New insights into the intricate taxonomy and phylogeny of the Sylvia curruca complex

Full text

New insights into the intricate taxonomy and phylogeny
of the Sylvia curruca complex
Urban Olsson
a,
⇑
, Paul J. Leader
b
, Geoff J. Carey
b
, Aleem Ahmed Khan
c
, Lars Svensson
d
, Per Alström
e,f
a
Section of Systematics and Biodiversity, Biology and Environmental Sciences, University of Gothenburg, Box 463, SE-405 30 Gothenburg, Sweden
b
AEC Ltd., 127 Commercial Centre, Palm Springs, Hong Kong
c
Zoology Division, Institute of Pure and Applied Biology, Bahauddin Zakariya University, Multan 60800, Pakistan
d
S:ta Toras väg 28, SE-269 77 Torekov, Sweden
e
Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, Chaoyang District, Beijing 100101, China
f
Swedish Species Information Centre, Swedish University of Agricultural Sciences, Box 7007, SE-750 07 Uppsala, Sweden
article info
Article history:
Received 26 March 2012
Revised 16 November 2012
Accepted 31 December 2012
Available online 12 January 2013
Keywords:
Phylogeography
Demography
Species limits
abstract
We use the mitochondrial cytochrome bfrom 213 individuals and the three nuclear introns BRM 15,
myoglobin 2 and ODC 6–7 from a smaller subsample to evaluate the taxonomy of the Lesser Whitethroat
Sylvia curruca (Aves, Passeriformes, Sylviidae) complex, which has long been controversial. We sequenced
type material of the taxa althaea,blythi,margelanica and minula, and used topotypical material of cauca-
sica,chuancheica,curruca and telengitica. The nuclear introns fail to resolve the complex, but cytochrome b
recovers six major clades, revealing genetically identifiable populations corresponding to previously
named taxa, and we propose that the names althaea,blythi,curruca,halimodendri,margelanica and minula,
respectively, should be used for these. The margelanica clade is suggested to have a more extensive dis-
tribution than previously known, including both the taxon telengitica and a population in eastern Mon-
golia. The taxon minula is found to have a more restricted range than generally believed, only breeding
in China. According to the mitochondrial gene tree, there is a basal dichotomy, with the taxa althaea,bly-
thi,halimodendri and margelanica being part of one clade, well separated from a clade containing curruca
and minula. Dating analysis suggests that a basal divergence separating curruca and minula from the other
four taxa occurred between 4.2 and 7.2 mya; these two then diverged between 2.3 and 4.4 mya. The splits
between the althaea,blythi,halimodendri and margelanica lineages is inferred to have occurred later,
approximately between 1.0 and 2.5 mya (all 95% HPD). The nucleotide data suggest significant departure
from demographic equilibrium in blythi (clade 1a), halimodendri (clade 2a) and minula, whereas tenden-
cies are weaker for other clades. We propose that the names althaea,blythi,curruca,halimodendri,margel-
anica and minula should be used for the major clades. However, whether these are treated as subspecies
or species is largely a matter of species definition and is not resolved by our data.
Ó2013 Elsevier Inc. All rights reserved.
1. Introduction
The Lesser Whitethroat Sylvia curruca complex (Aves, Passeri-
formes, Sylviidae) is a group of morphologically similar insectivo-
rous warblers, breeding across almost the entire Palearctic, from
western Europe to eastern Siberia, and southwards through
north-western China and Central Asia to south-western Iran and
eastern Turkey (Cramp, 1992; Dickinson, 2003; del Hoyo et al.,
2006; Mayr and Cottrell, 1986; Shirihai et al., 2001; Vaurie,
1959;Fig. 1). The morphological variation among taxa is subtle
and the taxonomy complex, with fifteen names introduced for
thirteen proposed taxa (Table 1), but little consensus as to circum-
scription and taxonomic rank (Table 2) (cf. del Hoyo et al., 2006;
Loskot, 2005; Mayr and Cottrell, 1986; Shirihai et al., 2001; Vaurie,
1959). Due to the complex taxonomic situation, we will in the fol-
lowing use epithets in isolation when referring to a certain taxon,
without any indication of its taxonomic position or rank. For exam-
ple, telengitica refers to the population from which the holotype of
this taxon was drawn, but whether it is a junior synonym, subspe-
cies or species is not implied. Linnaean binominals or trinominals
will only be used to explicitly denote cases where the taxonomic
rank is suggested. For example, Sylvia margelanica telengitica would
imply that telengitica is treated as a subspecies of Sylvia
margelanica.
The taxa differ mainly in shades of brown and grey, wing length,
tail length and wing formula, and in many cases single individuals
are difficult to diagnose with certainty. The situation is further
1055-7903/$ - see front matter Ó2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.ympev.2012.12.023
⇑
Corresponding author.
E-mail addresses: urban.olsson@bioenv.gu.se (U. Olsson), pjl@aechk.hk
(P.J. Leader), gjc@aechk.hk (G.J. Carey), aleembzu25@yahoo.com.au (A.A. Khan),
lars@lullula.se (L. Svensson), per.alstrom@slu.se (P. Alström).
Molecular Phylogenetics and Evolution 67 (2013) 72–85
Contents lists available at SciVerse ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev

complicated as some taxa have been described from specimens
collected away from the breeding grounds. Both blythi and minula
were collected on the wintering grounds in India, and the two syn-
types of margelanica were collected on migration (in April and
October) in Uzbekistan (del Hoyo et al., 2006; Shirihai et al.,
2001; Stolzmann, 1897; Vaurie, 1954). Inferences of their breeding
areas thus rest on morphological comparison between the holo-
types and specimens collected on their respective breeding
grounds, adding a degree of uncertainty both concerning which
names should be applied to breeding populations and their exact
ranges. The taxa halimodendri,jaxartica and snigirewskii were all
described from specimens collected relatively close together, in
an area where there are no apparent geographical or ecological
barriers to dispersal between the type localities. It is thus unclear
whether they represent separate evolutionary entities occupying
parapatric ranges; separate evolutionary entities, some of which
were on migration to unknown breeding areas; extremes of mor-
phological variation within a single lineage; or whether they were
described from insufficient or in other ways poorly representative
series.
Many authors judged the morphological variation within the
Lesser Whitethroat complex to be clinal, from the palest desert
form minula to the dark form curruca of the temperate forests
and the even darker mountainous althaea. This gradual variation
further complicates the taxonomy and has, together with other
considerations, led some authors to treat the complex as a single
polymorphic species (e.g. Cramp, 1992; Dementiev and Gladkov,
1968; Mayr and Cottrell, 1986; Portenko, 1960; Ripley, 1982;
Svensson, 1992; Williamson, 1976), two species (Loskot, 2005),
or three different species (del Hoyo et al., 2006; Martens and Steil,
1997; Vaurie, 1954). Shirihai et al. (2001) divided the complex into
four allospecies: Sylvia [c.]curruca (with subspecies halimodendri),
S.[c.]minula,S.[c.]althaea and S.[c.] margelanica. Most authors
separated northern and southern taxa on the basis of differences
in habitat choice and colour differences. For similar reasons, the
mountainous althaea was almost always treated as a valid species.
