya FC Klütsch on Dec 19, 2016
Content may be subject to copyright.
ORIGINAL
ARTICLE
Ancient diversification in glacial refugia
leads to intraspecific diversity in a
Holarctic mammal
Jean L. Polfus
1
*, Micheline Manseau
1,2
, Cornelya F.C. Kl€
utsch
3
,
Deborah Simmons
4,5
and Paul J. Wilson
3
1
Natural Resources Institute, University of
Manitoba, Winnipeg, Manitoba R3T 2M6,
Canada,
2
Protected Areas Establishment and
Conservation Directorate, Parks Canada,
Gatineau, Quebec J8X 0B3, Canada,
3
Biology
Department, Trent University, Peterborough,
Ontario K9J 7B8, Canada,
4
Saht
u Renewable
Resources Board, Tulita NT X0E 0K0,
Canada,
5
Aboriginal Studies, University of
Toronto, Toronto Ontario M5S 2J7, Canada
*Correspondence: Jean L. Polfus, Natural
Resources Institute, University of Manitoba,
303 Sinnott Building, 70 Dysart Road,
Winnipeg, Manitoba R3T 2M6, Canada.
E-mail: jeanpolfus@gmail.com
This is an open access article under the terms
of the Creative Commons Attribution-
NonCommercial-NoDerivs License, which
permits use and distribution in any medium,
provided the original work is properly cited,
the use is non-commercial and no
modifications or adaptations are made.
ABSTRACT
Aim Glacial-interglacial cycles influenced the contemporary genetic structure of
many North American species. While phylogeographical lineage divergence among
Pleistocene refugia has been proposed as a significant driver of subspecific and
ecotypic differentiation, emerging evidence highlights the role of diversification
within refugia in producing post-glacial variation. Caribou (Rangifer tarandus)
exhibit significant morphological, ecological and behavioural phenotypic variation
and occurred within Beringian and sub-Laurentide refugia. More specifically, the
boreal ecotype of woodland caribou ranges from the southern regions of Canada
to the Northwest Territories (NWT). Woodland caribou are generally accepted to
have evolved south of the glacial extent, but the boreal ecotype in the northern
part of their range co-occurs with caribou that have a Beringian origin. This prox-
imity provides an opportunity to test whether woodland caribou colonized boreal
habitats from a single southern refugial source or if independent evolution to a
common ecotype resulted from diversification within refugia.
Location Northwestern Canada.
Methods We used approximate Bayesian computation to discriminate between
alternate evolutionary histories of caribou belonging to boreal, northern mountain
and barren-ground ecotypes using microsatellite and mtDNA markers.
Results Our analysis indicates that unlike the southern-evolved boreal ecotype,
the boreal ecotype of central NWT has Beringian origins and arose from a
common lineage with barren-ground and mountain caribou. Importantly, the
divergence of the lineage resulting in the boreal ecotype of central NWT signif-
icantly pre-dates the Last Glacial Maximum.
Main conclusions We demonstrate that independent evolutionary trajectories
can converge on a similar phenotype and for the first time show that the bor-
eal ecotype of caribou in North America contains two phylogeographical
assemblages. The ancient divergence suggests that diversification within Berin-
gia could have resulted in ecological specialization. An eco-evolutionary focus
will be essential to designing biodiversity conservation strategies for caribou
that maximize genetic diversity and preserve adaptive potential in this
intraspecifically diverse species.
Keywords
approximate Bayesian computation, Beringia, caribou, convergent evolution,
ecotype, glacial refugia, ice age, parallel evolution, Rangifer
INTRODUCTION
Intraspecific variation is recognized as a significant driver in
the establishment and function of ecological dynamics
including population persistence, competition and responses
to environmental change (Bolnick et al., 2011). However, the
evolutionary processes that lead to the development and per-
sistence of intraspecific variation, especially for vagile species
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
http://wileyonlinelibrary.com/journal/jbi 1
doi:10.1111/jbi.12918
Journal of Biogeography (J. Biogeogr.) (2016)
in continuous habitats, can be difficult to identify
(Fitzpatrick et al., 2015; Puckett et al., 2015). Glacial cycles
during the Pleistocene have had a significant impact on spe-
cies distributions and genetic diversity (Hewitt, 2000). In
North America, vicariant divergence associated with the
North American Laurentide and Cordilleran ice sheets facili-
tated phylogeographical lineage diversification in several spe-
cies (Dyke, 2004; Weksler et al., 2010 and references
therein). Subsequent isolation and divergent selection pres-
sures in the physiographic conditions of refugia are com-
monly considered to influence intraspecific diversification
through genetic drift and adaptive evolution (Richardson
et al., 2014). However, recent research also points to the
importance of divergence within single large refugia as a
source of contemporary genetic variation and structure
(Galbreath et al., 2011; Lanier et al., 2015).
Northern cold-adapted species experienced extensive range
expansions, and in some cases increased population sizes
during glacial periods (Flagstad & Røed, 2003; Lorenzen
et al., 2011). The extensive Beringian refugium, that stretched
from eastern Siberia across the land bridge to Alaska and
into the Yukon, fostered considerable genetic diversity and
endemism (Weksler et al., 2010; Galbreath et al., 2011). Fol-
lowing glacial retreats, the reunification of divergent popula-
tions may have increased adaptive evolution through
introgression, or alternatively, disrupted local adaptation and
caused replacement or extinction of genealogical lineages
(Lanier et al., 2015; Kl€
utsch et al., 2016). During warm inter-
glacial periods, the ranges of cold-adapted species contracted
as viable tundra and boreal habitat were redistributed, which
contrasts with the pattern of expansion out of refugia dis-
played by many temperate species (Stewart et al., 2010).
Molecular techniques provide an opportunity to reconstruct
the population dynamics of cold-adapted species and predict
how phylogeographical patterns influence the contemporary
population structure (Stewart et al., 2010; Galbreath et al.,
2011; Esp
ındola et al., 2012).
In North America, caribou (Rangifer tarandus) persisted in
both high- and low-latitude habitats over the course of the
Pleistocene glaciations. The series of range oscillations and
repeated demographic fluctuations associated with the
expansion and retraction of continental glaciers produced
conspecific populations with distinct morphological, ecologi-
cal and behavioural traits (Flagstad & Røed, 2003). The
diverse spatial-temporal evolutionary histories that character-
ize caribou have made taxonomic clarity within the species
challenging and are evident in extensive intraspecific genetic
structure (Serrouya et al., 2012; Weckworth et al., 2012;
Kl€
utsch et al., 2016; Polfus et al., 2016). Genetic evaluations
have attributed the most pronounced intraspecific split (first
formally described as subspecies by Banfield in 1961) to two
distinct mitochondrial DNA (mtDNA) phylogeographical
lineages that originated south of the ice sheets (North Ameri-
can lineage; NAL) and north of the ice sheets (Beringian–
Eurasian lineage; BEL, Flagstad & Røed, 2003; Cronin et al.,
2005; McDevitt et al., 2009; Kl€
utsch et al., 2012; Weckworth
et al., 2012; Yannic et al., 2014). Finer-scale subdivisions fur-
ther classify North American caribou into ecotypes based on
geography and natural history traits (regardless of genealogi-
cal relationships), however, naming conventions do not
always correspond between jurisdictions and ecotype identifi-
cation can be ambiguous (COSEWIC, 2011; Pond et al.,
2016).
