Comparative phylogeography reveals a shared impact of ... · ROBERT VERITY,* ANIL PRAKASH,–– PE THAN HTUN,* THAUNG HLAING,* SIMONE NAMBANYA,††† DUONG SOCHEAT,‡‡‡

Molecular Ecology (2011) 20, 4533–4549 doi: 10.1111/j.1365-294X.2011.05268.x

Comparative phylogeography reveals a shared impact ofpleistocene environmental change in shaping geneticdiversity within nine Anopheles mosquito species acrossthe Indo-Burma biodiversity hotspot

KATY MORGAN,*†1 SAMANTHA M. O’LOUGHLIN,*‡1 BIN CHEN,§ YVONNE-MARIE LINTON,–

DAMRONGPAN THONGWAT,** PRADYA SOMBOON,†† MUN YIK FONG,‡‡ ROGER BUTLIN,§§

ROBERT VERITY,* ANIL PRAKASH,–– PE THAN HTUN,*** THAUNG HLAING,*** SIMONE

NAMBANYA,††† DUONG SOCHEAT,‡‡‡ TRUNG HO DINH§§§ and CATHERINE WALTON*

*Faculty of Life Sciences, University of Manchester, Manchester M13 9PT, UK, †Max Planck Institute for Developmental

Biology, 72076 Tuebingen, Germany, ‡NERC Centre for Population Biology, Imperial College London, Silwood Park, Ascot,

Berkshire SL5 7PY, UK, §College of Life Sciences, Chongqing Normal University, Chongqing 400047, China, –Natural History

Museum, London SW7 5BD, UK, **Department of Microbiology and Parasitolgy, Faculty of Medical Sciences, Naresuan

University, Phitsanulok 65000, Thailand, ††Department of Parasitology, Faculty of Medicine, Chiang Mai University, Chiang

Mai 50200, Thailand, ‡‡Department of Parasitology, University of Malaya, 50603 Kuala Lumpur, Malaysia, §§Department of

Animal and Plant Sciences, The University of Sheffield, Sheffield S10 2TN, UK, ––Regional Medical Research Centre, Dibrugarh

786001, Assam, India, ***Department of Medical Research, Yangon 11191, Myanmar, †††Centre of Malariology, Parasitology

and Entomology, Vientiane, Lao People‘s Democratic Republic, ‡‡‡National Centre for Malaria, Parasitology and Entomology,

Phnom Penh, Cambodia, §§§National Institute of Malariology, Parasitology and Entomology (NIMPE), Hanoi, Vietnam

Corresponde1These autho

� 2011 Black

Abstract

South-East Asia is one of the world’s richest regions in terms of biodiversity. An

understanding of the distribution of diversity and the factors shaping it is lacking, yet

essential for identifying conservation priorities for the region’s highly threatened biodiver-

sity. Here, we take a large-scale comparative approach, combining data from nine forest-

associated Anopheles mosquito species and using statistical phylogeographical methods to

disentangle the effects of environmental history, species-specific ecology and random

coalescent effects. Spatially explicit modelling of Pleistocene demographic history supports

a common influence of environmental events in shaping the genetic diversity of all species

examined, despite differences in species’ mtDNA gene trees. Populations were periodically

restricted to allopatric northeastern and northwestern refugia, most likely due to Pleistocene

forest fragmentation. Subsequent southwards post-glacial recolonization is supported by a

north–south gradient of decreasing genetic diversity. Repeated allopatric fragmentation and

recolonization have led to the formation of deeply divergent geographical lineages within

four species and a suture zone where these intraspecific lineages meet along the Thai–

Myanmar border. A common environmental influence for this divergence was further

indicated by strong support for simultaneous divergence within the same four species,

dating to approximately 900 thousand years ago (kya). Differences in the geographical

structuring of genetic diversity between species are probably the result of varying species’

biology. The findings have important implications for conservation planning; if the refugial

regions and suture zone identified here are shared by other forest taxa, the unique and high

levels of genetic diversity they house will make these areas conservation priorities.

Keywords: Anopheles, MsBayes, Phylogeography, Population genetics, South-East Asia, SPLATCHE

Received 6 April 2011; revision received 9 July 2011; accepted 20 July 2011

nce: Catherine Walton, Fax: +161 275 5082; E-mail: [email protected]

rs contributed equally to this work and are joint first authors.

well Publishing Ltd

4534 K. MORGAN ET AL.

Introduction

Considerable efforts have been made to explain the pro-

cesses of diversification in the Tropics, which harbour

the vast majority of the World’s biodiversity (reviewed

in Gaston 2000). South-East Asia encompasses four bio-

diversity hot spots: Wallacea, The Philippines, Sunda-

land and Indo-Burma (Myers et al. 2000). Of these, the

Indo-Burmese hot spot covers the greatest area (all of

mainland South-East Asia excluding peninsular Malay-

sia) yet is the least well studied. The dynamic climatic

and environmental history of South-East Asia is likely

to have played a key role in shaping the distribution of

biodiversity across the region (Sodhi et al. 2004); an

understanding of this is not only of fundamental scien-

tific interest but also important for specific potential

applications to malaria control and more generally for

the long-term conservation of biodiversity. The matter

is urgent as most tropical diversity is contained within

forests yet recent rates of deforestation are reportedly

higher in South-East Asia than in any other tropical

region (Achard et al. 2002). Biodiversity studies are

important for discovering unique genetic lineages and

cryptic species, but also to identify areas and landscape

features that have been influential in driving diversifica-

tion, so that processes of diversification may continue

in the future (Moritz et al. 2000; Moritz 2002). Despite

mainland South-East Asia’s rich biodiversity and fragile

conservation status, phylogeographical studies into the

origin and distribution of tropical biodiversity have

mainly focused on other regions (Fjeldsa 1994; Fjeldsa

& Lovett 1997; Colinvaux et al. 2000; Moritz et al. 2000;

Hewitt 2004). Here, we investigate the biogeography of

mainland South-East Asia using comparative phyloge-

ography, a powerful approach for determining the dif-

ferential effects of environment and species-specific

ecologies on the distribution of intraspecific genetic

diversity across a landscape (Bermingham & Moritz

1998; Carnaval et al. 2009).

Two major historical environmental factors that

potentially shape genetic diversity are climatic and tec-

tonic change (Crisci et al. 2003). South-East Asia has a

very complex tectonic history involving collisions of the

Indian and Australian plates into the South-East Asian

plate during the Cenozoic with subsequent movements

of these and other minor plates (Hall 2002). Tectonic

activity in the Pliocene and Pleistocene, i.e. on the time

frame where we might expect to detect effects at the

intraspecific level, is confined largely to the insular

region where it resulted in island formation and uplift.

This has primarily generated diversity at the species,

rather than intraspecific, level because of dispersal and

allopatric isolation on islands (e.g. Steppan et al. 2003;

Esselstyn et al. 2009). On the mainland, a more impor-

tant factor shaping intraspecific genetic diversity is

likely to be the periodic climatic fluctuations of the

Pleistocene, and this is the hypothesis we test here. In

more extreme latitudes, glacial maxima were experi-

enced as the extension of ice sheets, and in South-East

Asia, they resulted in substantial forest contraction and

fragmentation (Heaney 1991; Hewitt 2004; Hope et al.

2004).

The importance of the Pleistocene climatic fluctua-

tions and associated changes in tropical forest cover in

driving tropical diversification has long been debated,

primarily in relation to neotropical diversity (Haffer

1969; Haffer & Prance 2001). According to the ‘Pleisto-

cene refuge’ hypothesis, during the arid, cool periods of

the Pleistocene, the fragmentation of tropical forest hab-

itat isolated populations in allopatric refugia, with con-

sequent divergence because of genetic drift and local

adaptation (Haffer & Prance 2001). This hypothesis has

been disputed as a cause of neotropical diversity,

because of palynological evidence supporting the exis-

tence of continuous forest habitat throughout South

America during the Pleistocene (Colinvaux et al. 2000,

2001; Mayle et al. 2004). This may, however, not be the

case in South-East Asia. Palynological and sedimento-

logical data indicate substantial reductions in tropical

forest cover across South-East Asia during the last gla-

cial maximum (LGM), accompanied by expansions of

grassland and savannah habitat (van der Kaars 1991;

Kealhofer & Penny 1998; Penny 2001; Hope et al. 2004;

White et al. 2004). Aridity levels across South-East Asia,

and hence the fragmentation of tropical forest habitat,

were amplified by the large impact of glacio-eustatic

sea level changes on landmass configuration during the

Pleistocene (Heaney 1991). These sea level fluctuations

caused the repeated exposure and submergence of land

bridges between the mainland and insular regions

(Voris 2000). This not only drove dispersal and diver-

gence in land-based fauna (Ziegler et al. 2007; Reddy

2008) but also reduced the surface area of ocean across

which evaporation could occur, so lowering the mois-

ture content of the monsoon rains (Heaney 1991).

Because of the relatively small number of phylogeo-

graphical studies conducted in South-East Asia, it is dif-

ficult to identify the general biogeographical patterns in

the region. However, the wide distribution of several

forest-associated taxa (Colobinae monkeys (Brandon-

Jones 1996), Asian elephants (Vidya et al. 2009), dholes

(Iyengar et al. 2005), piculet birds (Fuchs et al. 2008)

and Anopheles mosquitoes (Morgan et al. 2009; Chen

et al. 2011) from Sri Lanka to Vietnam coupled with the

presence of distinct, divergent lineages within them,

indicates that forests throughout this region have some-

times been continuous but at other times fragmented.