Studies of the latter have suggested that it is reproductively iso-
lated from the parapatric halimodendri which breeds within ‘‘cruis-
ing range’’ at lower altitudes (Korelov, 1972; Loskot, 2001a;
Stepanyan, 1983). The northern birds inhabiting temperate forest,
curruca and blythi, are generally larger and darker on the upper-
parts than lowland desert and steppe forms, and taxonomic prob-
lems mainly concern whether or not to recognise the latter as a
valid taxon. Snigirewski (1929) remarked that the Siberian popula-
tions differ on average in wing formula from birds from Europe and
appear to approach the level of subspecies, and these were some-
times referred to as ‘‘morpha affinis’’. Ticehurst and Whistler
(1933) considered affinis Blyth, 1845 a junior synonym of althaea,
and the name has been suppressed according to opinion 1037 by
the ICZN (1976). Instead, Ticehurst and Whistler (1933) introduced
the name blythi for this form. Most authors now recognise blythi as
a valid taxon, but Shirihai et al. (2001) argued that more than 25%
of the individuals are impossible to separate from curruca and that
the taxon should thus not be recognised. Many authors separated
the populations in eastern Turkey, Caucasus and northern Iran
(caucasica (Ognev and Ban’kovski, 1910)), and some also recogni-
sed those in the Zagros mountains (zagrossiensis (Zarudny,
1911)), but both these were treated as junior synonyms of curruca
by Shirihai et al. (2001).
The desert forms are even more difficult from a taxonomic per-
spective. The taxa halimodendri (originally described from the bor-
ders between deserts and steppe north-east of the Aral Sea, and
apparently widespread in much of Kazakhstan and adjacent areas
south of it), minula (centered around the Tarim Basin, Xinjiang
province, China) and margelanica (centered around the Gansu
and Qinghai provinces, China) were recognised by all authors, al-
beit at different taxonomic levels (Table 2). Much less consensus
Table 2
Taxonomic treatment of the taxa in the Lesser Whitethroat complex by different authors.
Taxon Vaurie (1959) Mayr and Cottrell
(1986)
Cramp (1992) Shirihai et al. (2001) Loskot (2005) del Hoyo et al. (2006)
althaea S. a. althaea S. c. althaea S. c. althaea S. [c.] althaea S. althaea S. althaea
blythi S. c. blythi Synonym of curruca S. c. blythi Synonym of curruca S. c. blythi Synonym of curruca
caucasica Synonym of curruca S. c. caucasica S. c. caucasica Synonym of curruca S. c. caucasica Synonym of curruca
chuancheica Synonym of
margelanica
S. c. chuancheica – Synonym of
margelanica
Synonym of
margelanica
Synonym of
margelanica
curruca S. c. curruca S. c. curruca S. c. curruca S. [c.] curruca curruca S. c. curruca S. c. curruca
halimodendri S. c. halimodendri S. c. halimodendri S. c. halimodendri S. [c.] curruca
halimodendri
S. c. halimodendri S. c. halimodendri
jaxartica Synonym of minula S. c. jaxartica S. c. jaxartica Synonym of
halimodendri
Synonym of
halimodendri
Synonym of minula
margelanica S. m. margelanica S. c. margelanica S. c. margelanica S. [c.] margelanica S. c. margelanica S. m. margelanica
minula S. m. minula S. c. minula S. c. minula S. [c.] minula S. c. minula S. m. minula
monticola S. a. monticola S. c. monticola –– ––
snigirewskii S. c. snigirewskii Synonym of jaxartica Synonym of S. c.
jaxartica
Synonym of
halimodendri
S. c. snigirewskii Synonym of minula
telengitica S. c. telengitica S. c. telengitica S. c. telengitica Synonym of
halimodendri
S. c. telengitica Synonym of
halimodendri
zagrossiensis S. a. zagrossiensis Synonym of caucasica – – – Synonym of curruca
Table 1
List of names proposed for subspecies in the Lesser Whitethroat complex in
chronological order, with authors and type localities.
curruca (L., 1758); type locality ‘‘Sweden’’
affinis Blyth, 1845; type locality ‘‘southern India’’
minula Hume, 1873; type locality Bahwalpur, ‘‘India’’ (now in Pakistan)
althaea Hume, 1878; type locality Kashmir, India
margelanica Stolzmann, 1897; type locality Fergana, Uzbekistan
halimodendri Sushkin, 1904; type locality lower Irgiz and Turgay Rivers,
Kazakhstan
caucasica Ognev and Ban’kovski, 1910; type locality Mtskheta, near Tbilisi,
Georgia
zagrossiensis Zarudny, 1911; type locality Zagros Mts, Iran
telengitica Sushkin, 1925; type locality Chuya steppe, Russia
turkmenica Snigirewski, 1927; type locality near Repetek, Turkmenistan
chuancheica Portenko, 1929; type locality Gomi, Qinghai Province, China
jaxartica Snigirewski, 1929; type locality lower Syrdar’ya River valley,
Kazakhstan
snigirewskii Stachanow, 1929; new name for turkmenica Snigirewski, 1927,
preoccupied by Sylvia mystacea turcmenica Zarudny and Bilkevich, 1918.
blythi Ticehurst and Whistler, 1933; type locality Cawnpore, India
monticola Portenko, 1955; type locality Kvak, near Dushanbe, Tadjikistan
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 73

is found concerning their exact ranges and whether or not to rec-
ognise intervening and peripheral populations. A comparison be-
tween the two most recent studies of these taxa illustrates this
point (Fig. 1). Loskot (2005) considered the range of halimodendri
to extend to the lower Amudar’ya River in the south and treated
telengitica as a valid taxon. Shirihai et al. (2001) restricted the
range of halimodendri to the northern parts of Kazakhstan and ex-
tended the range in the east to include western Mongolia, treating
telengitica as a junior synonym of halimodendri.Shirihai et al.
(2001) considered the range of minula to extend from Turkmenis-
tan to Ningxia Province, China, south of the range of halimodendri,
while Loskot (2005) restricted the range of minula to the Tarim Ba-
sin and explicitly stated that the taxon does not breed west of Chi-
na. Shirihai et al. (2001) considered jaxartica and snigirewskii to be
junior synonyms of minula. Loskot (2005) concluded that jaxartica
is a junior synonym of halimodendri, but treated snigirewskii in the
Karakum Desert as a valid taxon. The range of margelanica is less
contentious, but there is disagreement between Loskot (2005)
and Shirihai et al. (2001) regarding the identity of the birds inhab-
iting the Qaidam depression in China.