Woodland caribou (R. t. caribou) belong predominately to
the NAL and were isolated in habitats south of the Lauren-
tide ice sheet during the Last Glacial Maximum (LGM: 26.5–
19 thousand calendar years before present; kyr bp (Dyke,
2004)). Specifically, the boreal ecotype of woodland caribou
are forest-dwelling animals known for their sedentary beha-
viour, dark pelage, large body and long legs, small group-size
and low-population densities across their current range
within the Canadian boreal zone (Fig. 1). The boreal ecotype
is considered a Designatable Unit (DU; COSEWIC, 2011)
and is listed as threatened by the Canadian Species at Risk
Act as a result of population declines that are generally
attributed to extensive habitat loss and fragmentation (Envi-
ronment Canada, 2012).
In north-western Canada, at their northern range margin,
the boreal ecotype co-occurs with barren-ground caribou (R.
t. groenlandicus) that aggregate in large numbers to calve on
the tundra and migrate to the boreal forest during the winter
(Nagy et al., 2011) and the northern mountain ecotype (R. t.
caribou) that occur throughout the mountains of the North-
west Territories (NWT), northern British Columbia and
Yukon Territory (COSEWIC, 2011). However, even in the
face of range overlaps and known mixing between the types,
recent genetic analysis has shown that in central NWT, the
boreal ecotype can be differentiated (Polfus et al., 2016).
Likewise, indigenous Dene First Nation and M
etis people of
central NWT classify to
̨dzı ‘boreal woodland caribou’ based
on identifiable physical features and behavioural traits, fur-
ther supporting the boreal ecotype as a distinctive group in
the northern extent of their range (Polfus et al., 2016).
Interestingly, the boreal ecotype in central NWT (hereafter
NWT boreal) assigns predominately to the BEL based on
mitochondrial patterns (Polfus et al., 2016), similar to sym-
patric barren-ground and northern mountain animals
(Weckworth et al., 2012), and unlike the boreal ecotype from
southern provinces that assign to the NAL (Kl€
utsch et al.,
2012). This proximity provides an ideal opportunity to test
competing refugial hypotheses. If the boreal phenotype arose
independently from distinct evolutionary lineages as a result
of parallel phenotypic evolution, it would suggest that natu-
ral selection plays an important role in caribou intraspecific
variation (Schluter et al., 2004; Elmer & Meyer, 2011). In
particular, ecological traits may be expected to converge in
closely related genomes when certain environmental condi-
tions strongly favour particular evolutionary outcomes
(Rosenblum et al., 2014). Alternatively, the boreal phenotype
in central NWT may be a result of shared ancestry or his-
toric introgression with NAL animals. Genetic drift may also
be an important mechanism causing intraspecific
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
2
J. L. Polfus et al.
differentiation in caribou, especially in small isolated popula-
tions (Serrouya et al., 2012; Mager et al., 2014).
Given signatures of significant BEL ancestry in central
NWT caribou (Polfus et al., 2016), our goal was to test alter-
native evolutionary models to assess the origin of the boreal
ecotype at the northern range margin. Specifically, we tested
the following two alternative hypotheses: (1) the NWT boreal
ecotype diverged from the BEL and converged to a boreal
phenotype within Beringia; (2) the NWT boreal ecotype rep-
resents NAL woodland caribou that subsequently colonized
the northern boreal zone following retraction of the ice
sheets and experienced some level of introgression from BEL
caribou at the northern range margin. To discriminate
between these alternate evolutionary histories, we applied
approximate Bayesian computation (ABC) of nuclear and
mitochondrial genetic markers in contemporary caribou
populations representing the boreal ecotype in central NWT,
the barren-ground subspecies, the northern mountain
ecotype, and the nearest population of boreal ecotype with
NAL origins and little evidence of introgression from BEL.
We also evaluated whether estimated divergence times coin-
cided with significant glacial events. Ultimately, we tested
whether the boreal ecotype of woodland caribou evolved
from a single refugial lineage or independently from two
refugial lineages.
MATERIALS AND METHODS
Study area and sample collection
A description of the central NWT study area and sample col-
lection can be found in Polfus et al. (2016) and a description
of central Saskatchewan (SK) can be found in Galpern et al.
(2012a; Fig. 1). We assembled a dataset of caribou faecal and
Figure 1 The range of the boreal ecotype of woodland caribou occurs within the boreal zone in Canada from the Northwest Territories
to eastern Labrador (Brandt, 2009; COSEWIC, 2011; Environment Canada, 2012). Small black dots represent locations of caribou faecal,
tissue and blood strip samples collected in the Mackenzie Mountains (within the range of the northern mountain ecotype), the boreal
forest of the Saht
u region, central Northwest Territories (within the overlapping ranges of the boreal ecotype and barren-ground
caribou) and the boreal forest of central Saskatchewan (Smoothstone–Wapeweka population of boreal ecotype).
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
3
Diversification in glacial refugia in a Holarctic mammal
tissue samples from animals belonging to four major groups:
(1) barren-ground caribou from the Bluenose East and Blue-
nose West herds of central NWT; (2) northern mountain
ecotype from the Mackenzie Mountains, NWT; and two
populations of boreal ecotype from (3) central NWT and (4)
the Smoothstone–Wapeweka population, SK.
Microsatellite DNA genotyping
We followed protocols for microsatellite DNA extraction,
amplification and genotyping that were developed as part of
a long-term caribou genetics database (Galpern et al., 2012a;
Kl€
utsch et al., 2012, 2016). We genotyped a panel of nine
microsatellite loci (BM848, BM888, MAP2C, RT5, RT6, RT7,
RT9, RT24 and RT30; Bishop et al., 1994; Wilson et al.,
1997). We used genemarker 1.9.1 (SoftGenetics, LLC) to
determine allele size. Two people evaluated all electrophero-
grams and scores were compared on an online server. We
used allelematch 2.5 (Galpern et al., 2012b) to check for
genotyping errors, remove duplicate profiles and identify
individuals. Samples included in the final dataset had a mini-
mum of eight successfully amplified loci.
Mitochondrial DNA sequencing
We amplified and sequenced 429 bp of the mtDNA control
region using the primers L15394 and H15947 (Flagstad &
Røed, 2003) following Kl€
utsch et al. (2012, 2016). We used
bioedit 7.2.5 (Hall, 1999) to check and align sequences and
dnasp 5 (Librado & Rozas, 2009) to distinguish haplotypes.