Where examined, divergence dated to within the

� 2011 Blackwell Publishing Ltd

PHYLOGEOGRAP HY OF ANOPHELES IN IND O-BURM A 4535

Pleistocene and has been attributed to the isolation of

populations within allopatric forest refugia during gla-

cial maxima. Although refugial regions have been pro-

posed in Sri Lanka ⁄ southern India (Morgan et al. 2009;

Vidya et al. 2009), Nepal (Fuchs et al. 2008), northeast-

ern India (O’Loughlin et al. 2008; Morgan et al. 2010),

Myanmar (Fuchs et al. 2008), Yunnan (Chen et al. 2004,

2011) and eastern South-East Asia (O’Loughlin et al.

2008; Morgan et al. 2009, 2010; Chen et al. 2011), there

is no firm consensus as to the likely number or location

of these refugia.

Further phylogeographical studies to identify regions

of high and ⁄ or unique genetic diversity could greatly

benefit ecosystem-based conservation planning in

South-East Asia. However, it can be difficult to perform

these directly on species of conservation concern,

because of their rarity and often incomplete taxonomic

and distributional information. In the Wet Tropics of

Australia, biogeographical patterns within a wide range

of invertebrate taxa strongly predicted conservation pri-

orities in vertebrate species, indicating that invertebrate

taxa may be used as ‘taxon surrogates’ for endangered

species (Moritz et al. 2001). Because of the medical

importance of Anopheles mosquitoes, the taxonomy, dis-

tribution and ecological characteristics of many species

are relatively well documented (Manguin et al. 2008).

Several Anopheles species show strong association with

tropical forest habitats (Reid 1968), and previous phylo-

geographical studies have supported a role of environ-

mental change in shaping genetic diversity within

Anopheles mosquitoes (Chen et al. 2004, 2011; O’Lough-

lin et al. 2008; Morgan et al. 2009, 2010). Forest-depen-

dent mosquitoes are therefore likely to be a good model

taxon for elucidating the processes underlying, and pat-

terns of, co-distributed forest biodiversity in South-East

Asia.

Under the Pleistocene refuge hypothesis, the genetic

diversity of forest taxa is expected to have been shaped

by a common environmental influence, the fragmenta-

tion of forest habitat during periods of increased arid-

ity. Traditionally, identifying a shared environmental

influence on genetic diversity has involved looking for

common patterns across the gene trees of different spe-

cies (e.g. Schneider et al. 1998; Feldman & Spicer 2006).

Species are expected to show a genetic signature of

their population history in the distribution of genetic

diversity among populations; additionally, where there

are distinct geographical lineages within species, these

are expected to be both spatially congruent across spe-

cies and to have simultaneous divergence times (Endler

1982; Moritz et al. 2009). However, because of the sto-

chastic nature of mutational, coalescent and demo-

graphic processes, similar population histories may give

rise to very different gene trees, and vice versa (Nielsen


& Beaumont 2009). Because of the variance expected in

the time to the most recent common ancestor of diver-

gent lineages, according to coalescent theory, comparing

independently estimated divergence times across

multiple taxon pairs may be misleading (Nielsen &

Beaumont 2009). Several recently developed coalescent-

based techniques account for the inherent stochasticity

of gene tree coalescence, enable alternative evolutionary

scenarios to be statistically evaluated and in some cases

allow multi-taxa data sets to be analysed simulta-

neously (Knowles 2004; Hickerson et al. 2006, 2007;

Nielsen & Beaumont 2009). These latest methodologies,

combined with comparative phylogeography, provide

an ideal toolkit for testing the expectations of the Pleis-

tocene refuge hypothesis described earlier.

Our previous studies of Anopheles species have sug-

gested some similarity in cross-species phylogeographi-

cal patterns, notably in east–west divergence within

mainland South-East Asia: in intraspecific geographical

lineages of An. minimus (Chen et al. 2011), An. annular-

is and An. splendidus (Morgan et al. 2009); and the allo-

patric divergence and subsequent speciation of

An. baimaii and An. dirus (Morgan et al. 2010). How-

ever, the degree of congruence between phylogeograph-

ical patterns of different species, and the likely locations

of Pleistocene forest refugia that underlie these patterns,

is unclear. Here, we draw together mitochondrial CO2

sequences from 1032 specimens comprising nine

mosquito species, sampled throughout mainland South-

East Asia. In addition to four newly sampled species

(An. aconitus, An. philippinensis, An. maculatus and

An. sawadwongporni), we have substantially extended

the geographical sampling of five previously studied

species (An. annularis, An. baimaii, An. splendidus,

An. jeyporiensis and An. minimus), most notably from

the little studied regions of northeastern India and

Myanmar. The use of multiple species counteracts the

potential problem of inferences from a single marker

being owing to locus-specific effects, because any

shared patterns of genetic diversity must necessarily

indicate a common cause. To our knowledge, this

study, with its use of multiple species and extensive

geographical coverage, is the first large-scale phylogeo-

graphical study performed in this region.

Five of the analysed species (An. minimus, An. bai-

maii, An. maculatus, An. sawadwongporni and An. splen-

didus) show a strong association with forest habitat;

Anopheles annularis, An. philippinensis and An. aconitus

can also be found in forest habitat; however, they are

also common in deforested habitats such as rice pad-

dies; An. jeyporiensis prefers hilly and mountainous

areas and is limited to northern latitudes (Covell 1927;

Reid 1968; Darsie & Pradhan 1990; Rattanarithikul et al.

2006). The sampled species come from three Series (an


informal taxonomic grouping) within the subgenus Cel-

lia (genus Anopheles): the Neomyzomyia Series (Anophe-

les baimaii); the Myzomyia Series (An. minimus,

An. aconitus and An. jeyporiensis all from the Funestus

Group); and the Neocellia Series (An. maculatus,

An. sawadwongporni, An. annularis, An. philippinensis

and An. splendidus (Garros et al. 2005; Harbach & Kit-

ching 2005). Within the Neocellia Series, Anopheles mac-

ulatus and An. sawadwongporni are most closely related

to each other, both being classified within the Macula-

tus Group, and An. philippinensis and An. annularis are

most closely related, both being within the Annularis

Subgroup. Anopheles splendidus (within the Splendidus

Subgroup) is relatively distantly related to these group-

ings (Harbach & Kitching 2005; Morgan et al. 2009).

Shared association with forest habitat is therefore not

particularly dependent upon taxonomic affinity of the

species.

This combined data set provides the opportunity to

differentiate the effects of shared environmental and

species-specific ecological effects on population history.

We will test two key predictions of the refuge hypothe-

sis: congruent patterns of genetic diversity and simulta-

neous vicariance of geographical lineages across

species. Differentiation between alternate biogeographi-

cal scenarios is made possible by the use of spatially

explicit coalescent-based modelling techniques and hier-

archical approximate Bayesian computation (hABC)

An. maculatusAn. aconitus

An. annularis

An. minimus

An. philippinensisAn. baimaii

An. sawadwongporniAn. splendidusAn. jeyporiensis

Location of modelledwestern refugia

Location of modelledeastern refugia

Fig. 1 Topographical map of mainland South-East Asia, indicating th

analysis, as implemented in the SPLATCHE (Currat

et al. 2004) and MsBayes (Hickerson et al. 2007) soft-

ware, respectively.

Materials and methods

Data collection

Nine Anopheles mosquito species were sampled from

across South-East Asia (Fig. 1; Table S1, Supporting

information). Some 499 sequences of the mitochondrial

CO2 gene were included from previous studies; 148

An. baimaii (O’Loughlin et al. 2008), 164 An. minimus

(Chen et al. 2011), 107 An. annularis (Morgan et al.

2009), 9 An. splendidus (Morgan et al. 2009) and 71

An. jeyporiensis individuals (Chen et al. 2004), and the

CO2 gene fragment was also sequenced from a further

533 Anopheles mosquitoes. This included many of the

same species from previously unsampled locations: 8

An. annularis; 37 An. baimaii and 15 An. jeyporiensis

from Myanmar; and 24 An. minimus from Myanmar,

Laos and Cambodia. In addition, sequences from four

previously unsampled species were included: 140

An. aconitus, 169 An. maculatus, 66 An. sawadwongporni

and 74 An. philippinensis from 16, 25, 9 and 7 geograph-

ical locations across South-East Asia, respectively (see

Fig. 1 and Table S1, Supporting information). Not all

species were present or available from each locality

e sampled locations for each species within this study.



sampled. Sequences generated in this study are avail-

able in GenBank under accession numbers HQ403680–

HQ404165.

DNA was extracted from individual mosquitoes

using the phenol–chloroform method (Sambrook 1989).

A 636–730 bp fragment of the CO2 gene was amplified

using the primers leu and lys (Sharpe et al. 2000),

according to the PCR protocol detailed within Morgan

et al. (2009). Sequences were analysed and trimmed

using Sequencher (Gene codes, Ann Arbor, USA) and

aligned in Clustal X (Thompson et al. 1997). No indels

were present within alignments; thus, alignments were

unambiguous.

Haplotype networks and summary statistics

An appropriate evolutionary model and associated

parameters were determined for each intraspecific

alignment using MODELTEST (Posada & Crandall

1998). Haplotype networks were produced using NET-

WORK 4.51 (Bandelt et al. 1999). After specifying the

evolutionary model parameters identified by MODEL-

TEST, FST matrices and hp and hs diversity statistics

were calculated for each intraspecific data set using Ar-

lequin 3.11 (Excoffier & Schneider 2005). To ensure that

estimates of population differentiation and diversity

were reliable, only populations with sample sizes of five

or more were included in the analysis. To assess the

similarity of geographical patterns of genetic structure

between species, the matrices of population pairwise

FST values were compared for all species pairs that

shared at least five sampling locations. The comparisons

were made using partial Mantel tests, as performed in

Arlequin 3.11, to remove any confounding effects of iso-

lation by distance.