We here present a phylogenetic and taxonomic evaluation, a
dating analysis and demographic observations of the Lesser White-
throat complex, based on comprehensive molecular data from the
mitochondrial cytochrome bgene, with additional data from the
nuclear introns BRM 15, myoglobin 2 and ODC 6–7, sampled
throughout the range of the Sylvia curruca complex. To ascertain
the correct use of names, we sequenced the cytochrome bgene
for type specimens of the taxa althaea,blythi,margelanica and min-
ula, and used topotypical material of caucasica,chuancheica,curru-
ca,jaxartica and telengitica.
2. Methods
2.1. Sampling
A total of 232 Lesser Whitethroats were sampled from northern
Europe, Central Asia, Siberia, Mongolia and China (Fig. 2 and
Supplementary Table 1), as well as from wintering grounds in West
Africa and the Arabian Peninsula. We used Sylvia cantillans as out-
group for rooting the phylogeny. An effort was made to include
samples of all named taxa in the Lesser Whitethroat complex,
but we were unable to procure samples from anywhere in Iran,
lowland Uzbekistan or Turkmenistan, despite searching suitable
habitat during the breeding season in the latter, including the type
locality of snigirewskii. Iran is of particular interest, as it is often in-
cluded in the range of minula.
We sequenced the holotypes of althaea,blythi and minula, kept
in the Natural History Museum (NHM), Tring, UK, and the two syn-
types of margelanica, kept in the Museum and Institute of Zoology,
Polish Academy of Sciences, Warszawa, Poland. We applied for per-
mission to sample the holotypes of halimodendri,jaxartica,snig-
irewskii and telengitica, which are kept in the Zoological Institute
of Russian Academy of Sciences, St Petersburg (Loskot, 2001b),
but this request was denied, as no destructive sampling is allowed
from holotypes (Vladimir Loskot, personal communication). How-
ever, we obtained samples from the type locality of telengitica on
the Chuya steppe, Russia; from Qinghai Province, China near the
type locality of chuancheica; from Syrdar’ya river, the type locality
of jaxartica; and from a variety of localities in Kazakhstan repre-
senting the presumed breeding range of halimodendri. Territorial
summer birds of the taxon curruca, with type locality ‘‘Sweden’’
were sampled from various localities in Sweden, including the
province Uppland where Carl von Linné (Linnaeus), the author of
this taxon, was active. The taxon caucasica, was sampled from
the type locality Tbilisi, Georgia, and adjacent areas, also territorial
summer birds.
2.2. DNA extraction and sequencing
DNA was extracted from toepads, muscle, blood or feathers,
using QIA Quick DNEasy Kit (Qiagen, Inc.) according to the manu-
facturer’s instructions, but with 30
l
l DTT added to the initial incu-
bation step of the extraction of toepads and feathers.
Approximate eastern
limit of halimodendri
Approximate western and
northern limit of minula
Approximate extension
of hybrid zone between
telengitica and blythi
Approximate northern
limit of snigirewskii
Approximate extension
of hybrid zone between
minula and margelanica
Russia
Kazakhstan Mongolia
China
Iran
Tu
Uzb
Afgh
Tu
Pakistan
Fig. 1. Breeding range of the Lesser Whitethroat Sylvia curruca complex based on Shirihai et al. (2001; used with permission). Ranges according to Loskot (2005) are very
approximately indicated by blue lines and boxes when differing from Shirihai et al. (2001). Afgh = Afghanistan, Tu = Turkmenistan, Uzb = Uzbekistan.
74 U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85

Amplification and sequencing of the cytochrome bgene (cyt b)
followed the protocols described in Olsson et al. (2005). The cyt b
gene was amplified as one fragment to decrease the risk of ampli-
fying nuclear pseudocopies (cf. e.g. Arctander, 1995; Sorenson and
Quinn, 1998). The cyt bgene of the toepad samples from the type
specimens were sequenced in small fragments, using a range of
specific primer combinations (Supplementary Table 2). We also se-
quenced intron sequences from the Z-linked brama protein gene
(BRM) and the autosomal genes ornithine decarboxylase (ODC)
and myoglobin (myo). Amplification and sequencing of intron 15
of BRM followed Borge et al. (2005); of introns 6–7 of the ODC gene
Allen and Omland (2003), Friesen et al. (1999), and Irestedt et al.
(2006); and of intron 2 of the myo gene Olsson et al. (2005).
2.3. Phylogenetic analyses
Sequences were aligned using MegAlign 4.03 in the DNAstar
package (DNAstar Inc.). For the nuclear data sets some manual
adjustments were necessary.
The choice of substitution model was determined based on the
Akaike Information Criterion (Akaike, 1973) and a hierarchical like-
lihood ratio test (Posada and Crandall, 1998), both calculated in
MrModeltest (Nylander, 2004). The preferred model for cyt bwas
HKY 85 (Hasegawa et al., 1985), assuming rate variation across
sites according to a discrete gamma distribution with four rate cat-
egories (
C
;Yang, 1994) and an estimated proportion of invariant
sites (I; Gu et al., 1995). For BRM the preferred model was HKY +
C
,
and for ODC and myo HKY + I.
The cyt bdataset was analysed using BEAST version 1.7.0
(Drummond et al., 2012a; Drummond and Suchard, 2012). Xml
files for the BEAST analyses were generated in BEAUti version
1.7.0 (Drummond et al., 2012a,b). Analyses were run using (a) a
fixed clock rate of 2.1%/MY (Weir and Schluter, 2008) or (b) an
uncorrelated lognormal relaxed clock (Drummond et al., 2006)
with the same mean rate, in all possible combinations with (i) a
constant population size prior or (ii) a population expansion
growth prior. All sequences, including identical haplotypes (as sug-
gested at http://beast.bio.ed.ac.uk/FAQ#Should_I_remove_identi-
cal_sequences.3F), but no outgroups, were included. Other priors
were used with default values. 5 10
7
generations were run, sam-
pled every 1000 generation. Every analysis was run twice. The
MCMC output was analysed in Tracer version 1.5.0 (Rambaut and
Drummond, 2009) to evaluate whether valid estimates of the pos-
terior distribution of the parameters had been obtained. The first
25% of the generations were discarded as ‘‘burn-in’’, well after sta-
tionarity of chain likelihood values had been established. Trees
were summarized using TreeAnnotator version 1.7.0 (Rambaut
and Drummond, 2012), choosing ‘‘Maximum clade credibility tree’’
and ‘‘Mean heights’’, and displayed in FigTree version 1.3.1 (Ram-
baut, 2009).
Gene trees for the nuclear introns were estimated by Bayesian
inference (BI) using MrBayes 3.1.2 (Huelsenbeck and Ronquist,
2001, 2005; Ronquist and Huelsenbeck, 2003). All loci were ana-
lysed separately (single-locus analyses, SLAs). Default priors were
used. Four Metropolis-coupled MCMC chains were run for
510
7
generations, and sampled every 1000 generations; the
heating temperature was set to 0.1. Two independent analyses
were run simultaneously, starting from random trees (per default).