Statistical data analysis
We tested each locus and population for significant devia-
tions from Hardy–Weinberg equilibrium (HWE) and linkage
disequilibrium (LD) using genepop 4.2 (Rousset, 2008). We
used structure 2.3.4 (Pritchard et al., 2000) to identify
population clusters (K) for K=1 through K=15 under the
admixture model with correlated allele frequencies. We con-
ducted five iterations for each K with 1,000,000 burn-ins and
10,000,000 Markov chain Monte Carlo repetitions on a high-
performance computing cluster (www.sharcnet.ca). We sum-
marized run statistics using structure harvester 0.6.94
(Earl & vonHoldt, 2012). We used spagedi 1.5 (Hardy &
Vekemans, 2002) to test microsatellite pairwise differentia-
tion.
Approximate Bayesian computation
We used ABC simulations to test competing evolutionary
models. ABC analysis allows rapid tests of different scenarios
by calculating summary statistics rather than exact likeli-
hoods (Csill
ery et al., 2010). Deviations between the simu-
lated and observed summary statistics are evaluated to
measure fit for each model investigated (Lopes & Boessen-
kool, 2010). We used the software diyabc 2.0.4 (Cornuet
et al., 2014) to explore whether the NWT boreal ecotype
diverged from the BEL or the NAL. Alternative scenarios
tested also included admixture between populations at vari-
ous time-scales. We divided the evolutionary scenarios into
two major groups: (1) admixture models with divergence
and admixture events (Fig. 2; scenarios 1–3) and, (2) split
models with no admixture events (scenarios 4–5). First, we
tested a set of split scenarios to identify the most likely can-
didates. The top three split models were added to a series of
preliminary runs that included admixture models. We nar-
rowed down the supported models to the top five (Fig. 2)
and included them in a final run to test support with three
datasets: microsatellites, mtDNA and a combined dataset.
We initially set the mutation model parameters in diyabc
to a stepwise mutation model as identified by Kl€
utsch et al.
(2016) and then fine-tuned the parameters to the dataset.
We set the prior range for the split between the two phylo-
geographical lineages to t
4
=10–25,000 generations. The
prior range of the divergence events were set to t
3
=100–
17,000 and t
2
=10–10,000 generations. The youngest event
had a prior range of t
1
=10–3000 generations for the com-
bined dataset. To convert time estimates to years we assumed
a generation time of 7 years for female caribou. We chose
summary statistics (i.e. mean number of alleles, mean size
variance of alleles, mean number of haplotypes, etc.) based
on their success in previous analyses on caribou (see Kl€
utsch
et al., 2016). Approximately 3 million simulations were used
to test scenarios on a high-performance computing cluster.
We compared simulations through logistic regression and
linear discriminant analysis in diyabc. We used the model-
checking option to assess the goodness-of-fit of model
parameter posterior combination (Appendix S1 of the Sup-
porting Information, Figs S1.1 & S1.2).
RESULTS
Population structure and diversity
We amplified 655 samples (Appendix S1, Table S1) from
individual caribou. There was no evidence that certain loci
deviated from HWE (6/36 cases significant after Bonferroni
correction) or expressed LD (1/144 cases significant after
Bonferroni correction). structure analysis revealed K=2
(ΔKcriterion) that corresponded to a NWT group and SK
group. The mean likelihood also supported additional sub-
structure at K=4 (all four groups showing differentiation;
Appendix S1, Figs S1.3–S1.5). Pairwise comparisons (F
ST
and
R
ST
) supported divergence among groups with the strongest
differentiation found between the NWT boreal and the SK
boreal (Appendix S1, Table S1.2).
We sequenced 370 caribou at the mtDNA control region
and found 79 haplotypes that fit into the well-resolved phy-
logeny of NAL and BEL (Kl€
utsch et al., 2012). Most haplo-
types were non-overlapping between groups (Appendix S1,
Fig. S1.6). We found only three NAL haplotypes (in 11 cari-
bou; 3.9%) in the NWT dataset, and only one BEL haplotype
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
4
J. L. Polfus et al.
(in three caribou; 3.4%) in the SK dataset (Appendix S1,
Table S1.3).
ABC analysis
All top models identified through ABC analysis suggested
that the NWT boreal ecotype has a BEL origin (Fig. 2). Sce-
nario 1 was identified as the most likely evolutionary model
for the microsatellite and combined dataset based on the
posterior probability values, credible intervals and logistic
regression (Appendix S1, Table S1.4, Fig. S1.7). Scenario 1
suggests that the NWT boreal ecotype diverged from the
BEL c. 60.5 kyr bp (CI: 19.5–109.2 kyr bp; combined dataset;
Table 1). This model also estimates that the northern moun-
tain ecotype arose relatively recently at c. 4.2 kyr bp (CI:
0.2–16.4 kyr bp; combined dataset) through admixture
between two divergent populations that had initially split
from the barren-ground and NWT boreal lineages of the
BEL c. 45.9 kyr bp (CI: 20.8–66.7 kyr bp; combined
dataset). Models that included divergence of the SK boreal
ecotype from the NWT boreal ecotype (or vice versa) were
not supported.
The most likely evolutionary model for the separate
mtDNA dataset was scenario 4. This model suggests that
both the northern mountain and the NWT boreal ecotypes
diverged from the barren-ground lineage of the BEL at dif-
ferent time points. This result can be explained in part by
the fact that NWT boreal caribou include primarily BEL
haplotypes. In contrast, caribou mtDNA data from central
Canada include more phylogenetically differentiated
Table 1 Time estimates in calendar years before present for
scenario 1 (found in Fig. 2) produced with approximate
Bayesian computation for the combined dataset that includes
caribou (Rangifer tarandus) microsatellite and mtDNA data
from central Northwest Territories and central Saskatchewan,
Canada.
Time point Mean Median
95% confidence
interval
t
1
4193 2688 206 16450
t
2
45920 46900 20790 66710
t
3
60550 59360 19460 109200
t
4
135800 141400 68600 173600
Figure 2 Top five approximate Bayesian computation scenarios tested with diyabc that model the evolutionary history of four
contemporary caribou (Rangifer tarandus) groups: barren-ground caribou, boreal ecotype of central Northwest Territories (NWT),
northern mountain ecotype and boreal ecotype of central Saskatchewan (SK), Canada.
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
5
Diversification in glacial refugia in a Holarctic mammal
haplogroups and therefore, more haplotypic diversity
(Kl€
utsch et al., 2016). Since the majority of haplotypes in
this analysis came from the BEL, the average number of
mtDNA substitutions in this dataset was also lower than
Kl€
utsch et al. (2016), which could influence time estimates
and model choice to a certain degree. Furthermore, replace-
ment events may have resulted in the loss of ancient haplo-
types.