Bayesian skyline plots

If species have experienced population bottlenecks dur-

ing glacial periods and undergone subsequent range

expansions during interglacial periods, because of the

expansion of forest habitat, they are expected to show

signals of population growth with a shared date of

onset. The hypothesis of recent population growth was

tested for each species through the construction of

Bayesian skyline plots (BSP) (Heled & Drummond

2008), using the software BEAST (Drummond & Rambaut

2007). The BSP method enables the inference of popula-

tion demographic history without the prior specification

of a restrictive and potentially inappropriate parametric

model (Heled & Drummond 2008). For the species with

deeply divergent lineages (An. annularis, An. minimus

and An. philippinensis), each lineage was analysed indi-

vidually, as pronounced genetic structure can bias


growth estimates. Anopheles splendidus and the western

lineage of An. philippinensis were omitted from the

Bayesian skyline analysis because of low sample num-

bers. To estimate the timing of changes in effective pop-

ulation size, the 2.3% divergence per million years rate

of mitochondrial evolution, as estimated using a range

of arthropod taxa (Brower 1994), was applied. Analyses

were performed three times to confirm convergence.

Spatially explicit modelling

To test rigorously the hypothesis that the evolutionary

histories of all species are consistent with a shared his-

tory of environmental change, we performed spatially

explicit modelling of post-glacial range expansion under

a range of evolutionary scenarios, using the software

SPLATCHE 1.1 (Currat et al. 2004). SPLATCHE simu-

lates in a spatially explicit manner demographic and

spatial expansions, from one or several origins, for-

wards in time. Prior to the modelled expansion, a per-

iod of stability may be included, with specified levels of

migration between origin populations. The modelled

expansion and the pre-expansion period of stability are

hereafter referred to as the expansion and the stability

phases, respectively. After the expansion phase, coales-

cent theory is used to generate simulated genealogies

backwards in time, until the most recent common

ancestor of all individuals is reached.

To generate levels of genetic diversity in the simu-

lated data that are comparable to those in the real data,

all models included a stability phase, from the mid-

Pleistocene [600 thousand years ago (kya)] until the

start of the expansion phase, 100 kya. Several alterna-

tive geographical locations were used for the origin

populations (Fig. 1) to reflect different putative loca-

tions for glacial refugia suggested by the literature (see

Introduction). Where multiple refugia were included in

the model, levels of migration between refugial regions

during the stability phase were varied, from 1 · 10)2 to

1 · 10)5 migrants per generation, to reflect periodic

gene flow between allopatric populations following

interglacial expansion and secondary contact. Following

the stability phase, the expansion phase involved the

modelling of each demographic and spatial population

expansion scenario from 100 kya through the penulti-

mate interglacial and LGM, into the current interglacial

until the present time. Because SPLATCHE has high

memory requirements, the range expansion was mod-

elled in steps of 200 generations. A generation time of

0.1 years, as has been estimated for several Anopheles

mosquito species (e.g. Maharaj 2003), was assumed. The

expansion was modelled from 100 000 ya until present

time; thus, a total of 5000 steps were performed.

Because SPLATCHE is only able to model population


expansions, it was necessary to maintain a glacial cli-

mate until the onset of the most recent interglacial. This

is considered analogous to populations repeatedly colo-

nizing and becoming extinct from the low-altitude

regions during each of the proceeding interglacials. A

switch from a glacial to an interglacial climate was initi-

ated after 4250 steps, corresponding to the onset of the

last deglaciation in the northern hemisphere, 15 kya

(Clark et al. 2009).

During the expansion phase, SPLATCHE allows the

incorporation of relevant environmental information

into the model. The species within this study exhibit

varying degrees of forest dependency and association;

thus, forest cover is likely to have been important for

population survival and dispersal. As the available pal-

ynological data provide little specific information

regarding the distribution of habitat types across South-

East Asia during the last 100 ky (van der Kaars 1991;

Penny 2001; Hope et al. 2004; White et al. 2004), altitude

was used as a proxy for Pleistocene forest cover. The

rationale for this is that mid-elevation regions are likely

to have retained sufficient moisture for forest habitats

to survive during the glacial periods as they are suffi-

ciently high to intercept the monsoon rains and fog

with subsequent run-off of moisture, without being so

high that low temperatures prevent forest growth (Hea-

ney 1991; Gathorne-Hardy et al. 2002). Topographical

information taken from the National Geophysical Data

Centre (NGDC) service (Amante & Eakins 2009) was

therefore used to define and incorporate environmental

heterogeneity into a total of 21 679 demes, each of

which represents an area of approximately 325 km2,

across mainland South-East Asia. Each deme was classi-

fied into one of 25 equally sized categories. Carrying

capacity and friction values were then assigned to each

deme, according to the altitudinal category within that

deme. The carrying capacity and friction parameters

represent, respectively, the size of the population that

any given deme can support and the ease with which

individuals can travel through that deme. The carrying

capacity of intermediate elevations was kept constant

throughout glacial and interglacial periods, at an inter-

mediate carrying capacity of 1000, to reflect the persis-

tence of forest habitat in these regions throughout the

Pleistocene (Heaney 1991; Gathorne-Hardy et al. 2002).

High-elevation regions are likely to have become too

cold and low-elevation regions too arid for the survival

of forest habitat (Gathorne-Hardy et al. 2002), so these

regions were allocated low carrying capacity and high

friction values during the glacial periods (Fig. 2a,b).

The low-altitude regions were allocated zero carrying

capacity values during the glacial period to reflect the

replacement of forest with savannah and grassland

(Heaney 1991; van der Kaars 1991; Kealhofer & Penny

1998; Penny 2001; Hope et al. 2004; White et al. 2004),

and high carrying capacity values (2000) during the

interglacial period, to reflect the expansion of forest

habitat as the climate became more humid. Two alter-

native glacial carrying capacity schemes were modelled:

a narrow carrying capacity scheme in which only a lim-

ited range of altitudes could support mosquito popula-

tions and a wide carrying capacity scheme in which

mosquito populations could survive over a wider range

of altitudes (Fig. 2a,b). In total, 17 evolutionary scenar-

ios were modelled, as detailed in Table 1.

To compare the fit of the real data from each species

to the modelled evolutionary scenarios, the demo-

graphic information generated by SPLATCHE during

the expansion modelling phase was used to simulate,

backwards in time, the genealogy of a set of 340 indi-

viduals sampled from 34 geographical locations across

South-East Asia (10 individuals sampled per location)

until the coalescence of their Most Recent Common

Ancestor (MRCA). This was repeated to generate 1000

simulated data sets for each range expansion model.

Evolutionary model parameters were specified as esti-

mated in MODELTEST. The mutation rate was calcu-

lated from the 2.3% per million year (pmy) divergence

rate (Brower 1994) and adjusted to reflect the modelling

of the range expansion in steps of 200 generations (i.e.

data were simulated 200 generations at a time). Thus, a

modified substitution rate, u, of 1.4697 · 10)4 ⁄ locus ⁄generation was used to generate simulated sequences of

639 bp in length. The fit of the real to the simulated

data was determined following the method of Ray et al.

(2005), for the seven species with widespread distribu-

tions and sampling. The 1000 genealogies simulated

under each range expansion scenario were compared

with each real intraspecific data set using FST matrices,

all of which were calculated in Arlequin 3.11. Pairwise

correlations of the real data FST matrix with each of the

1000 simulated data FST matrices were performed to

produce a distribution of Pearson correlation coeffi-

cients for each scenario. The mean correlation coeffi-

cient was used as an estimate of the goodness of fit of

real to simulated data sets. The correlation coefficients

were calculated using the statistical package R (Ihaka &

Gentleman 1996).

Patterns of diversity

Pleistocene glacial refugial regions are generally

expected to have accumulated greater genetic diversity

relative to those regions recolonized after the LGM,

because of the larger long-term effective population

sizes within refugia (Hewitt 1999, 2001). To further test

the hypothesis of long-term population stability of

Anopheles mosquitoes in northern forest refugia, and the


Carrying capacityscale

0

100

500

1000

1500

2000

(a) (b)

(c)

Fig. 2 Representation of the carrying capacity values for the individual demes within South-East Asia. a and b show the glacial car-

rying capacities modelled under the narrow altitudinal range and wide altitudinal range scenarios, respectively; c shows the intergla-

cial carrying capacities.


inferred recent colonization of the south, we tested the

expectation of a correlation between latitude and

genetic diversity. Anopheles jeyporiensis was omitted

from this analysis as this species is distributed solely at

high latitudes, and An. splendidus was omitted because

of insufficient sample sizes. The diversity statistics hp

and hs were calculated for each population, after remov-

ing those populations with fewer than five individuals.

Correlations between diversity and latitude were car-

ried out independently for each species; however,

because the numbers of populations per species were

typically low, we also maximized the power to test the

hypothesis by combining the populations of all species

into a single analysis. To enable this, the diversity sta-

tistics were normalized by allocating a value of one to

the population within each species with the highest


diversity statistic and calculating the relative diversity

statistics of all other populations within that species.

The significance of correlations between diversity statis-

tics and latitude was determined using SPSS.