The first 25% of the generations were discarded as ‘‘burn-in’’, well
after stationarity of chain likelihood values had been established
by manual inspection, and the posterior probability was estimated
for the remaining generations. To establish how well each model fit
the data, we calculated Bayes Factors (BF; Kass and Raftery, 1995;
Newton and Raftery, 1994) in Tracer 1.5.0 (Rambaut and Drum-
mond, 2009) using the harmonic mean as an approximation of
the marginal likelihood of a model.
We also analysed a data set based on one representative of each
unique cyt bhaplotype, which were included in a Bayesian Infer-
ence (BI) phylogenetic analysis. Four Metropolis-coupled MCMC
chains were run for 5 10
7
generations, sampled every 1000 gen-
eration. The first 25% of the generations were discarded as ‘‘burn-
in’’, well after stationarity of chain likelihood values had been
Fig. 2. Sampling sites of birds presumed to be on their breeding grounds. For clade 2, unfilled triangles indicate localities we visited but where we were unable to observe any
representatives of the Lesser Whitethroat complex. Samples from migrating birds are not shown. The same information is summarized in Fig. 4. (Based on Shirihai et al., 2001;
used with permission.)
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 75

established, and the posterior probability estimated for the
remaining generations.
Genetic distances were calculated for the cyt bdata set of un-
ique haplotypes of each pair of clades separately, following the rec-
ommendations of Fregin et al. (2012) of using ‘‘complete deletion’’,
perfectly homologous parts of the sequences and optimal substitu-
tion models.
2.4. Demographic analyses
Descriptive statistics of nucleotide variation and population dif-
ferentiation were calculated for the seven main clades separately
using DNAsp 5.0 (Librado and Rozas, 2009). Subclades 1a and 1b
within the blythi clade and clades 2a and 2b were analysed both
separately and combined.
We performed three statistical tests to evaluate evidence for or
against the populations being in demographic equilibrium, using
different sources of information. We chose one method using
information from the distribution of mutation frequencies, R
2
(Ramos-Onsins and Rozas, 2002). In a growing population, an
excess of recent mutations would be expected to produce an excess
of singletons (substitutions present in only one sampled sequence)
(Slatkin and Hudson, 1991; Tajima, 1989). R
2
is based on the differ-
ence between the number of singleton mutations and the average
number of nucleotide differences. The null hypothesis is a popula-
tion at demographic equilibrium, and after a recent strong popula-
tion growth, values of R
2
are expected to become lower. Fu’s F
S
test
statistic (Fu, 1997), on the other hand, is based on haplotype distri-
bution. Also in this case, the null hypothesis is a population at
demographic equilibrium, and an excess of singleton mutations
caused by an expansion would result in low values. The third
statistic, the raggedness index, r, uses information from the distri-
bution of the pairwise sequence differences (the mismatch distri-
bution) and evaluates the fit of the data to a model of population
expansion. A recent population expansion is expected to result in
a unimodal distribution of pairwise differences among haplotypes
with a smooth distribution, while a stable (non-expanding)
population at demographic equilibrium is expected to result in a
multi-modal, ragged, signal (Harpending, 1994; Rogers and
Harpending, 1992; Slatkin and Hudson, 1991). A deviation from
the null hypothesis of population expansion results in lower
values. According to Ramos-Onsins and Rozas (2002), the most
powerful tests for detecting population growth are Fu’s F
S
test
and their R
2
test, whereas the raggedness is relatively weak. They
found the behaviour of the R
2
test to be superior for small sample
sizes, whereas F
S
was better for large sample sizes (Ramos-Onsins
and Rozas, 2002). The raggedness index is nevertheless included
here as it represents a different approach.
For the three clades for which population expansion was in-
ferred, we calculated the time of the expansion event from the
equation
s
=2
l
t, where
l
is the mutation rate per sequence and
per generation, and t measures expansion time in generations.
The
l
value was calculated by multiplying the substitution rate
of cyt b(1.05 10
8
substitutions/site/year (Weir and Schluter,
2008); by 1040 (the sequence length). The generation time was
estimated at 1.7 years per generation, which reflects the average
for several passerine species (Sæther et al., 2005).
3. Results
We obtained a contiguous 1076 base pair portion of the cyt b
gene and flanking region of tRNA-Thr from 209 individuals con-
taining 97 unique haplotypes in the Sylvia curruca complex and
one outgroup species. A further four sequences were incomplete.
The phylogenetic analyses of the light strand of the cyt bgene con-
tains 1077 characters, of which 143 (13%) are parsimony informa-
tive, 853 (79%) are constant, and 81 (8%) are variable but
parsimony-uninformative. No unexpected start or stop codons that
could indicate the presence of a nuclear copy were observed in the
cyt bsequences.
The nuclear data sets all contained very few parsimony infor-
mative sites (BRM 1.8%, myo 1.2%, ODC 1.5%), and failed to resolve
the different populations. Furthermore, the nuclear haplotypes
showed no evidence of geographic or taxonomic structure, and
were scattered among representatives of different taxa in an
apparently random pattern (Supplementary Figs. S1–S3).
Genetic pairwise distances between clades are shown in Table 3,
as average distances and total range.
Details of origin and GenBank accession numbers are given in
Supplementary Table 1.
3.1. Phylogenetic analyses
The BF comparison favoured the analysis with a strict clock rate
and constant population size, although the support for the choice
between this model and a strict clock with an expansion growth
population or an uncorrelated lognormal relaxed clock with a con-
stant population size was not strong (Table 4).
Six major clades were identified, both in the analysis of the
complete cyt bdata set (Fig. 3) and the analysis of unique haplo-
types (Fig. 4). The branching pattern among clades 1–4 differ be-
tween these analyses, but none of the conflicts are well
supported. The samples included in clade 1 that were obtained
during the breeding season are consistent with the published range
of blythi (del Hoyo et al., 2006; Mayr and Cottrell, 1986; Shirihai
et al., 2001; Vaurie, 1959). Clade 1 is divided into two well sup-
ported subclades, 1a and 1b, in the analysis of the entire data set
(Fig. 3), but clade 1b is not recovered in the phylogeny based on
unique haplotypes (Fig. 4). The holotype of blythi is part of clade
1a, and although many samples in this clade were obtained from
migrating individuals, no representatives of any other haplotype
clade were obtained from temperate forest within the Asian range
of the S. curruca complex (Fig. 2,Supplementary Table 1). We have
a single sample from clade 1b from a possible breeding area in
Mongolia.