DISCUSSION
The role of parallel evolution in intraspecific
diversity
We show for the first time that multiple evolutionary routes
can converge on a similar phenotype in an intraspecifically
diverse Holarctic species. Our analysis points to the role of
the Beringia refugium on genetic variation and structure in
contemporary caribou populations. Boreal caribou of central
NWT are specialized for survival in the boreal forest and are
phenotypically and behaviourally similar to southern boreal
ecotype animals (i.e. display sedentary behaviour, dark pelage
and large body size; COSEWIC, 2011). However, ABC simu-
lations and mtDNA lineage assignment support a BEL origin
for the NWT boreal ecotype, distinct from the NAL of the
boreal ecotype that diversified south of the ice sheets during
the LGM. Because divergence between the Beringian-derived
barren-ground and boreal lineages extends to c. 60.5 kyr bp,
an alternative model is possible where the northern boreal
lineage colonized southern habitats when the ice-free corri-
dor between the Laurentide and Cordilleran first opened –
c. 14–15 kyr bp (Dyke, 2004; Dixon, 2015) –or perhaps even
predating the LGM. However, our results show that the rep-
resentative southern boreal ecotype from SK diverged before
the Beringian-derived barren-ground and boreal lineages.
The order of divergence does not support a Pleistocene colo-
nization model, but rather implies independent convergence
to a similar boreal ecotype in separate refugia north and
south of the ice sheets. Thus, caribou from distinct poly-
phyletic groups converged on a shared phenotype.
Our study suggests that natural selection has influenced
the evolution of the boreal ecotype because a similar suite
of traits evolved independently in association with the envi-
ronmental pressures of the boreal forest. While we could
not test the timing of ecological diversification compared to
lineage divergence, we can infer that adaptation to Berin-
gian microhabitat was likely an adaptive driver of this lin-
eage. Furthermore, although genetic drift is suspected to
play a role in genetic diversification in caribou (Serrouya
et al., 2012; Mager et al., 2014), genetic drift would not be
expected to produce parallel phenotypic traits in multiple
lineages in correlation with specific environments (Schluter
et al., 2004). Thus, ecological variation and adaptive evolu-
tion may be significant drivers in caribou ecotype evolution
to the extent that independent lineages converged to similar
phenotypic outcomes.
Our results contrast with Banfield’s classic Rangifer taxo-
nomic interpretation, based largely on craniometrical mea-
surements, that included western mountain and boreal
ecotypes in the woodland subspecies (R. t. caribou) that orig-
inated in sub-Laurentide refugia. Rather, we show that the
mountain and boreal ecotype of central NWT are distinct
groups with BEL origins. Our results support the intuition of
Geist (2007), who, using pelage characteristics and taxo-
nomic inferences, suggested that the mountain and boreal
woodland caribou north of 60˚latitude were more likely
“splinter populations of barren ground caribou, which have
adapted to a more sessile life style, increased in body size
and assumed some ‘woodland mannerisms’”.
In fact, the NWT boreal ecotype may be similar to seden-
tary caribou that occur in the boreal zone of Alaska. In gen-
eral, Alaskan caribou belong to the BEL, but have
behavioural strategies that have been classified into migratory
and sedentary ecotypes (Hinkes et al., 2005; Mager et al.,
2014). However, the sedentary Alaskan caribou display sig-
nificantly less genetic structure than we found in the NWT
boreal ecotype. Using 19 microsatellites from caribou across
the Alaskan mainland, Mager et al. (2014) found little
genetic differentiation between migratory or sedentary herds
that also ranged greatly in population size and used both for-
est and tundra habitats. Thus, local behavioural strategies
may be relatively plastic within Alaska (Hinkes et al., 2005).
Similarly, genetic evidence suggests that Eurasian forest rein-
deer (R. t. fennicus) arose from the large continuous popula-
tion of BEL reindeer in the vast palaeo-tundra of Siberia and
central Eurasia during the Pleistocene (Flagstad & Røed,
2003). Thus, it is possible that the forest reindeer, the NWT
boreal ecotype, and the Alaskan sedentary ecotype may have
arisen through similar processes of parallel phenotypic evolu-
tion.
Among ungulates, caribou and reindeer display high levels
of microsatellite heterozygosity (C^
ot
eet al., 2002; Boulet
et al., 2007). The extensive standing genetic variability in
Rangifer may be essential to the evolution of convergent phe-
notypes (Barrett & Schluter, 2008; Elmer & Meyer, 2011).
Understanding the source of variation (selection on new
mutations or pre-existing genetic variation) can help explain
how intraspecific variation is maintained in natural popula-
tions (Barrett & Schluter, 2008; Esp
ındola et al., 2012). Par-
allel phenotypic evolution may be common in Rangifer.
Genetic evidence suggests that Peary caribou (R. t. pearyi)
and Svalbard reindeer (R. t. platyrhynchus) may have con-
verged to a shared small-bodied, short-legged phenotype
from two evolutionary lineages (Gravlund et al., 1998). The
high arctic islands represent a severe and unpredictable envi-
ronment with selection pressures that could have indepen-
dently produced the phenotypically divergent characteristics
of the Peary and Svalbard animals (Flagstad & Røed, 2003).
Recent analysis suggests that mtDNA introgression (ad-
mixture of BEL and NAL) does not correspond to the pres-
ence of migratory behaviour in caribou (Kl€
utsch et al.,
2016). If ecotypic adaptations to different environments are
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
6
J. L. Polfus et al.
a result of parallel phenotypic evolution then some beha-
vioural traits, like migratory behaviour, may not match pat-
terns of neutral marker genetic structure (Pond et al., 2016).
Furthermore, unique phenotypes and behavioural adapta-
tions are likely to be influenced by behavioural plasticity,
pleiotropy or interacting gene pathways (R
eale et al., 2003;
Kopp, 2009). Cases of potential parallel evolution present an
ideal opportunity for future genomic research to illuminate
the genetic basis for adaptive traits (Elmer & Meyer, 2011).
For example, grey wolves (Canis lupus), like caribou, are
highly mobile and display divergent ecotypic adaptations
(Carmichael et al., 2001; Musiani et al., 2007). Recently,
Schweizer et al. (2016) used single-nucleotide polymorphisms
to examine phenotypic diversity in wolves and found pat-
terns of selection on morphological genes that were corre-
lated with environmental gradients suggesting that local
adaptation is important to ecotype divergence. Genomic
research in non-model species holds the promise of exposing
synergies among intraspecific diversity, local adaptations and
population persistence; however, real-world conservation
applications are still speculative (Shafer et al., 2015).