Hierarchical Bayesian analysis of simultaneousdivergence

An expectation of the Pleistocene refuge hypothesis is

that vicariance between geographical lineages will be

simultaneous across multiple co-distributed taxa, as that

vicariance has been caused by a common environmen-

tal event (Endler 1982; Moritz et al. 2009). For those

species with geographically congruent divergent lin-

eages (An. annularis, An. philippinensis, An. splendidus

and An. minimus; see Fig. 3a–d), the hypothesis of

Table 1 Scenarios of Pleistocene demographic history, as modelled in SPLATCHE

Scenario No. of origins Location of origins

Pre-expansion

migration

Range of glacial

period altitudes

with high carrying

capacity and low friction

Null model

1 10 Throughout South-East Asia High NA

One northern refugium

2 1 East (northern Vietnam) NA Narrow

3 1 East (northern Vietnam) NA Wide

4 1 West (northeastern India) NA Narrow

5 1 West (northeastern India) NA Wide

One southern refugium

6 1 South (southern Thailand) NA NA

Two northern refugia

7 2 East (northern Vietnam) and West (northeastern India) Low Narrow

8 2 East (northern Vietnam) and West (northeastern India) Medium Narrow

9 2 East (northern Vietnam) and West (northeastern India) High Narrow

10 2 East (northern Vietnam) and West (northeastern India) Low Wide

11 2 East (northern Vietnam) and West (northeastern India) Medium Wide

12 2 East (northern Vietnam) and West (northeastern India) High Wide

13 2 East (Yunnan) and West (northeastern India) Medium Narrow

14 2 East (Yunnan) and West (northeastern India) Medium Wide

15 2 East (northern Vietnam) and West (northern Myanmar) Medium Narrow

16 2 East (northern Vietnam) and West (northern Myanmar) Medium Wide

Two southern refugia

17 2 Southeast (southern Vietnam) and southwest (southern

Myanmar)

Medium NA


simultaneous vicariance was tested using MsBayes. The

hierarchical Bayesian model implemented in MsBayes

uses approximate Bayesian computation (ABC) to test

the hypothesis of simultaneous divergence across multi-

ple taxon pairs. Several hyper-parameters relating to

the divergence of the four taxon pairs, i.e. the intraspe-

cific lineages within An. annularis, An. aconitus, An. phi-

lippinensis and An. splendidus, were estimated using the

hABC framework. These hyper-parameters include the

number of distinct divergence times across all taxon

pairs (W) (if this is estimated to be 1, then simultaneous

divergence is inferred), the mean divergence time

across all taxon pairs (E(s)) and the ratio of the variance

in s to the mean of s (X = var(s) ⁄ E(s)) (Hickerson et al.

2006, 2007). See Hickerson et al. (2007) for full details of

the implemented model. From the K = 1 000 000 simu-

lated draws from the hyper-prior, 2000 were accepted

according to the acceptance ⁄ rejection with local regres-

sion algorithm (Beaumont et al. 2002) and used to con-

struct the posterior probability distribution of the

hyper-parameters. Five runs were carried out using a

range of priors and convergence between runs con-

firmed.

To convert the estimated divergence time into units

of years, a generation time of 0.1 years was assumed

and a substitution rate of 1.15 · 10)8 ⁄ year ⁄ bp, as

derived from the 2.3% divergence rate estimated for

arthropod mitochondrial DNA (Brower 1994). As an

alternative divergence rate estimation, we also used the

slower rate of 1.5% pmy that was estimated for the

mitochondrial CO1 gene using a broad range of arthro-

pod taxa (Quek et al. 2004).

Results

Haplotype networks and patterns of diversity

Median joining haplotype networks indicated the pres-

ence of two deeply divergent clades within the genealo-

gies of An. annularis (Fig. 3a), An. philippinensis

(Fig. 3b), An. splendidus (Fig. 3c) and An. minimus

(Fig. 3d). In all cases, an association between genetic

structure and geography is apparent: haplotypes

derived from the west of South-East Asia (Myanmar,

northeastern India and, where sampled, Sri Lanka) fall

primarily in clade 1, whereas haplotypes derived from

the east (eastern Thailand, Cambodia, Laos and Viet-

nam) fall primarily in the divergent clade 2. In both

An. annularis (Fig. 3a) and An. minimus (Fig. 3d), the

greatest degree of admixture between eastern and

western lineages is seen in central South-East

Asia (western Thailand and eastern Myanmar). The


An. philippinensis

An. minimus

An. splendidus

(a) An. annularis

An. aconitus

COLOUR KEY

An. sawadwongporni

An. maculatus

An. baimaii

(f)

(b)

(d)

(h)

(g)

(i)

(e)

(c)

An. jeyporiensis

Fig. 3 Median joining haplotype networks for each of the nine species included within this study. The size of the circles is propor-

tional to the number of individuals with that haplotype; circles are coloured according to the locality from which individuals were

sampled. The length of the branches separating haplotypes is proportional to the number of mutational steps between them. Haplo-

type networks were constructed using NETWORK 4.51.


haplotype networks of An. aconitus (Fig. 3e), An. jeypo-

riensis (Fig. 3f), An. sawadwongporni (Fig. 3g) and

An. maculatus (Fig. 3i) do not show deeply divergent

eastern and western lineages. Substantial genetic diver-

sity can be seen in the networks of both An. aconitus

and An. jeyporiensis with many unique and divergent

haplotypes. By contrast, haplotypes of the closely

related An. maculatus and An. sawadwongporni from

across most of mainland South-East Asia form predomi-

nantly star-like genealogies indicative of recent popula-

tion expansion. The network of An. baimaii, which has

a distribution limited to the west of South-East Asia,

also displays a star-like appearance (Fig. 3h).

The partial Mantel tests comparing the matrices of

population pairwise FST values between species pairs,

controlling for the effect of isolation by distance,

revealed no statistically significant correlations. How-

ever, the following partial correlations showed a posi-

tive trend that approached significance: An. minimus

and An. annularis (r = 0.602, P = 0.064), An. minimus


and An. philippinensis (r = 0.623, P = 0.094) and

An. minimus and An. aconitus (r = 0.374, P = 0.059),

suggesting some similarity between these species in the

distribution of genetic diversity across the landscape.

Hence, both the haplotype networks and FST matrices

revealed that whilst there are some similarities between

certain species, there are also considerable incongruenc-

es in the overall phylogeographical patterns.

Population expansion

Similarly timed population expansions were detected in

An. aconitus, An. jeyporiensis, An. baimaii, An. maculatus

and the eastern lineages of An. annularis, An. minimus

and An. philippinensis, starting in the mid- to late Pleis-

tocene, between 300 and 100 kya (see Fig. S1, Support-

ing information). Within the western lineages of

An. annularis and An. minimus, the Bayesian skyline

analysis supported a stable population size rather than

a recent expansion (Fig. S1d,f, Supporting information).


Although the BSP of An. sawadwongporni suggests a

lack of recent population growth (Fig. S1k, Supporting

information), this is misleading as it results from the

almost complete lack of genetic variation in this data

set, which is itself indicative of very recent expansion

(either demographic or selective sweep).

Spatially explicit modelling of demographic history

The FST matrices of An. annularis and An. baimaii best

fitted the demographic scenarios 15 and 8, respectively,

which both include two refugia in the northeast and

northwest of South-East Asia, limited adaptability to

glacial environments, and medium levels of pre-

expansion migration between refugia (Tables 1 and 2,

Fig. S2a,b, Supporting information). Because An. bai-

maii is restricted to western South-East Asia, the genetic

data were compared with simulated data from only the

corresponding western populations; hence, the best sup-

ported scenario for this species effectively represents a

single northwestern refugium. Anopheles minimus and

An. philippinensis best fitted scenario 7, which was simi-

lar to scenario 8 but allowed only low levels of pre-

expansion migration between refugia (Tables 1 and 2,

Fig. S2b,e, Supporting information). Scenario 11, which

was similar to scenario 8 but included adaptability to a

wider range of glacial environments, showed the best

fit to the FST matrices of An. aconitus and An. saw-

adwongporni (Tables 1 and 2, Fig. S2c,g, Supporting

information). The best fitting scenario for all species

was always amongst those with two allopatric northern

refugia (scenarios 7–16), rather than the following alter-

native scenarios: the null model, which included multi-

ple origins and high levels of pre-expansion migration

between them (i.e. no refugia); scenarios including a

single northern refugium; or scenarios including single

or multiple southern refugia (Tables 1 and 2). The spe-

cific location of eastern and western refugia was varied

between northern Vietnam and Yunnan, and between

northern Myanmar and northeastern India, respectively.

There was little difference between the fit of the real

data in each species to those simulated using various

combinations of these putative northeastern and north-

western refugia (Tables 1 and 2).

Patterns of diversity

No significant correlations were detected between

genetic diversity and latitude when each species was

considered independently. This is probably because of

low power associated with small sample sizes because

the combined analysis with the normalized diversity

statistics from all species detected significant and posi-

tive correlations between latitude and hp (r = 0.347,

P = 0.002, n = 77; see Fig. S3, Supporting information)

and between latitude and hs (r = 0.343, P = 0.002,

n = 77). The hypothesis of greater genetic diversity in

the north relative to the south was therefore supported.

Simultaneous divergence

A simultaneous divergence time for the lineages within

all four species was strongly supported using a range

of priors; the estimated number of divergence times (W)

was close to one (1.064, 95% CI: 1.00–1.898), and the

estimated variance in divergence times (X) was close to

zero (0.0095, 95% CI: 0.00–0.119) (Fig. 4). The mean sestimated for all taxon pairs (E(s)) is 0.199. Assuming

the 2.3% pmy mitochondrial divergence rate estimated

for a range of arthropod taxa, including other Diptera

(Brower 1994), this translates to a divergence time of

880 075 ya (95% CI: 460 353–1 338 907 ya). If we

assume the slower rate of 1.5% divergence pmy, this

translates to an older divergence time of 1 349 448 ya

(95% CI: 705 875–2 052 990 ya).