Table 3
Genetic pairwise distances between clades calculated following the recommendations of Fregin et al. (2012). Average distances between clades are given to the lower left of
diagonal, and total range to the upper right.
blythi halimodendri 2a halimodendri 2b margelanica althaea curruca minula
blythi – 2.1–2.9 2.5–3.4 3.3–5.8 2.2–3.5 11.4–14.0 13.3–16.3
halimodendri 2a 2.4 – 1.0–1.5 3.6–4.9 1.9–2.8 9.9–12.0 17.4–22.6
halimodendri 2b 2.9 1.3 – 3.0–3.6 1.9–2.5 9.4–10.9 8.2–9.1
margelanica 4.7 4.1 3.3 – 3.4–5.3 9.2–11.4 10.8–13.1
althaea 2.9 2.4 2.2 4.4 – 8.2–9.9 13.0–16.1
curruca 13.0 10.9 10.2 10.0 9.2 – 6.2–7.3
minula 14.8 20.0 8.7 11.9 14.6 6.7 –
76 U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85

Clade 2 (Figs. 3 and 4) best matches the published range of hal-
imodendri (del Hoyo et al., 2006; Mayr and Cottrell, 1986; Shirihai
et al., 2001; Vaurie, 1959). This could not be confirmed by compar-
ison to the holotype or topotypical specimens, but summer sam-
ples obtained from the presumed breeding range of this taxon all
belonged in this clade. During migration, also representatives of
clades 1 and 4 were obtained in the same region. Among the sam-
ples obtained from this area during summer, there is no indication
that there is more than one taxon inhabiting the area. One sample
from the type locality of jaxartica at the Syrdar’ya river is nested in
halimodendri clade 2a. No Lesser Whitethroats were encountered
by us in Turkmenistan or the central or eastern parts of Uzbekistan,
despite active search for birds breeding in the area. Clade 2 is di-
vided into two subclades, 2a and 2b (Figs. 3 and 4). Very little is
known about the population making up clade 2b, as all except
one of the samples are of migrating or wintering birds. No distin-
guishing features are known, but our single summer record of an
individual representing clade 2b was obtained in Gansu, China, at
a locality where margelanica was a common breeding species. It
is thus possible that this clade may occupy a different breeding
range than clade 2a, but this needs further study.
Clade 3 (Figs. 3 and 4) matches the published range of margela-
nica (del Hoyo et al., 2006; Mayr and Cottrell, 1986; Shirihai et al.,
2001; Vaurie, 1959). The haplotypes of the two syntypes of margel-
anica are part of this clade, and match samples from the published
distribution of this taxon. This clade also includes a sample from
the type locality of telengitica. One sample obtained from eastern
Mongolia, outside the published range of margelanica, is firmly
placed in this clade. Some samples were obtained in Xinjiang prov-
ince, China at the end of the summer, but at locations where only
representatives of either clade 2 or clade 6 were present during the
breeding season; these samples are considered to have been col-
lected from migrants.
Clade 4 (Figs. 3 and 4) matches the published range of althaea
(del Hoyo et al., 2006; Mayr and Cottrell, 1986; Shirihai et al.,
2001; Vaurie, 1959), and the holotype of althaea is part of this
clade. However, the sequence of the holotype is incomplete, and
was excluded from the final analysis.
Clade 5 (Figs. 3 and 4) matches the published range of curruca
(del Hoyo et al., 2006; Mayr and Cottrell, 1986; Shirihai et al.,
2001; Vaurie, 1959), and several samples are topotypical. Samples
from eastern Turkey and the Caucasus, including from the type
locality of caucasica, are part of this clade and form a subclade
which, however, receives insufficient support.
Clade 6 (Figs. 3 and 4) matches minula, sensu Loskot (2005), and
the holotype of minula is part of this clade. It also includes several
samples from the Qaidam.
3.2. Dating analysis
The results of the dating analysis (Fig. 3) suggest that a basal
divergence separating curruca and minula from the other four taxa
occurred approximately 5.7 million years ago (mya) (95% HPD:
4.2–7.2 mya); these two then diverged from each other around
3.3 mya (95% HPD: 2.3–4.4 mya). The splits between the althaea,
blythi,halimodendri and margelanica lineages are inferred to have
occurred later, beginning approximately 2 mya (95% HPD: 1.4–
2.5 mya). Resolution between the althaea,blythi,halimodendri
and margelanica is not sufficiently supported, and the dating of
individual nodes uncertain. Nevertheless, all these clades appear
to have diverged during a time period between 1.0 and 2.5 mya,
i.e. the first half of the Pleistocene.
3.3. Demographic analyses
The population genetics analyses of the sequence data, summa-
rized in Table 5, suggest significant departures from the null
hypotheses in several cases, both for Fu’s F
S
and R
2
, for which the
null hypothesis is a population at demographic equilibrium, and
for raggedness, r, for which the null hypothesis is population
expansion.
For clade 1, Fu’s F
S
and R
2
indicate departure from the null
hypothesis of a population at demographic equilibrium for cyt b,
although the level of significance is low for R
2
. The raggedness r,
points in the same direction in not departing from the null hypoth-
esis of a population expansion. When the cyt bdata is divided into
two groups, 1a and 1b, the historical demographic signal differs be-
tween these. For clade 1a, Fu’s F
S
and R
2
indicate significant depar-
ture from demographic equilibrium, whereas r does not deviate
from the null hypothesis of population expansion. The onset of
an expansion is estimated to 89 kya, which predates the last glacial
maximum (LGM) and coincides with the marine isotope stage 5
(MIS5) interglacial. For clade 1b, support for deviation from demo-
graphic equilibrium in Fu’s F
S
and R
2
is weak, and, in addition, the
raggedness index, r, indicates weak support for deviation from the
null hypothesis of population expansion. For BRM, Fu’s F
S
indicate
significant departure from the null hypothesis of a population at
demographic equilibrium, while R
2
and r do not depart from the
null hypothesis of a population at demographic equilibrium or of
a population expansion, respectively. For myo, Fu’s F
S
and R
2
indi-
cate departure from the null hypothesis of a population at demo-
graphic equilibrium, at the same time as there is weak indication
of departure from the null hypothesis of a population expansion
in r. For ODC, no deviation from the null hypothesis is detected,
but this could be an artefact due to smaller sample size.
For clade 2, there are indications in the cyt bdata of deviation
from the null hypothesis of demographic equilibrium in Fu’s F
S
and R
2
, and no deviation from the null hypothesis of population
expansion in the raggedness index, r. With the cyt bdata analysed
separately for the two subclades 2a and 2b, Fu’s F
S
and R
2
both sug-
gest significant departure from demographic equilibrium for clade
2a, supported by lack of deviation from the null hypothesis of pop-
ulation expansion in the raggedness index, r. The onset of the in-
ferred expansion is estimated to 90 kya, predating the LGM and
coinciding with the MIS5 interglacial. For clade 2b, no deviation
from the different null hypotheses is suggested for any indicator,
Table 4
log
10
Bayes Factors for BEAST analyses.
Model ln P
(model|data)
Strict clock, constant
pop.
Strict clock, expansion
growth pop.