Timing of divergence
Phylogeographical reconstructions provide context for cur-
rent molecular patterns and allow for interpretation of the
impact of past climatic cycles on caribou (Flagstad & Røed,
2003; Kl€
utsch et al., 2012, 2016; Røed et al., 2014). Our ABC
analysis suggests that the BEL and NAL split c. 135.8 kyr bp
(CI: 68.6–173.6 kyr bp), which is comparable to Kl€
utsch
et al. (2016) at 97.3 kyr bp (CI: 44.6–135.8; combined
microsatellite and control region mtDNA). Our estimates are
more recent than those predicted by Yannic et al. (2014) at
300 kyr bp (184–430 kyr bp) using cyt bsequences and sig-
nificantly older than McDevitt et al. (2009) at 37.5 kyr bp
(CI: 28.1–46.7 kyr bp) using mtDNA control region. How-
ever, the coalescence estimates reveal that it is important to
consider multiple scales of cyclic climatic change, not just
the LGM (Barnosky, 2008). The interstadial periods of warm
climate between the early, middle and late Wisconsin glacial
periods likely resulted in reunification and introgression
between lineages (Fig. 3). In support of this assessment, an
ancient caribou mtDNA sample dated to 29,775 564
(IntCal09 years bp) recovered from the Yukon (Lorenzen
et al., 2011), is ancestral to the NAL (western clade sensu
Kl€
utsch et al., 2012), and suggests that potential connections
may have occurred prior to the LGM (Fig. 3).
Genetic diversification within refugia may be a source of
post-glacial variation in cold-adapted species (Weksler et al.,
2010). The palaeoenvironment of Beringia included pockets
of low-elevation spruce forests (especially during interglacials
and interstadials) among the extensive steppe-tundra and
grass-dominated ecosystem (Zazula et al., 2007). The internal
complexity of Beringia is thought to have influenced small
mammal diversity (Weksler et al., 2010; Galbreath et al.,
2011; Lanier et al., 2015), and could also have facilitated eco-
logical divergence of caribou. For example, our results reveal
that the split between the NWT boreal ecotype and barren-
ground caribou occurred prior to the LGM, which implies
that genetic subdivision likely persisted within Beringia.
Figure 3 Timeline of last 140 thousand calendar years (kyr) before present. Blue bubbles represent the estimates (t
1
–t
4
) associated with
the approximate Bayesian computation scenario 1 (found in Fig. 2). The timeline includes associated caribou (Rangifer tarandus)
histories in Canada (Lorenzen et al., 2011; Kl€
utsch et al., 2012; Letts et al., 2012), palaeogeographical events (Carlson, 2013; Dixon,
2015), palaeoenvironmental reconstructions (Vardy et al., 1998; Galloway et al., 2012) and glacial maps (Dyke, 2004) for North
America. The scale of the timeline shifts from 10-kyr increments to 1-kyr increments around the Last Glacial Maximum or at c. 20 kyr
before present. BEL –Beringian–Eurasian lineage, NA –North America, NAL –North American lineage, NWT –Northwest Territories,
YT –Yukon Territory.
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
7
Diversification in glacial refugia in a Holarctic mammal
While microgeographical adaptation to forested versus
steppe-tundra habitats may have played a critical role in the
development of caribou ecotypes during the Pleistocene,
there is also the possibility that the ancient lineages of NWT
boreal and barren-ground caribou experienced more pro-
nounced geographical separation associated with the divide
between the Eurasian and American landmasses.
The substantial sympatric phenotypic diversification in
caribou suggests that some genetic signals can withstand con-
tact zones. The Holocene has not been long enough for dis-
placement or admixture to completely mask the genetic
legacy of Pleistocene glacial vicariance in caribou. Interest-
ingly, while overlapping ranges (Roffler et al., 2012; Mager
et al., 2014) and large-scale merging between sedentary and
migratory herds are common in Alaska (Hinkes et al., 2005),
population merging between the boreal ecotype and barren-
ground caribou is not presently common in the western
Canadian boreal zone (Nagy et al., 2011). The genetic struc-
ture evident between barren-ground and the NWT boreal
ecotype suggests that any mixing that does occur is not suffi-
cient to prevent the perpetuation of distinct genetic signa-
tures (Appendix S1, Figs S1.5 & S1.6).
The clear microsatellite genetic structure across fine spa-
tial scales in central NWT are likely a result of ancestral
genetic signals and current ecological adaptations or beha-
vioural mechanisms that promote reproductive isolation
(Rundle & Nosil, 2005). The relatively low genetic diversity
in the modern NWT boreal ecotype may also suggest a
recent expansion into the ice-free region of central NWT
and potential founder effects. Likewise, the behaviours asso-
ciated with the boreal ecotype likely confer increased fitness
in the boreal forest, especially since similar phenotypes are
expressed by Eurasian forest reindeer and Alaskan sedentary
caribou. The genetic structure among neighbouring caribou
types suggests that microgeographical adaptation and its
driving mechanisms could promote the persistence of local
diversification (Rundle & Nosil, 2005; Richardson et al.,
2014). Future research is needed to understand how long
periods of isolation need to exist for genetic differentiation
to arise and remain divergent when contact is re-estab-
lished.
Similar to Weckworth et al. (2012) our results also contra-
dict the inclusion of the northern mountain ecotype of west-
ern Canada in the woodland subspecies. Our analysis
suggests that caribou in the Mackenzie Mountains arose
c. 4000 years bp from ancient BEL populations. This corre-
sponds to Letts et al. (2012) who found low mtDNA differ-
entiation between ancient (up to 3790 years bp) and modern
mountain caribou (Fig. 3). However, weak microsatellite
structure between the barren-ground and the northern
mountain ecotype implies that historic exchange or incom-
plete lineage sorting is influencing differentiation between
the groups (Letts et al., 2012; Polfus et al., 2016). In north-
ern Alberta, boreal ecotype caribou share BEL and NAL phy-
logeographical lineages (Weckworth et al., 2012) as do both
boreal and mountain ecotypes in the central Rockies which
suggest that zones of contact have occurred (McDevitt et al.,
2009). Future ancient DNA approaches may provide more
insight into the history of post-glacial contact and illuminate
geographical events that influenced population persistence at
transitional periods during the late Pleistocene and early
Holocene.
Conservation implications
We demonstrate that the boreal ecotype of caribou in
North America contains two phylogeographical assemblages
that compose an irreplaceable component of Canada’s bio-
diversity. Importantly, our results also show that southern
boreal ecotype animals belonging to the NAL represent an
independent evolutionary unit of caribou. As was initially
suggested by Geist (2007), protecting the ‘true woodland
caribou’ becomes even more critical if the group includes
only NAL animals along the southern edge of caribou dis-
tribution. The southern extent of the boreal forest also faces
threats related to anthropogenic disturbance, fragmentation
and shifting predatory–prey dynamics (Environment
Canada, 2012). Because the contiguous habitat of the boreal
forest and the dispersal capabilities of caribou are likely
critical components to the long-term persistence of the bor-
eal ecotype, the genetic variation in the boreal ecotype of
central NWT could help prevent the extinction of beha-
vioural adaptations in declining southern populations
through evolutionary rescue (Bell & Gonzalez, 2009). Fur-
thermore, as managers consider the re-introduction of pro-
grammes for declining caribou populations, our results
indicate that attention must be paid to the evolutionary
history of putative source populations.