Discussion

The data presented support a common role for climatic

fluctuation, and the consequent fragmentation of forest

habitat within South-East Asia (van der Kaars 1991;

Kealhofer & Penny 1998; Penny 2001; Hope et al. 2004;

White et al. 2004), in driving diversification and shap-

ing the distribution of genetic diversity in nine species

of Anopheles mosquito across the region. The best fitting

demographic models for all six widespread species

tested were those that included two allopatric refugia

within the mountainous regions of northeast and north-

west South-East Asia. With the exception of An. saw-

adwongporni, which may have increased in population

size very recently, all species with sufficient sample

sizes exhibited late Pleistocene (100–300 kya) demo-

graphic expansions, which are likely to have been trig-

gered by the spread of forest habitat from refugia

during warm, moist interglacial periods (Heaney 1991).

Allopatric isolation within these refugia has presumably

led to the simultaneous divergence of the eastern and

western lineages of the CO2 gene observed in An. annu-

laris, An. minimus, An. splendidus and An. philippinensis.

This pattern of east–west divergence is also shared by

An. baimaii and An. dirus, which have western and

eastern distributions, respectively, although divergence

in this case has resulted in speciation (Morgan et al.

2010). In contrast, no deep divergences were observed

in An. aconitus, An. sawadwongporni or An. maculatus.

Despite such different outcomes in patterns of genetic

structure, the demographic modelling process was

nonetheless able to detect a common influence of envi-


Tab

le2

Mea

nco

rrel

atio

nco

effi

cien

tfo

rth

ere

alF

ST

mat

rix

of

each

spec

ies

toea

cho

fth

e10

00m

atri

ces

sim

ula

ted

un

der

each

SP

LA

TC

HE

scen

ario

Sce

nar

io

Sp

ecie

sm

ean

corr

elat

ion

coef

fici

ent

An

.ac

onit

us

An

.an

nu

lari

sA

n.

baim

aii

An

.m

acu

latu

sA

n.

min

imu

sA

n.

phil

ippi

nen

sis

An

.sa

wad

won

gpor

ni

Nu

llm

od

el

10.

033

(0.0

20to

0.04

5)0.

033

(0.0

21to

0.04

5)0.

026

(0.0

13to

0.03

8))

0.01

6()

0.02

4to

)0.

006)

)0.

030

()0.

041

to)

0.01

7)0.

095

(0.0

71to

0.11

8)0.

013

()0.

006

to0.

032)

On

en

ort

her

nre

fug

ium

20.

127

(0.1

16to

0.13

9)

0.05

4

(0.0

43to

0.06

4)

0.25

2

(0.2

32to

0.27

2)

)0.

038

()0.

049

to)

0.02

6)

0.03

1

(0.0

19to

0.04

3)

0.06

5

(0.0

45to

0.08

5)

0.05

5

(0.0

32to

0.07

7)

30.

130

(0.1

16to

0.14

2)

0.19

2

(0.1

80to

0.20

5)

0.15

9

(0.1

44to

0.17

5)

0.03

0

(0.0

18to

0.04

3)

0.07

2

(0.0

60to

0.08

5)

0.07

2

(0.0

55to

0.08

9)

0.06

8

(0.0

48to

0.08

8)

40.

141

(0.1

29to

0.15

4)

0.06

4

(0.0

53to

0.07

6)

0.11

3

(0.0

98to

0.12

7)

)0.

050

()0.

060

to)

0.03

9)

0.00

6

()0.

005

to0.

017)

0.00

3

()0.

007

to0.

019)

)0.

162

()0.

180

to)

0.14

6)

50.

143

(0.1

29to

0.15

6)

0.05

7

(0.0

45to

0.06

8)

0.15

7

(0.1

42to

0.17

3)

0.02

3

(0.0

13to

0.03

5)

0.05

8

(0.0

46to

0.07

0)

0.14

3

(0.1

33to

0.15

4)

0.09

7

(0.0

76to

0.12

6)

On

eso

uth

ern

refu

giu

m

60.

057

(0.0

43to

0.07

0)

0.09

8

(0.0

85to

0.11

0)

0.05

3

(0.0

40to

0.06

7)

0.00

4

()0.

007

to0.

015)

)0.

023

()0.

037

to)

0.00

9)

0.15

1

(0.1

25to

0.17

7)

0.02

5

(0.0

06to

0.04

5)

Tw

on

ort

her

nre

fug

ia

7)

0.08

3

()0.

067

to0.

110)

0.49

5

(0.4

80to

0.51

1)

0.19

7

(0.1

87to

0.20

7)

0.18

3

(0.1

80to

0.18

7)

0.43

6

(0.4

25to

0.44

7)

0.72

6

(0.7

11to

0.74

1)

0.17

8

(0.1

64to

0.19

1)

8)

0.02

6

()0.

035

to)

0.01

8)

0.50

6

(0.4

97to

0.51

4)

0.36

4

(0.3

55to

0.37

2)

0.26

6

(0.2

57to

0.27

4)

0.27

3

(0.2

65to

0.28

1)

0.63

1

(0.6

16to

0.64

6)

0.11

5

(0.1

06to

0.12

4)

90.

008

()0.

002

to0.

018)

0.46

6

(0.4

56to

0.47

6)

0.33

1

(0.3

20to

0.34

2)

0.23

3

(0.2

23to

0.24

2)

0.23

4

(0.2

24to

0.24

4)

0.58

0

(0.5

63to

0.59

7)

0.11

0

(0.0

99to

0.12

1)

10)

0.01

7

()0.

026

to)

0.00

7)

0.46

1

(0.4

53to

0.46

9)

0.24

4

(0.2

35to

0.25

2)

0.16

2

(0.1

53to

0.17

1)

0.37

1

(0.3

60to

0.38

2)

0.66

8

(0.6

57to

0.68

1)

0.11

3

(0.1

05to

0.12

2)

110.

301

(0.2

88to

0.31

4)

0.07

1

(0.0

60to

0.08

2)

0.08

8

(0.0

78to

0.09

9)

)0.

089

( )0.

098

to)

0.08

0)

0.05

4

(0.0

44to

0.06

4)

)0.

009

()0.

035

to0.

017)

0.26

9

(0.2

49to

0.29

0)

120.

272

(0.2

59to

0.28

5)

0.12

0

(0.1

07to

0.13

3)

0.11

4

(0.1

03to

0.12

5)

)0.

057

()0.

068

to)

0.04

6)

0.09

7

(0.0

85to

0.10

8)

0.10

7

(0.0

79to

0.13

5)

0.25

7

(0.2

37to

0.27

6)

130.

078

(0.0

67to

0.09

0)

0.40

4

(0.3

96to

0.41

2)

0.33

8

(0.3

28to

0.34

8)

0.18

2

(0.1

71to

0.19

3)

0.25

2

(0.2

42to

0.26

2)

0.63

6

(0.6

20to

0.65

1)

0.09

0

(0.0

78to

0.10

0)

140.

127

(0.1

14to

0.14

1)

0.41

8

(0.4

09to

0.42

7)

0.18

1

(0.1

70to

0.19

2)

0.05

4

(0.0

43to

0.06

5)

0.28

1

(0.2

68to

0.29

4)

0.44

1

(0.4

16to

0.46

6)

0.20

6

(0.1

91to

0.22

1)

15)

0.02

4

()0.

032

to)

0.01

5)

0.51

0

(0.5

02to

0.51

9)

0.34

0

(0.3

32to

0.34

8)

0.22

7

(0.2

19to

0.23

6)

0.24

3

(0.2

35to

0.25

0)

0.42

3

(0.4

13to

0.44

3)

0.11

6

(0.1

07to

0.12

5)

160.

117

(0.1

03to

0.13

0)

0.42

9

(0.4

21to

0.43

8)

0.18

0

(0.1

70to

0.19

0)

0.06

0

(0.0

49to

0.07

1)

0.28

6

(0.2

73to

0.29

9)

0.45

8

(0.4

33to

0.48

3)

0.17

5

(0.1

60to

0.19

0)

Tw

oso

uth

ern

refu

gia

170.

252

(0.2

39to

0.26

4)

0.32

2

(0.3

10to

0.33

5)

0.07

2

(0.0

60to

0.08

4)

)0.

081

()0.

090

to)

0.07

3)

0.07

4

(0.0

64to

0.08

4)

)0.

028

()0.

049

to)

0.00

6)

0.02

3

(0.0

02to

0.04

4)

95%

con

fid

ence

inte

rval

s[c

alcu

late

das

mea

n±

(sta

nd

ard

erro

ro

fth

em

ean

·1.

96)]

are

sho

wn

inb

rack

ets.



Den

sity

01

23

41.0 1.5 2.0 2.5 3.0 3.5 4.0

Psi (= number of possible divergence times)

50

100

150

200

Pr(Om

ega, E(t) I X)

00.0

0.1

0.2

0.3

0.4

0.5

E(t)

1.2

1.0

0.8

0.6

0.40.2

0.0

Omega

(a)

(b)

Fig. 4 The posterior distributions of parameters estimated using MsBayes [35]. a shows the posterior probability distribution of W,

the number of divergence times across multiple taxon pairs; b shows the three-dimensional joint-probability distribution of E(s), the

mean divergence time across multiple taxon pairs, and X, which represents the variance in estimated divergence times across taxon

pairs.


ronmental history on the distribution of genetic diver-

sity across all species.