Relaxed clock, constant
pop.
Relaxed clock, expansion
growth pop.
Strict clock, constant pop. 3778.552 – 6.021 5.361 82.098
Strict clock, expansion growth
pop.
3784.573 6.021 – 0.66 76.076
Relaxed clock, constant pop. 3783.913 5.361 0.66 – 76.737
Relaxed clock, expansion
growth pop.
3860.649 82.098 76.076 76.737 –
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 77

rendering the evidence contradictory. The nuclear loci were only
analysed with both clades combined. For BRM the raggedness
index, r, indicates deviation from the null hypothesis of population
expansion, whereas there is some indication of departure from
demographic equilibrium for both Fu’s F
S
and R
2
. For myo, both
Fu’s F
S
and R
2
suggest departure from demographic equilibrium,
while the raggedness index, r, does not suggest departure from
population expansion. For ODC, the raggedness index, r, does not
indicate deviation from the null hypothesis of population
expansion, but there is only weak support for departure from
demographic equilibrium for Fu’s F
S
and none from R
2
.
For clade 3, Fu’s F
S
for all loci, particularly cyt band BRM, indi-
cate possible departure from demographic equilibrium, but R
2
pro-
vides little support for this. The raggedness index, r, for the ODC
data indicates deviation from the null hypothesis of population
expansion, whereas for other loci no such deviation is suggested.
For clade 4, Fu’s F
S
and R
2
for the cyt bdata both indicate weak
support for a departure from demographic equilibrium, whereas
Fig. 3. Dated cytochrome btree of the Lesser Whitethroat complex for all samples. Numbers above nodes indicate posterior probabilities (

posterior probability1.00).
Numbers below nodes indicate estimated age, with 95% confidence interval in parentheses. Clade support and age estimates have been removed for minor clades.
78 U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85

the raggedness index, r, indicates weak support for a deviation
from the null hypothesis of population expansion. For the BRM
data, there is slight indication for a departure from demographic
equilibrium for Fu’s F
S
and R
2
, with no support for a deviation from
the null hypothesis of population expansion from r. For the myo
data, r indicates a more pronounced support for a deviation from
the null hypothesis of population expansion, at the same time as
there is only slight indication for a departure from demographic
equilibrium for Fu’s F
S
and none from R
2
. For ODC, none of the indi-
cators suggest deviation from their respective null hypotheses,
rendering the evidence contradictory.
For clade 5, both Fu’s F
S
and R
2
for the cyt bdata provide support
for departure from demographic equilibrium (although weak from
Fu’s F
S
), whereas the raggedness index, r, indicates departure from
Fig. 4. Cytochrome btree of the Lesser Whitethroat complex based on unique haplotypes. Maps on the right indicate sampling localities for individuals on their presumed
breeding grounds. Posterior probability for major clades is given above nodes (posterior probability of 1.00 is denoted by

). Dashed lines connect individual samples of
particular significance with their sampling localities. The phylogenetic position of type specimens are indicated, as are topotypical samples of curruca, for which the holotype
is lost, and for caucasica,chuancheica,jaxartica and telengitica, for which holotypes could not be sequenced. In the case of the holotype of althaea, the sequence obtained was
short. It was firmly placed in the althaea clade in preliminary analyses (not shown), but was excluded from the final analysis as the lower number of informative sites of a
short sequence affects both topology and branch support of the phylogeny.
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 79

a null hypothesis of population expansion. For the nuclear data,
there is no support for a deviation from the null hypothesis of pop-
ulation expansion from r. for any of the three loci. Fu’s F
S
and R
2
suggest weak support for departure from demographic equilibrium
for all loci except for BRM, where there is no indication of depar-
ture from a null hypothesis for R
2
.
For clade 6, Fu’s F
S
and R
2
for the cyt bdata both provide
strong support for departure from demographic equilibrium, at
the same time as there is weak support for deviation from the
null hypothesis of population expansion for the raggedness in-
dex, r. The onset of the inferred expansion is estimated to
90 kya, predating the LGM and coinciding with the MIS5 inter-
glacial. For the BRM data, there is indication of a departure from
demographic equilibrium for Fu’s F
S
and R
2
, with no support for
a deviation from the null hypothesis of population expansion
from r. For the myo and ODC data, the sample size was too
small to calculate Fu’s F
S
, but R
2
for myo indicates a departure
from demographic equilibrium, combined with lack of support
for a deviation from the null hypothesis of population expansion
from r. For ODC, R
2
does not indicate a departure from demo-
graphic equilibrium, and r does not deviate from the null
hypothesis of population expansion, rendering the evidence
somewhat contradictory, albeit based on weak signals.
4. Discussion
The Lesser Whitethroat complex provides an example of a
group of organisms that is rather well studied from a morpholog-
ical point of view (e.g. del Hoyo et al., 2006; Mayr and Cottrell,
1986; Shirihai et al., 2001; Svensson, 1992; Vaurie, 1959). Still,
there is no taxonomic consensus among different authors, neither
concerning circumscription of taxa nor regarding species limits.
This study thus adds important evidence to the taxonomic situa-
tion surrounding the Lesser Whitethroat complex, by demonstrat-
ing the existence of well supported clades based on mitochondrial
data, corresponding to previously named taxa. However, the nucle-
ar loci do not discriminate the different populations, thus failing to
support the divergence into separate evolutionary lineages sug-
gested by the mitochondrial data.
4.1. Taxonomy
4.1.1. Clade 1 – blythi
The taxon blythi differs morphologically very slightly from curr-
uca, and its validity has been questioned (del Hoyo et al., 2006;
Shirihai et al., 2001). These two taxa occur in similar habitat in
the temperate forest region that extends from Europe to Siberia,
Table 5
Descriptive statistics on genetic variation in the Sylvia curruca complex, with significance determined using coalescent simulations; all calculated in DNASP 5.10. (Librado and
Rozas, 2009).