Environmental change due to anthropogenic influence is
an increasing threat to many species, especially cold-adapted
species (Berteaux et al., 2004; Post et al., 2009). Rangifer’s
adaptation to a wide range of environments across the
Holarctic and continuance through the glacial cycles of the
Pleistocene suggests that a continuous geographical distribu-
tion and genetic mixing may be imperative to their success
(Hinkes et al., 2005; Boulet et al., 2007; Lorenzen et al.,
2011). In particular, caribou show substantial adaptive capac-
ity and potential phenotypic plasticity that seem to make the
species as a whole especially tolerant of changing conditions,
however, more information is needed to understand how
caribou will respond to future environmental change (Yannic
et al., 2014). Understanding the synergies between ecology
and evolution may facilitate the design of biodiversity con-
servation strategies for caribou that prepare for future
responses to restrictions on current interglacial climate refu-
gia (Stewart et al., 2010). Dividing species into units (sub-
species, ecotypes or DUs) that confine policies to particular
groups in isolation, may misrepresent genetic histories and
be an insufficient conservation approach. Rather, a focus on
large-scale eco-evolutionary processes could provide a frame-
work for maximizing genetic diversity and preserving adap-
tive potential.
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
8
J. L. Polfus et al.
ACKNOWLEDGEMENTS
The research was part of a collaborative community project.
The Ɂehdzo Got’i
zne
zand the Saht
u Renewable Resources
Board (SRRB) facilitated sample collection and contributed
financial, logistical and administrative support. We are grate-
ful to local participants, Parks Canada staff, SK Minister of
Environment staff and others for help collecting caribou
samples (full list of contributors included in Appendix S2).
Research was conducted under a multiyear research license
from the Aurora Research Institute, NWT Department of
Environment and Natural Resources (ENR) wildlife research
licenses, and a University of Manitoba Ethics Protocol.
Funding was provided by the SRRB, ENR, Cumulative
Impact Monitoring Program, Environmental Studies
Research Fund, Parks Canada, University of Manitoba, and
an NSERC Strategic Grant held by M.M. and P.J.W. J.L.P.
thanks Claire Polfus for reviewing the manuscript and the
Wilburforce Fellowship in Conservation Science.
REFERENCES
Barnosky, A.D. (2008) Climatic change, refugia, and biodi-
versity: where do we go from here? An editorial comment.
Climatic Change,86,29–32.
Barrett, R.D. & Schluter, D. (2008) Adaptation from standing
genetic variation. Trends in Ecology and Evolution,23,38–44.
Bell, G. & Gonzalez, A. (2009) Evolutionary rescue can pre-
vent extinction following environmental change. Ecology
Letters,12, 942–948.
Berteaux, D., R
eale, D., McAdam, A.G. & Boutin, S. (2004)
Keeping pace with fast climate change: can arctic life
count on evolution? Integrative and Comparative Biology,
44, 140–151.
Bishop, M.D., Kappes, S.M., Keele, J.W., Stone, R.T., Sun-
den, S., Hawkins, G.A., Toldo, S.S., Fries, R., Grosz, M.D.
& Yoo, J. (1994) A genetic linkage map for cattle. Genetics,
136, 619–639.
Bolnick, D.I., Amarasekare, P., Ara
ujo, M.S., B€
urger, R.,
Levine, J.M., Novak, M., Rudolf, V.H., Schreiber, S.J.,
Urban, M.C. & Vasseur, D.A. (2011) Why intraspecific
trait variation matters in community ecology. Trends in
Ecology and Evolution,26, 183–192.
Boulet, M., Couturier, S., C^
ot
e, S.D., Otto, R.D. & Ber-
natchez, L. (2007) Integrative use of spatial, genetic, and
demographic analyses for investigating genetic connectivity
between migratory, montane, and sedentary caribou herds.
Molecular Ecology,16, 4223–4240.
Brandt, J. (2009) The extent of the North American boreal
zone. Environmental Reviews,17, 101–161.
Carlson, A. (2013) The younger Dryas climate event. Encyclo-
pedia of Quaternary Science,3, 126–134.
Carmichael, L., Nagy, J., Larter, N. & Strobeck, C. (2001)
Prey specialization may influence patterns of gene flow in
wolves of the Canadian Northwest. Molecular Ecology,10,
2787–2798.
Cornuet, J.-M., Pudlo, P., Veyssier, J., Dehne-Garcia, A.,
Gautier, M., Leblois, R., Marin, J.-M. & Estoup, A. (2014)
DIYABC v2.0: a software to make approximate Bayesian
computation inferences about population history using
single nucleotide polymorphism. DNA sequence and
microsatellite data. Bioinformatics,30, 1187–1189.
COSEWIC (2011) Designatable units for caribou (Rangifer
tarandus) in Canada. Committee on the Status of Endan-
gered Wildlife in Canada, Ottawa, Ontario, Canada.
C^
ot
e, S., Dallas, J., Marshall, F., Irvine, R., Langvatn, R. &
Albon, S. (2002) Microsatellite DNA evidence for genetic
drift and philopatry in Svalbard reindeer. Molecular Ecol-
ogy,11, 1923–1930.
Cronin, M.A., MacNeil, M.D. & Patton, J.C. (2005) Varia-
tion in mitochondrial DNA and microsatellite DNA in
caribou (Rangifer tarandus) in North America. Journal of
Mammalogy,86, 495–505.
Csill
ery, K., Blum, M.G., Gaggiotti, O.E. & Franc
ois, O.
(2010) Approximate Bayesian computation (ABC) in prac-
tice. Trends in Ecology and Evolution,25, 410–418.
Dixon, E.J. (2015) Late Pleistocene colonization of North
America from Northeast Asia: new insights from large-
scale paleogeographic reconstructions. Mobility and ancient
society in Asia and the Americas (ed. by M.D. Frachetti
and R. Spengler III), pp. 169–184. Springer, New York,
NY, USA.
Dyke, A.S. (2004) An outline of North American deglaciation
with emphasis on central and northern Canada. Quater-
nary Glaciations: Extent and Chronology,2, 373–424.
Earl, D.A. & vonHoldt, B.M. (2012) Structure Harvester: a
website and program for visualizing structure output and
implementing the Evanno method. Conservation Genetics
Resources,4, 359–361.
Elmer, K.R. & Meyer, A. (2011) Adaptation in the age of
ecological genomics: insights from parallelism and conver-
gence. Trends in Ecology and Evolution,26, 298–306.
Environment Canada (2012) Recovery strategy for the wood-
land caribou (Rangifer tarandus caribou), boreal population,
in Canada. Species at Risk Act Recovery Strategy Series.