Differences between species in the patterns of genetic

diversity, which are present despite a common influ-

ence of Pleistocene environmental change, can be

explained by a combination of stochastic effects and dif-

ferences in species biology. The SPLATCHE modelling

approach can provide some insight into the latter. The

fit of real to simulated data sets was influenced by lev-

els of pre-expansion migration between refugial regions,

representing levels of gene flow as populations

expanded from allopatric refugia during the intergla-

cials. Whereas An. minimus and An. philippinensis best

fitted a model of low pre-expansion migration, all

remaining species best fitted a model with higher levels

of gene flow, potentially representing differences in dis-

persal capacity.

The model fitting was also influenced by the breadth

of altitudinal range of high carrying capacity and low

friction during glacial periods, corresponding to a spe-

cies’ tolerance to glacial habitats. Anopheles minimus and

An. baimaii best fit models that have a low altitudinal

range, consistent with our understanding that these spe-

cies are highly forest dependent (Reid 1968; Darsie &

Pradhan 1990; Rattanarithikul et al. 2006). Anopheles

aconitus and An. sawadwongporni are the only species

that best fit evolutionary scenarios with a wider altitu-

dinal range, indicating greater flexibility to glacial habi-

tats. Anopheles aconitus is more of an ecological

generalist, utilizing exposed larval habitats such as rice

paddies and ditches (Overgaard et al. 2003). During gla-

cial periods, An. aconitus may have been able to survive

in a relatively wide range of habitats, providing they

retained sufficient moisture to provide appropriate



larval habitat. A similar argument may apply to

An. sawadongporni as its larval habitats, shaded streams

(Rattanarithikul et al. 2006), could also potentially occur

across a wider altitudinal range than forest habitats.

However, it would then be surprising that An. macula-

tus, which is apparently ecologically similar and closely

related to An. sawadongporni (Rattanarithikul et al. 2006;

Morgan et al. 2009), best fitted models with narrow alti-

tudinal ranges of suitable glacial habitat. This apparent

difference is also striking given that over the same area

of mainland South-East Asia (i.e. excluding eastern

China, Taiwan and Sri Lanka), the gene trees of both

species indicate recent range expansions. This suggests

that the differences in the altitudinal ranges of suitable

glacial habitat for these two species may be due to limi-

tations in the SPLATCHE modelling approach, for

example, in the limited number of scenarios modelled

and ⁄ or the greater geographical sampling coverage for

An. maculatus. If An. aconitus, An. sawadongporni and

perhaps also An. maculatus are more plastic in their

habitat requirements, this may underlie their lack of

formation of deeply divergent lineages, presumably

owing to lower levels of isolation during glacial periods

compared to the other more forest-associated species in

this study.

The spatially explicit modelling approach provided

strong evidence for the persistence of mosquito popula-

tions in two allopatric northern refugia, in the east and

the west of South-East Asia, during the Pleistocene gla-

cial periods. The approach had insufficient power, how-

ever, to distinguish between the potential eastern

regions of northern Vietnam and Yunnan, and between

the potential western regions of northern Myanmar and

northeastern India. Northern Vietnam, northeastern

India and northern Myanmar were all identified as

putative forest refugial regions, based on a phylogenetic

study of the Neocellia Series of Anopheles mosquitoes

(Morgan et al. 2009). Northern Vietnam has also been

proposed as a refugial region for the following forest-

associated taxa: Anopheles baimaii (Morgan et al. 2010),

the bird species Sasia ochracea (Fuchs et al. 2008) and

Colobinae monkeys (Brandon-Jones 1996). The substan-

tial population structure within Anopheles jeyporiensis,

which is limited to the north of South-East Asia, further

supports the persistence of suitable forest habitat in the

north throughout the Pleistocene. Increased, high-den-

sity sampling throughout the mountainous northern

regions may enable the locations of Pleistocene forest

refugia to be more precisely pinpointed, the relative

importance of those refugia defined and any additional

complexity, such as microrefugia (Rull 2010), identified.

Because of the long-term persistence of populations

within glacial refugia, such regions are expected to

accumulate greater diversity than recently colonized,


less stable regions (Hewitt 1999, 2001). For example,

within continental Europe, the long-term survival of

many species in southern refugia has led to greater

genetic diversity in the south relative to the more

recently colonized north (Hewitt 2001, 2004). In the for-

est-dependent South-East Asian taxa studied here, a

similar pattern is evident, but genetic diversity

increases with increasing latitude. North–south clines of

decreasing diversity along similar regions of the Thai–

Myanmar border have also been noted at the intraspe-

cific level in the black fly, Simulium tani (Pramual et al.

2005) and at the species level in sphingid moths (Beck

et al. 2007). This commonality of latitudinal diversity

gradient together with the demographic modelling sup-

ports a scenario of long-term survival in mountainous

northern refugia followed by more recent colonization

of the south for a range of taxa. All of the putative refu-

gial areas, northeast India, northern Myanmar, Yunnan

and northern Vietnam, are not only rich in genetic

diversity as observed here but are also noted regions of

high species biodiversity and endemicity (Pramual et al.

2005; Sterling & Hurley 2005; Ying-shan et al. 2007).

The results of this study support the importance of

montane regions as sources and sinks of tropical biodi-

versity, as has previously been documented in Borneo,

Malaya and Sumatra (in Crematogaster ants; Quek et al.

2007), in Africa (Fjeldsa & Lovett 1997; Roy 1997; Bowie

et al. 2006) and in South America (Fjeldsa & Rahbek

2006).

The Thai–Myanmar border is a suture zone where

the deeply divergent eastern and western lineages of

An. minimus and An. annularis overlap, and also a

potential hybrid zone where the sister taxa An. dirus

and An. baimaii meet (Morgan et al. 2010). Increased

sampling within the region is necessary to determine

whether An. splendidus and An. philippinensis also have

suture zones along the Thai–Myanmar border. Simulta-

neous divergence of eastern and western lineages

within An. annularis, An. minimus, An. philippinensis

and An. splendidus is strongly supported by hABC anal-

ysis and estimated to have occurred within the mid-

Pleistocene, �900 or �1300 kya, depending on the

divergence rate used. The �900 kya estimate is likely to

be more accurate because this used the divergence rate

of 2.3% pmy (Brower 1994), which was estimated from

groups taxonomically similar to Anopheles (e.g. several

Diptera) and the same gene, CO2. Conversely, the 1.5%

pmy divergence rate estimated by Quek et al. (2004)

used a very broad range of arthropods (including crabs)

and a different gene, CO1. Allopatric divergence in the

speciation of An. dirus and An. baimaii also dates to a

similar time period, �1 Ma using the faster Brower

(1994) rate of divergence (Morgan et al. 2010). The

simultaneous and apparently coincident formation of


suture ⁄ hybrid zones across several species suggests that

this divergence was indeed driven by a common influ-

ence of environmental change. The estimated date

(using either molecular clock) is too recent to have been

triggered by tectonic activity or orogony, which dates

back to 5 Ma or earlier in this mainland region (Hall

2002). A more likely explanation for the initial driving

of divergence is the change in Pleistocene glacial cycles

approximately 1 Ma, such that the cycles occurred more

rapidly and with increased severity (Imbrie et al. 1992).

The increased environmental pressures resulting from

this change may have intensified the fluctuations in for-

est cover, triggering diversification within forest-associ-

ated taxa.

The positioning of the suture zone may result from

the southwards projection of the northern mountainous

region into the narrow mountainous formation along

the Thai–Myanmar border (see Fig. 1). At the transition

into the post-glacial period, the regions close to the

mountains are likely to have been reforested prior to

the more distant and initially more arid areas of central

Myanmar and Thailand. Consequently, the mountain-

ous projection along the Thai–Myanmar border could

facilitate early southwards dispersal from eastern and

western refugia along its eastern and western flanks to

form the suture zone.

If the common patterns in the distribution of genetic

diversity detected in this study are also found in other

forest-associated taxa, the refugial regions of the moun-

tainous north of South-East Asia and the suture zone

on the Thai–Myanmar border would have particularly

high conservation value. These regions not only cur-

rently possess high and unique genetic diversity but

would also be important for the generation and mainte-

nance of biodiversity in the long term. Thus, northern

Vietnam, Yunnan, northern Myanmar and northeastern

India, as well as the border region of Thailand and

Myanmar, are highlighted as important regions for fur-

ther study. In this respect, Anopheles mosquitoes will be

useful in more fine-scale studies of these regions to

characterize in greater detail the effects of Pleistocene

environmental change. Above all, it is imperative that

comparable studies of other forest-dependent taxa,

including vertebrates, are conducted to determine the

taxonomic generality of the findings of this study and

ensure that this understanding can be taken into

account in planning conservation policy.

Acknowledgements

Thank you to all staff at the following institutions, for their

invaluable help in mosquito collection: Gauhati University in

northeastern India; the Office of Disease Prevention and Control

in Chiang Mai, Thailand; the Department for Medical Research,

Myanmar; the Centre of Malariology, Parasitology and Ento-

mology, Laos; the National Institute of Malariology, Parasitol-

ogy and Entomology, Vietnam; and the Center of Malariology,

Parasitology and Entomology in Phnom Penh, Cambodia. We

also thank David Robertson, University of Manchester, and

Michael Bruford, Cardiff University, for helpful discussion and

NERC, NSFC (Grant No. 31071968) and the Edinburgh NERC

Biomolecular Analysis Facility for funding and support.

References

Achard F, Eva HD, Stibig H-J et al. (2002) Determination of

deforestation rates of the world’s humid tropical forests.

Science, 297, 999–1002.

Amante C, Eakins BW (2009) ETOPO1 1 Arc-Minute Global

Relief Model: Procedures, Data Sources and Analysis. NOAA

Technical Memorandum NESDIS NGDC-24, 19 pp.