Locus Taxon Length (bp) N
1
Hd
2
p
3
h
W4
S
5
P
6
Fu’s F
S7
R
28
r
9
s
10
CYTB (mtDNA) blythi 1077 65 0.818 0.00292 0.00685 33 3.144 16.043
***
0.0433
*
0.0587 0.921
blythi 1a 1077 49 0.716 0.00106 0.00396 19 1.145 20.713
***
0.0298
***
0.1143 1.145
blythi 1b 1077 16 0.625 0.00151 0.00364 13 1.625 2.373
*
0.1061 0.0519 0.226
halimodendri 1077 41 0.888 0.00481 0.00751 33 5.176 9.354
**
0.0693 0.0458 0.944
halimodendri clade 2a 1077 31 0.824 0.00155 0.00503 21 1.671 17.213
***
0.0391
***
0.0683 1.158
halimodendri clade 2b 1077 10 0.822 0.00140 0.00197 6 1.511 1.249 0.1638 0.1630 1.511
margelanica 1077 29 0.813 0.00164 0.00402 17 1.769 5.077
**
0.0811 0.0914 0.344
althaea 1077 15 0.638 0.00092 0.00143 5 0.991 1.548 0.1234 0.0488 0.990
curruca 1077 15 0.876 0.00265 0.00428 15 2.857 3.002 0.0849
**
0.0173
**
1.657
minula 1077 33 0.669 0.00117 0.00297 13 1.265 8.238
***
0.0506
***
0.0551 1.157
BRM (Z) blythi 291 12 1.000 0.01080 0.01260 10 3.121 11.049
***
0.1242 0.0487 3.121
halimodendri 291 10 0.848 0.00852 0.01150 10 2.455 3.251 0.1055
*
0.0110
**
1.682
margelanica 291 13 0.949 0.01109 0.01004 9 3.205 4.830
**
0.1594 0.0421 3.205
althaea 291 11 0.891 0.00906 0.01300 11 2.618 2.055 0.1172
*
0.0602 1.762
curruca 291 7 0.889 0.00810 0.01022 8 2.333 1.776 0.1680 0.1204 1.650
minula 291 7 0.952 0.00659 0.00847 6 1.905 3.409
**
0.1232
**
0.1293 1.905
MYO (A) blythi 617 10 1.000 0.00522 0.00745 13 3.222 7.942
***
0.0909
**
0.0494 3.222
halimodendri 617 9 0.972 0.00504 0.00656 11 3.111 4.215
**
0.0903
**
0.1597 3.111
margelanica 617 9 0.944 0.00549 0.00775 13 3.389 2.209 0.1441 0.0586 1.666
althaea 617 6 0.600 0.00162 0.00213 3 1.000 0.189 0.2546 0.0622
**
1.000
curruca 617 6 0.929 0.00486 0.00563 9 3.000 1.609 0.1242
*
0.2449 3.000
minula 617 3 1.000 0.00216 0.00216 2 1.333 NA 0.2357
*
0.6667 1.333
ODC (A) blythi 460 6 0.933 0.00493 0.00571 6 2.267 1.846 0.1848 0.1111 2.267
halimodendri 460 9 0.944 0.00640 0.00640 8 2.944 2.600
*
0.1544 0.0617 2.944
margelanica 460 7 1.000 0.00994 0.01154 12 4.571 3.232
*
0.1901 0.0317
**
3.060
althaea 460 5 0.400 0.00261 0.00313 3 1.200 1.688 0.4000 0.6800 0.105
curruca 460 7 0.944 0.00652 0.00800 10 3.000 2.547
*
0.1167
*
0.0895 3.000
minula 460 3 1.000 0.00399 0.00356 3 1.833 NA 0.2650 0.2778 1.833
(⁄)
0.1 > p> 0.05
*
p< 0.05.
**
p< 0.01.
***
p< 0.001.
1
N– Number of sequences analyzed.
2
Hd – Haplotype diversity (Nei, 1987).
3
p
– Nucleotide diversity.
4
h
W
– The population mutation parameter theta, estimated from No. of segregating sites.
5
S– No. of segregating sites.
6
P
– Average number of nucleotide differences between sequence pairs (Nei, 1987).
7
F
S
– Fu’s statistic (Fu, 1997).
8
R
2
– Ramos-Onsin and Roza’s statistic (Ramos-Onsins and Rozas, 2002).
9
r– Raggedness statistic (Harpending, 1994).
10
s
– Total mutations.
80 U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85

and due to their morphological similarity it is unclear where their
ranges meet. Our data suggest that blythi is a valid taxon, not clo-
sely related to curruca. It has its closest relatives to the south-east,
and may have colonised the eastern taiga from this direction, ulti-
mately coming into contact with curruca. We lack samples from
most of the area where the ranges of blythi and curruca would be
predicted to come into contact, but one sample from the Komi re-
gion in the north-easternmost corner of Europe (westernmost bly-
thi sample in Fig. 2) had a haplotype consistent with blythi.
4.1.2. Clade 2 – halimodendri
Most authors allocated halimodendri to the curruca group or
superspecies (Table 2). Presumably this was in most cases moti-
vated by the assumed hybridisation, manifested in the apparently
clinal transition between blythi and halimodendri (cf. Loskot, 2005;
Shirihai et al., 2001).
The taxonomy of the populations inhabiting lowland arid habi-
tats in the former Soviet republics in Central Asia could not be re-
solved by analysis of sequences from the holotypes of halimodendri,
jaxartica or snigirewskii, as no destructive sampling is allowed from
holotypes kept in St Petersburg (Vladimir Loskot, personal com-
ment). Loskot (2005), who studied the type series of all these, con-
sidered snigirewskii a valid taxon, but stated that the type series of
jaxartica is a mixture of blythi,halimodendri and minula, rendering
jaxartica a junior synonym of halimodendri. We have only found
evidence of one population breeding within the published range
of halimodendri (clade 2a, Figs. 3 and 4). A sample from the type
locality of jaxartica is firmly placed in clade 2a (Figs. 3 and 4), cor-
roborating the view of Loskot (2005) regarding jaxartica. However,
the validity of jaxartica and snigirewskii is neither supported nor re-
jected by this study. The combined evidence of our own data and
the analysis of Loskot (2005) support the assumption that the cor-
rect name of clade 2 is halimodendri.
4.1.3. Clade 2b – unnamed?
The individuals belonging to this clade have all been identified a
posteriori from their haplotypes. As it is not known whether this
clade represents a distinct population or not, suggesting a name
would be premature. We are not aware of any morphological or
other differences compared to halimodendri, but the possibility that
individuals representing clade 2b (Figs. 3 and 4) may be involved
among holotypes of jaxartica or snigirewskii cannot be rejected.
We have obtained several samples from Kazakhstan and Xinjiang
(China), but these were all taken at a time of year when the birds
could have been on migration. However, this is also true for the
holotypes of halimodendri,jaxartica or snigirewskii. A single individ-
ual from the breeding season belonging to this haplotype group
was obtained from within the breeding range of margelanica,ina
location where individuals of margelanica were also sampled
(Figs. 2–4). The possibility that the two populations of clades 2b
and 3 breed in sympatry should be further investigated. On present
evidence, this clade is best included under the name halimodendri.
4.1.4. Clade 3 – margelanica
The range of margelanica is not clear, but is clearly larger than
previously considered. The majority of our samples of breeding
birds come from the published range in Qinghai and Gansu prov-
inces, China. Beside these, one available sample from the Chuya
steppe, Russia, which is the type locality of telengitica, is part of
clade 3 with high support. This taxon has not previously been asso-
ciated with margelanica, although they are similar in size (Loskot,
2005; Shirihai et al., 2001). Furthermore, one sample collected dur-
ing the breeding season in grassland in eastern Mongolia near the
Chinese border, indicates that the range of the margelanica clade is
greater than previously assumed. This individual has a unique hap-
lotype somewhat diverged from other populations, but it is
unknown whether this is true for the entire local population. If
the population represents a valid taxon, it appears to be unnamed.