Environment Canada, Ottawa, Ontario, Canada. xi +
138pp.
Esp
ındola, A., Pellissier, L., Maiorano, L., Hordijk, W., Gui-
san, A. & Alvarez, N. (2012) Predicting present and future
intra-specific genetic structure through niche hindcasting
across 24 millennia. Ecology Letters,15, 649–657.
Fitzpatrick, S., Gerberich, J., Kronenberger, J., Angeloni, L. &
Funk, W. (2015) Locally adapted traits maintained in the
face of high gene flow. Ecology Letters,18,37–47.
Flagstad, Ø. & Røed, K.H. (2003) Refugial origins of reindeer
(Rangifer tarandus L.) inferred from mitochondrial DNA
sequences. Evolution,57, 658–670.
Galbreath, K.E., Cook, J.A., Eddingsaas, A.A. & DeChaine,
E.G. (2011) Diversity and demography in Beringia: multi-
locus tests of paleodistribution models reveal the complex
history of arctic ground squirrels. Evolution,65, 1879–
1896.
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
9
Diversification in glacial refugia in a Holarctic mammal
Galloway, J.M., Adamczewski, J., Schock, D.M., Andrews,
T.D., MacKay, G., Bowyer, V.E., Meulendyk, T., Moor-
man, B.J. & Kutz, S.J. (2012) Diet and habitat of moun-
tain woodland caribou inferred from dung preserved in
5000-year-old alpine ice in the Selwyn Mountains, North-
west Territories, Canada. Arctic,65,59–79.
Galpern, P., Manseau, M. & Wilson, P. (2012a) Grains of
connectivity: analysis at multiple spatial scales in landscape
genetics. Molecular Ecology,21, 3996–4009.
Galpern, P., Manseau, M., Hettinga, P., Wilson, P. & Smith,
K. (2012b) ALLELEMATCH: an R package for identifying
unique multilocus genotypes where genotyping error and
missing data may be present. Molecular Ecology Resources,
12, 771–778.
Geist, V. (2007) Defining subspecies, invalid taxonomic tools,
and the fate of the woodland caribou. Rangifer Special
Issue,17,25–28.
Gravlund, P., Meldgaard, M., P€
a€
abo, S. & Arctander, P.
(1998) Polyphyletic origin of the small-bodied, high-arctic
subspecies of tundra reindeer (Rangifer tarandus). Molecu-
lar Phylogenetics and Evolution,10, 151–159.
Hall, T.A. (1999) BioEdit: a user-friendly biological sequence
alignment editor and analysis program for Windows 95/
98/NT. Nucleic Acids Symposium Series,41,95–98.
Hardy, O.J. & Vekemans, X. (2002) SPAGeDi: a versatile
computer program to analyse spatial genetic structure at
the individual or population levels. Molecular Ecology
Notes,2, 618–620.
Hewitt, G. (2000) The genetic legacy of the Quaternary ice
ages. Nature,405, 907–913.
Hinkes, M.T., Collins, G.H., Van Daele, L.J., Kovach, S.D.,
Aderman, A.R., Woolington, J.D. & Seavoy, R.J. (2005)
Influence of population growth on caribou herd identity,
calving ground fidelity, and behavior. Journal of Wildlife
Management,69, 1147–1162.
Kl€
utsch, C.F.C., Manseau, M. & Wilson, P.J. (2012) Phylo-
geographical analysis of mtDNA data indicates postglacial
expansion from multiple glacial refugia in woodland cari-
bou (Rangifer tarandus caribou). PLoS ONE,7, e52661.
Kl€
utsch, C.F.C., Manseau, M., Trim, V., Polfus, J.L. & Wil-
son, P.J. (2016) The eastern migratory caribou: the role of
genetic introgression in ecotype evolution. Royal Society
Open Science,3, 150469.
Kopp, A. (2009) Metamodels and phylogenetic replication: a
systematic approach to the evolution of developmental
pathways. Evolution,63, 2771–2789.
Lanier, H.C., Gunderson, A.M., Weksler, M., Fedorov, V.B.
& Olson, L.E. (2015) Comparative phylogeography high-
lights the double-edged sword of climate change faced by
arctic and alpine-adapted mammals. PLoS ONE,10,
e0118396.
Letts, B., Fulton, T.L., Stiller, M., Andrews, T.D., MacKay,
G., Popko, R. & Shapiro, B. (2012) Ancient DNA reveals
genetic continuity in mountain woodland caribou of the
Mackenzie and Selwyn Mountains, Northwest Territories,
Canada. Arctic,65,80–94.
Librado, P. & Rozas, J. (2009) DnaSP v5: a software for com-
prehensive analysis of DNA polymorphism data. Bioinfor-
matics,25, 1451–1452.
Lopes, J.S. & Boessenkool, S. (2010) The use of approximate
Bayesian computation in conservation genetics and its
application in a case study on yellow-eyed penguins. Con-
servation Genetics,11, 421–433.
Lorenzen, E.D., Nogu
es-Bravo, D., Orlando, L., Weinstock,
J., Binladen, J., Marske, K.A., Ugan, A., Borregaard, M.K.,
Gilbert, M.T.P. & Nielsen, R. (2011) Species-specific
responses of Late Quaternary megafauna to climate and
humans. Nature,479, 359–364.
Mager, K.H., Colson, K.E., Groves, P. & Hundertmark, K.J.
(2014) Population structure over a broad spatial scale dri-
ven by nonanthropogenic factors in a wide-ranging migra-
tory mammal, Alaskan caribou. Molecular Ecology,23,
6045–6057.
McDevitt, A.D., Mariani, S., Hebblewhite, M., Decesare, N.J.,
Morgantini, L., Seip, D., Weckworth, B.V. & Musiani, M.
(2009) Survival in the Rockies of an endangered hybrid
swarm from diverged caribou (Rangifer tarandus) lineages.
Molecular Ecology,18, 665–679.
Musiani, M., Leonard, J.A., Cluff, H., Gates, C.C., Mariani,
S., Paquet, P.C., Vil
a, C. & Wayne, R.K. (2007) Differenti-
ation of tundra/taiga and boreal coniferous forest wolves:
genetics, coat colour and association with migratory cari-
bou. Molecular Ecology,16, 4149–4170.
Nagy, J.A., Johnson, D.L., Larter, N.C., Campbell, M.W.,
Derocher, A.E., Kelly, A., Dumond, M., Allaire, D. &
Croft, B. (2011) Subpopulation structure of caribou (Ran-
gifer tarandus L.) in arctic and subarctic Canada. Ecological
Applications,21, 2334–2348.
Polfus, J.L., Manseau, M., Simmons, D., Neyelle, M., Bayha,
W., Andrew, F., Andrew, L., Kl€
utsch, C.F.C., Rice, K. &
Wilson, P. (2016) Łegh
agots’enete
z(learning together): the
importance of indigenous perspectives in the identification
of biological variation. Ecology and Society,21(2), 18.