Bandelt H-J, Forster P, Rowhl A (1999) Median-joining

networks for inferring intraspecific phylogenies. Molecular

Biology and Evolution, 16, 37–48.

Beaumont MA, Zhang W, Balding DJ (2002) Approximate

Bayesian Computation in Population Genetics. Genetics, 162,

2025–2035.

Beck J, Kitching IJ, Haxaire J (2007) The latitudinal distribution

of Sphingid species richness in continental Southeast Asia:

what causes the biodiversity ‘hotspot’ in northern Thailand?

The Raffles Bulletin of Zoology, 55, 179–185.

Bermingham E, Moritz C (1998) Comparative phylogeography:

concepts and applications. Molecular Ecology, 7, 367–369.

Bowie RCK, Fjeldsa J, Hackett SJ, Bates JM, Crowe TM (2006)

Coalescent models reveal the relative roles of ancestral

polymorphism, vicariance, and dispersal in shaping

phylogeographical structure of an African montane forest

robin. Molecular Phylogenetics and Evolution, 38, 171–188.

Brandon-Jones D (1996) The Asian Colobinae (Mammalia:

Cercopithecidae) as indicators of Quaternary climatic

change. Biological Journal of the Linnean Society, 59, 327–350.

Brower A (1994) Rapid morphological radiation and convergence

among races of the butterfly Heliconius erato inferred from

patterns of mitochondrial DNA evolution. Proceedings of the

National Academy of Sciences, USA, 91, 6491–6495.

Carnaval AC, Hickerson MJ, Haddad CFB, Rodrigues MT,

Moritz C (2009) Stability predicts genetic diversity in the

Brazilian Atlantic forest hotspot. Science, 323, 785–789.

Chen B, Harbach RE, Butlin RK (2004) Genetic variation and

population structure of the mosquito Anopheles jeyporiensis in

southern China. Molecular Ecology, 13, 3051–3056.

Chen B, Pedro PM, Harbach RE, Somboon P, Walton C, Butlin

RK (2011) Mitochondrial DNA variation in the malaria

vector Anopheles minimus across China, Thailand and

Vietnam: evolutionary hypothesis, population structure and

population history. Heredity, 106, 241–252.

Clark PU, Dyke AS, Shakun JD et al. (2009) The last glacial

maximum. Nature, 325, 710–714.

Colinvaux PA, De Oliveira PE, Bush MB (2000) Amazonian

and neotropical plant communities on glacial time-scales: the

failure of the aridity and refuge hypotheses. Quaternary

Science Reviews, 19, 141–169.

Colinvaux PA, Irion G, Raesaenen ME, Bush MB (2001)

Geological and paleoecological data falsify the Haffer &



Prance refuge hypothesis of Amazonian speciation.

Amazoniana, 16, 609–646.

Covell G (1927) The Distribution of Anopheline Mosquitoes in

India and Ceylon. Indian Journal of Medical Research Memoirs,

5, 1–85.

Crisci JV, Katinas L, Posadas P (2003) Historical Biogeography:

An Introduction. Harvard University Press, Cambridge MA.

Currat M, Ray N, Excoffier L (2004) SPLATCHE: a program to

simulate genetic diversity taking into account environmental

heterogeneity. Molecular Ecology Notes, 4, 139–142.

Darsie RF, Pradhan SP (1990) The mosquitoes of Nepal: their

identification, distribution and biology. Mosquito Systematics,

22, 69–130.

Drummond AJ, Rambaut A (2007) BEAST: Bayesian

evolutionary analysis by sampling trees. BMC Evolutionary

Biology, 7, 214–221.

Endler JA (1982) Problems in distinguishing historical from

ecological factors in biogeography. American Zoologist, 22,

441–452.

Esselstyn JA, Timm RM, Brown RM (2009) Do geological or

climatic processes drive speciation in dynamic archipelagos?

The tempo and mode of diversification in Southeast Asia

shrews. Evolution, 63, 2595–2610.

Excoffier LGL, Schneider S (2005) Arlequin ver. 3.0: an

integrated software package for population genetics data

analysis. Evolutionary Bioinformatics Online, 1, 47–50.

Feldman CR, Spicer GS (2006) Comparative phylogeography of

woodland reptiles in California: repeated patterns of

cladogenesis and population expansion. Molecular Ecology,

15, 2201–2222.

Fjeldsa J (1994) Geographical patterns for relict and young

species of birds in Africa and South America and

implications for conservation priorities. Biodiversity and

Conservation, 3, 207–226.

Fjeldsa J, Lovett JC (1997) Geographical patterns of old and

young species in African forest biota: the significance of

specific montane areas as evolutionary centres. Biodiversity

and Conservation, 6, 325–346.

Fjeldsa J, Rahbek C (2006) Diversification of tanagers, a species

rich bird group, from lowlands to montane regions of South

America. Integrative and Comparative Biology, 46, 72–81.

Fuchs J, Ericson PGP, Pasquet E (2008) Mitochondrial

phylogeographic structure of the white-browed piculet (Sasia

ochracea): cryptic genetic differentiation and endemism in

Indochina. Journal of Biogeography, 35, 565–575.

Garros C, Harbach RE, Manguin S (2005) Systematics and

biogeographical implications of the phylogenetic

relationships between members of the Funestus and

Minimus Groups of Anopheles (Diptera: Culicidae). Journal of

Medical Entomology, 42, 7–18.

Gaston KJ (2000) Global Patterns in Biodiversity. Nature, 405,

220–227.

Gathorne-Hardy FJ, Syaukani, Davies RG, Eggleton P, Jones

DT (2002) Quaternary rainforest refugia in south-east Asia:

using termites (Isoptera) as indicators. Biological Journal of the

Linnean Society, 75, 453–466.

Haffer J (1969) Speciation in Amazonian forest birds. Science,

165, 131–137.

Haffer J, Prance GT (2001) Climatic forcing of evolution in

Amazonia during the Cenozoic: on the refuge theory of

biotic differentiation. Amazoniana, 16, 579–607.


Hall R (2002) Cenozoic geological and plate tectonic evolution

of SE Asia and the SW Pacific: computer-based

reconstructions, model and animations. Journal of Asian Earth

Sciences, 20, 353–431.

Harbach RE, Kitching IJ (2005) Reconsideration of anopheline

mosquito phylogeny (Diptera: Culicidae: Anophelinae) based

on morphological data. Systematics and Biodiversity, 3, 345–

374.

Heaney LR (1991) A synopsis of climatic and vegetational

change in Southeast Asia. Climatic Change, 19, 53–61.

Heled J, Drummond A (2008) Bayesian inference of population

size history from multiple loci. BMC Evolutionary Biology, 8,

289.

Hewitt GM (1999) Post-glacial re-colonization of European

biota. Biological Journal of the Linnean Society, 68, 87–112.

Hewitt GM (2001) Speciation, hybrid zones and

phylogeography—or seeing genes in space and time.

Molecular Ecology, 10, 537–549.

Hewitt GM (2004) Genetic consequences of climatic oscillations

in the Quaternary. Proceedings of the Royal Society, B, 359,

183–195.

Hickerson MJ, Dolman G, Moritz C (2006) Comparative

phylogeographic summary statistics for testing simultaneous

vicariance. Molecular Ecology, 15, 209–223.

Hickerson MJ, Stahl E, Takebayashi N (2007) MsBayes:

pipeline for testing comparative phylogeographic histories

using hierarchical approximate Bayesian computation. BMC

Bioinformatics, 8, 268.

Hope G, Kershaw AP, van der Kaars S et al. (2004) History of

vegetation and habitat change in the Austral-Asian region.

Quaternary International, 118–119, 103–126.

Ihaka R, Gentleman R (1996) R: a language for data analysis

and graphics. Journal of Computational and Graphical Statistics,

5, 299–314.

Imbrie J, Boyle EA, Clemens SC, Duffy A, Howard WR et al.

(1992) On the structure and origin of major glaciation cycles

1. Linear responses to Milankovitch forcing. Paleoceanography,

7, 701–738.

Iyengar A, Babu VN, Hedges S, Venkataraman AB, Maclean N,

Morin PA (2005) Phylogeography, genetic structure, and

diversity in the dhole (Cuon alpinus). Molecular Ecology, 14,

2281–2297.

van der Kaars WA (1991) Palynology of eastern Indonesian

marine piston-cores: a Late Quaternary vegetational and

climatic record for Australasia. Palaeogeography,

Palaeoclimatology, Palaeoecology, 85, 239–302.

Kealhofer L, Penny D (1998) A combined pollen and phytolith

record for fourteen thousand years of vegetation change in

northeastern Thailand. Review of Palaeobotany and Palynology,

103, 83–93.

Knowles LL (2004) The burgeoning field of statistical

phylogeography. Journal of Evolutionary Biology, 17, 1–10.

Maharaj R (2003) Life table characteristics of Anopheles

arabiensis (Diptera: Culicidae) under simulated seasonal

conditions. Journal of Medical Entomology, 40, 737–742.

Manguin S, Garros C, Dusfour I, Harbach RE, Coosemans M

(2008) Bionomics, taxonomy, and distribution of the major

malaria vector taxa of Anopheles subgenus Cellia in Southeast

Asia: an updated review. Infection, Genetics and Evolution, 8,

489–503.


Mayle FE, Beerling DJ, Gosling WD, Bush MB (2004)

Responses of Amazonian ecosystems to climatic and

atmospheric carbon dioxide changes since the last glacial

maximum. Philosophical Transactions of the Royal Society B,

359, 499–514.

Morgan K, O’Loughlin SM, Mun-Yik F, Linton Y-M, Somboon

P et al. (2009) Molecular phylogenetics and biogeography of

the Neocellia Series of Anopheles mosquitoes in the Oriental

Region. Molecular Phylogenetics and Evolution, 52, 588–601.

Morgan K, Linton Y-M, Somboon P et al. (2010) Inter-specific

gene flow dynamics during the Pleistocene-dated speciation

of forest-dependent mosquitoes in Southeast Asia. Molecular

Ecology, 19, 2269–2285.

Moritz C (2002) Strategies to protect biological diversity and

the evolutionary processes that sustain it. Systematic Biology,

51, 238–254.

Moritz C, Patton JL, Schneider CJ, Smith TB (2000)

Diversification of rainforest faunas: an integrated molecular

approach. Annual Review of Ecology and Systematics, 31, 533–

563.

Moritz C, Richardson KS, Ferrier S et al. (2001)

Biogeographical concordance and efficiency of taxon

indicators for establishing conservation priority in a tropical

rainforest biota. Proceedings of the Royal Society, B, 268, 1875–

1881.

Moritz C, Hoskin CJ, MacKenzie JB et al. (2009) Identification

and dynamics of a cryptic suture zone in tropical rainforest.

Proceedings of the Royal Society B., 276, 1235–1244.

Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB,

Kent J (2000) Biodiversity hotspots for conservation

priorities. Nature, 403, 853–858.

Nielsen R, Beaumont MA (2009) Statistical inferences in

phylogeography. Molecular Ecology, 18, 1034–1047.

O’Loughlin SM, Okabayashi T, Honda M, Kitazoe Y, Kishino

H et al. (2008) Complex population history of two Anopheles

dirus mosquito species in Southeast Asia suggests the

influence of Pleistocene climate change rather than human-

mediated effects. Journal of Evolutionary Biology, 21, 1555–

1569.

Overgaard HJ, Ekbom B, Suwonkerd W, Takagi M (2003)

Effect of landscape structure on anopheline mosquito

density and diversity in northern Thailand: implications for

malaria transmission and control. Landscape Ecology, 18,

605–619.

Penny D (2001) A 40,000 year palynological record from

north-east Thailand; implications for biogeography and

palaeo-environmental reconstruction. Palaeogeography Palae-

oclimatology Palaeoecology, 171, 97–128.

Posada D, Crandall KA (1998) MODELTEST: testing the model

of DNA substitution. Bioinformatics, 14, 817–818.

Pramual P, Kuvangkadilok C, Baimai V, Walton C (2005)

Phylogeography of the black fly Simulium tani (Diptera:

Simuliidae) from Thailand as inferred from mtDNA

sequences. Molecular Ecology, 14, 3989–4001.

Quek S-P, Davies SJ, Itino T, Pierce NE (2004) Codiversification

in an ant-plant mutualism: stem texture and the evolution of

host use in Crematogaster (FormicidaeL Myrmicinae) in

habitats of Macaranga (Euphorbiaceae. Evolution, 58, 554–

570.

Quek S-P, Davies SJ, Ashton PS, Itino T, Pierce NE (2007) The

geography of diversification in mutualistic ants: a gene’s-eye

view into the Neogene history of Sundaland rain forests.

Molecular Ecology, 16, 2045–2062.

Rattanarithikul R, Harrison BA, Harbach RE, Panthusiri P,

Coleman RE (2006) Illustrated keys to the mosquitoes of

Thailand IV. Anopheles. Southeast Asian Journal of Tropical

Medicine and Public Health, 37(Suppl. 2), 1–128.

Ray N, Currat M, Berthier P, Excoffier L (2005) Recovering the

geographic origin of early modern humans by realistic and

spatially explicit simulations. Genome Research, 15, 1161–1167.

Reddy S (2008) Systematics and biogeography of the shrike

babblers (Pteruthius): species limits, molecular phylogenetics,

and diversification patterns across southern Asia. Molecular

Phylogenetics and Evolution, 47, 54–72.

Reid JA (1968) Anopheline Mosquitoes of Malaya and Borneo.

Government of Malaysia, Malaysia.

Roy MS (1997) Recent diversification in African greenbuls

(Pycnonotidae: Andropadus) supports a montane speciation

model. Proceedings of the National Academy of Sciences, USA,

264, 1337–1344.

Rull V (2010) On microrefugia and cryptic refugia. Journal of

Biogeography, 37, 1623–1625.

Sambrook J (1989) Molecular Cloning: A Laboratory Manual. Cold

Spring Harbor Laboratory, Cold Spring Harbor, NY.

Schneider CJ, Cunningham M, Moritz C (1998) Comparative

phylogeography and the history of endemic vertebrates in

the Wet Tropics rainforests of Australia. Molecular Ecology, 7,

487–498.

Sharpe RG, Harbach RE, Butlin RK (2000) Molecular variation

and phylogeny of members of the Minimus Group of

Anopheles subgenus Cellia (Diptera: Culicidae). Systematic

Entomology, 25, 263–272.

Sodhi NS, Koh LP, Brook BW, Ng PKL (2004) Southeast Asian

biodiversity: an impending disaster. Trends in Ecology &

Evolution, 19, 654–660.

Steppan SJ, Zawadzki C, Heaney LR (2003) Molecular phylogeny

of the endemic Philippine rodent Apomys (Muridae) and the

dynamics of diversification in an oceanic archipelago.

Biological Journal of the Linnean Society, 80, 699–715.

Sterling EJ, Hurley MM (2005) Conserving biodiversity in

Vietnam: applying biogeography to conservation research.

Proceedings of the Californian Academy of Sciences, 56(Suppl. 1),

98–118.

Thompson J, Gibson T, Plewniak F, Jeanmougin F, Higgins D

(1997) The CLUSTAL_X windows interface: flexible

strategies for multiple sequence alignment aided by quality

analysis tools. Nucleic Acids Research, 25, 4876–4882.

Vidya TNC, Sukumar R, Melnick DJ (2009) Range-wide

mtDNA phylogeography yields insights into the origins of

Asian elephants. Proceedings of the Royal Society, B, 276, 893–

902.

Voris HK (2000) Maps of Pleistocene sea levels in Southeast

Asia: shorelines, river systems and time durations. Journal of

Biogeography, 27, 1153–1167.

White JC, Penny D, Kealhofer L, Maloney B (2004) Vegetation

changes from the late Pleistocene through the Holocene from

three areas of archaeological significance in Thailand.

Quaternary International: The Record of Human ⁄ Climate

Interaction in Lake Sediments, 113, 111–132.

Ying-shan P, Zhi-yi Z, Li-na P (2007) Strategic studies on the

biodiversity sustainability in Yunnan Province, Southwest

China. Forestry Studies in China, 9, 225–237.



Ziegler T, Abegg C, Meijaard E et al. (2007) Molecular

phylogeny and evolutionary history of Southeast Asian

macaques forming the M. silenus group. Molecular

Phylogenetics and Evolution, 42, 807–816.

This study forms part of the PhD thesis work of K.M. and

S.M.O’L. investigating the processes generating patterns of

genetic diversity in South-East Asia. B.C., D.T., P.S. and M.Y.F.

are mosquito biologists interested in the biodiversity of

mosquitoes in South-East Asia. Y-M.L. is a mosquito systema-

tist. R.B. is a population geneticist with research interests in

mechanisms of speciation. R.V. is currently studying for a PhD

in theoretical evolutionary genetics at Queen Mary, University

of London. A.P., P.T.H., T.H., S.N., D.S. and T.H.D. are inter-

ested in understanding mosquito biodiversity with respect to

the transmission of vector-borne disease in their respective

countries. C.W. is a population geneticist interested in under-

standing the processes generating and maintaining tropical

biodiversity.

Data accessibility

All sequence and sampling locality data, including GPS coordi-

nates, have been archived in Dryad with the reference

doi:10.5061/dryad.dc7m1. Sequence data generated in this

study are available from GenBank, under the accession num-

bers HQ403680–HQ404165.


Supporting information

Additional supporting information may be found in the online

version of this article.

Fig. S1 Bayesian Skyline Plots, as constructed in BEAST v1.4

[49], showing changes in effective population size within each

species over time.

Fig. S2 Correlation coefficients for the FST matrices of each

intraspecific dataset to the matrices of the 1000 datasets simu-

lated under various evolutionary scenarios in SPLATCHE. Ver-

tical bars indicate means. The red histogram shows the

correlation coefficients between the real data and data simu-

lated under the best fitting evolutionary model; the blue histo-

gram shows the coefficients under the scenario that differs

from the best fitting scenario only in the range of glacial carry-

ing capacities (Figure 2a or b); the black histogram shows the

coefficients under the null model (see Table 1 for details).

Fig. S3 The relationship between genetic diversity, as repre-

sented by hp, and latitude, across the seven species included

within this study.

Table S1 Details of sample locations and numbers of speci-

mens obtained for each species.

Please note: Wiley-Blackwell are not responsible for the content

or functionality of any supporting information supplied by the

authors. Any queries (other than missing material) should be

directed to the corresponding author for the article.

Comparative phylogeography reveals a shared impact of ... · ROBERT VERITY,* ANIL PRAKASH,–– PE THAN HTUN,*** THAUNG HLAING,*** SIMONE NAMBANYA,††† DUONG SOCHEAT,‡‡‡

Documents

Comparative phylogeography reveals a shared impact of ... · ROBERT VERITY,* ANIL PRAKASH,–– PE THAN HTUN,* THAUNG HLAING,* SIMONE NAMBANYA,††† DUONG SOCHEAT,‡‡‡