Pond, B.A., Brown, G.S., Wilson, K.S. & Schaefer, J.A. (2016)
Drawing lines: spatial behaviours reveal two ecotypes of
woodland caribou. Biological Conservation,194,
139–148.
Post, E., Forchhammer, M.C., Bret-Harte, M.S. et al. (2009)
Ecological dynamics across the arctic associated with
recent climate change. Science,325, 1355–1358.
Pritchard, J.K., Stephens, M. & Donnelly, P. (2000) Inference
of population structure using multilocus genotype data.
Genetics,155, 945–959.
Puckett, E.E., Etter, P.D., Johnson, E.A. & Eggert, L.S. (2015)
Phylogeographic analyses of American black bears (Ursus
americanus) suggest four glacial refugia and complex pat-
terns of post-glacial admixture. Molecular Biology and Evo-
lution,32, 2338–2350.
R
eale, D., McAdam, A.G., Boutin, S. & Berteaux, D. (2003)
Genetic and plastic responses of a northern mammal to
climate change. Proceedings of the Royal Society B: Biologi-
cal Sciences,270, 591–596.
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
10
J. L. Polfus et al.
Richardson, J.L., Urban, M.C., Bolnick, D.I. & Skelly, D.K.
(2014) Microgeographic adaptation and the spatial scale of
evolution. Trends in Ecology and Evolution,29, 165–176.
Røed, K.H., Bjørnstad, G., Flagstad, Ø., Haanes, H.,
Hufthammer, A.K., Jordhøy, P. & Rosvold, J. (2014)
Ancient DNA reveals prehistoric habitat fragmentation
and recent domestic introgression into native wild rein-
deer. Conservation Genetics,15, 1137–1149.
Roffler, G.H., Adams, L.G., Talbot, S.L., Sage, G.K. & Dale,
B.W. (2012) Range overlap and individual movements
during breeding season influence genetic relationships of
caribou herds in south-central Alaska. Journal of Mammal-
ogy,93, 1318–1330.
Rosenblum, E.B., Parent, C.E. & Brandt, E.E. (2014) The
molecular basis of phenotypic convergence. Annual Review
of Ecology, Evolution, and Systematics,45, 203–226.
Rousset, F. (2008) Genepop’007: a complete reimplementa-
tion of the Genepop software for Windows and Linux.
Molecular Ecology Resources,8, 103–106.
Rundle, H.D. & Nosil, P. (2005) Ecological speciation. Ecol-
ogy Letters,8, 336–352.
Schluter, D., Clifford, E.A., Nemethy, M. & McKinnon, J.S.
(2004) Parallel evolution and inheritance of quantitative
traits. The American Naturalist,163, 809–822.
Schweizer, R.M., vonHoldt, B.M., Harrigan, R., Knowles,
J.C., Musiani, M., Coltman, D., Novembre, J. & Wayne,
R.K. (2016) Genetic subdivision and candidate genes
under selection in North American grey wolves. Molecular
Ecology,25, 380–402.
Serrouya, R., Paetkau, D., McLellan, B.N., Boutin, S., Camp-
bell, M. & Jenkins, D.A. (2012) Population size and major
valleys explain microsatellite variation better than taxo-
nomic units for caribou in western Canada. Molecular
Ecology,21, 2588–2601.
Shafer, A.B., Wolf, J.B., Alves, P.C., Bergstr€
om, L., Bruford,
M.W., Br€
annstr€
om, I., Colling, G., Dalen, L., De Meester,
L. & Ekblom, R. (2015) Genomics and the challenging
translation into conservation practice. Trends in Ecology
and Evolution,30,78–87.
Stewart, J.R., Lister, A.M., Barnes, I. & Dal
en, L. (2010)
Refugia revisited: individualistic responses of species in
space and time. Proceedings of the Royal Society of London
B: Biological Sciences,277, 661–671.
Vardy, S.R., Warner, B.G. & Aravena, R. (1998) Holocene
climate and the development of a subarctic peatland near
Inuvik, Northwest Territories, Canada. Climatic Change,
40, 285–313.
Weckworth, B.V., Musiani, M., McDevitt, A.D., Hebblewhite,
M. & Mariani, S. (2012) Reconstruction of caribou evolu-
tionary history in western North America and its implica-
tions for conservation. Molecular Ecology,21, 3610–3624.
Weksler, M., Lanier, H.C. & Olson, L.E. (2010) Eastern Ber-
ingian biogeography: historical and spatial genetic struc-
ture of singing voles in Alaska. Journal of Biogeography,37,
1414–1431.
Wilson, G., Strobeck, C., Wu, L. & Coffin, J. (1997) Charac-
terization of microsatellite loci in caribou Rangifer taran-
dus, and their use in other artiodactyls. Molecular Ecology,
6, 697–699.
Yannic, G., Pellissier, L., Ortego, J. et al. (2014) Genetic
diversity in caribou linked to past and future climate
change. Nature Climate Change,4, 132–137.
Zazula, G.D., Froese, D.G., Elias, S.A., Kuzmina, S. & Math-
ewes, R.W. (2007) Arctic ground squirrels of the mam-
moth-steppe: paleoecology of Late Pleistocene middens
(24000–29450 14 C yr BP), Yukon Territory, Canada.
Quaternary Science Reviews,26, 979–1003.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the
online version of this article:
Appendix S1 Population genetic summary statistics and
ABC model evaluations.
Appendix S2 Full list of collaborators and contributors.
DATA ACCESSIBILITY
Genetic mtDNA data generated for this study are available
on GenBank: KX463387–KX463407.
BIOSKETCH
The team is interested in developing collaborative applied
research that integrates population genetics and ecological
data to monitor and study the origin, genetic diversity, status
and evolutionary history of wildlife populations. Their exper-
tise includes the use of non-invasive methods, DNA markers
and cross-cultural approaches in the management of wildlife
and protected areas. The team’s goals include developing
new analytical methods, synthesizing knowledge across juris-
dictional, cultural and political boundaries, and approaching
conservation problems in a way that respects the lives and
experiences of people that depend on natural resources for
their livelihood. You can find more at http://lecol-ck.ca and
http://srrb.nt.ca.
Author contributions: J.L.P. collected field samples, con-
ducted research, contributed to research design, analysis and
data interpretation, and wrote the first draft of the manu-
script. M.M., P.J.W. and D.S. contributed to analysis
approach, study design, and data interpretation. C.F.C.K.
performed the ABC analysis and aided in data interpretation.
All authors provided revisions to the manuscript and gave
final approval.
Editor: Robert Bryson
Journal of Biogeography
ª2016 The Authors. Journal of Biogeography Published by John
Wiley & Sons Ltd.
11
Diversification in glacial refugia in a Holarctic mammal