The two syntypes of margelanica were collected near Fergana in
Uzbekistan, approximately 1000 km from the nearest known
breeding area of this taxon, but their haplotypes match those of
the population in Qinghai and Gansu provinces, China. All our sam-
ples from this area are part of the same clade, and the taxon chu-
ancheica, which was described from the same area, must thus be
regarded as a junior synonym of margelanica.
4.1.5. Clade 4 – althaea
The taxon althaea is one of the least controversial taxa, accord-
ing to our data and previous studies (Korelov, 1972; Loskot, 2001a;
Stepanyan, 1983). All samples collected at high elevation during
the breeding season fall in the same clade as the holotype of alt-
haea, although there is substantial divergence within the clade.
None of our samples comes from the type locality of monticola
(Kvak, near Dushanbe, Tajikistan), but we have samples from with-
in 400 kms both to the north and south, and we consider it unlikely
that monticola would not be part of clade 4. However, the clade is
relatively divergent, and it is possible that there is a certain differ-
entiation between populations from different mountain areas,
although our sample is too limited to evaluate this.
4.1.6. Clade 5 – curruca
The curruca clade was sampled from two geographically well
separated areas, the UK and Sweden in the north-west represent-
ing curruca, and eastern Turkey and Armenia in the south-east rep-
resenting caucasica. This clade is rather diverse, but all haplotypes
from Turkey and Armenia are in one unsupported subclade. Fur-
ther studies are required to assess the implications of this.
South-eastern birds allocated to caucasica are often described as
being similar to althaea, but we have no indication in our data that
more than one form breeds in this area, and no evidence of althaea
occurring there.
4.1.7. Clade 6 – minula
The position of the minula clade as sister to curruca on a lineage
separate from the remainder of the Lesser Whitethroat complex
was unanticipated by all previous authors (e.g. del Hoyo et al.,
2006; Loskot, 2005; Mayr and Cottrell, 1986; Shirihai et al., 2001;
Vaurie, 1959). Another observation at odds with the opinion of
most previous authors, except Loskot (2005), is that there is no
indication that minula breeds outside of China. We have found no
evidence of minula in Kazakhstan or adjacent Tajikistan, and all
our samples from lowland arid areas in the former Soviet republics
in Central Asia, including from within the published range of min-
ula, belong in either the halimodendri clade as defined here, or to
probable migrants of clades 1, 3 or 4. We have searched in vain
in summer for Lesser Whitethroats in the southern parts of Tajiki-
stan and Turkmenistan, and the only evidence we have found of
birds inhabiting the purported range of snigirewskii is a specimen
in the American Museum of Natural History (AMNH) labelled hal-
imodendri from Repetek 3 August (AMNH 595898). According to
former staff at Repetek Nature Reserve (personal communication),
Turkmenistan, the type locality of snigirewskii, no Lesser White-
throats are known to breed there; instead, in suitable habitats,
Menetries’s Warbler S. mystacea occurs. Furthermore, we have
been unable to trace any published records of breeding from low-
land arid areas in Iran, although there are two AMNH summer
specimens from eastern Iran (AMNH 595906, 25 July, Biddadin
and AMNH 595907, 15 July, Garmab). During the breeding seasons
from 1971 to 1975 there were no records of birds in minula-type
plumage between the end of April and the middle of September,
and no indication of breeding anywhere in the country (Derek
Scott, personal communication). The only breeding records of
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 81

Lesser Whitethroats from Iran during this period are from popula-
tions in mountain regions, judged to be curruca or althaea (Derek
Scott, personal communication). Lesser Whitethroats often sing
on migration, and we suspect that such singing migrants may have
been mistaken for territory-holding males in the past. Martens and
Steil (1997) included in their analyses the songs of two males iden-
tified as minula from the Kopet Dagh in Iran, recorded at 2000 m on
7 May. Given the very high altitude and spring date, we consider it
most likely that these were migrants, if they were indeed minula.
We thus consider the southern limits of the Central Asian lowland
desert forms uncertain, and probably overestimated. However, our
own explorations have been restricted and future studies are
needed to clarify this.
4.2. Species limits
The congruent morphological differences (cf. del Hoyo et al.,
2006; Loskot, 2005; Shirihai et al., 2001; Vaurie, 1959), allopatric
ranges and substantial genetic divergence in the cyt bgene bear
evidence of long standing isolation between several populations.
If the cyt btree presented here represents the true phylogeny,
the possibility of several cases of cryptic speciation cannot be ruled
out. However, although the different taxa are generally identifiable
as unique clades, and vice versa, diagnostic morphological features
useful at an individual level are few and subtle, and many individ-
uals cannot be identified with certainty. As a result, the presence or
degree of interbreeding between different populations is difficult
to assess on current evidence. For some populations, such as be-
tween clade 5 and 6, ongoing gene flow seems highly unlikely
based on their widely allopatric ranges. Similarly, clade 4, althaea,
breeding in mountainous areas, has been judged to be reproduc-
tively isolated from parapatric populations breeding in the sur-
rounding lowlands, presumably clade 2a (halimodendri)as
circumscribed here (Korelov, 1972; Loskot, 2001a; Stepanyan,
1983). For other clades, the situation is less clear. The taxonomic
status among clades 1–3, and also their potential interactions with
clades 5 and 6 in areas of contact (if any) may not be satisfactorily
evaluated based on our data. The morphological similarity be-
tween many individuals from different clades is a confounding fac-
tor, and whether the haplotype of an individual is truly a marker of
phylogenetic affinity or the result of introgression cannot be deter-
mined with certainty in all cases. For example, the apparent con-
flict between morphology and phylogeny in the relations
between blythi and both curruca and halimodendri, respectively,
may be explained in different ways, with different consequences
for the taxonomic treatment. The cyt btree taken at face value sug-
gests that blythi is part of a clade including althaea,halimodendri
and margelanica and that the morphological similarity with curruca
might be due to parallel evolution of coloration, perhaps as an
adaptation to the temperate forest ecotone. Alternatively, the
mitochondrial gene tree does not reflect the species tree. In this
case, the morphological similarity between blythi and curruca
could be due to these being sister taxa, and their positions in the
gene tree being the result of ancient mitochondrial introgression.
A similar possible scenario was described by Olsson et al. (2010)
for Lanius excubitor, where the populations inhabiting temperate
regions from Scandinavia to Siberia are morphologically similar,
but where the mitochondrial gene tree indicates a deep split be-
tween lineages apparently coming into contact in more or less
the same geographical region as between curruca and blythi.I
n
both these cases a comprehensive study focusing on the species
phylogeny rather than gene trees would be required to resolve
the origins of these populations. Our own attemps in this respect
failed, as the nuclear loci we used all failed to resolve the
phylogeny.
Analysis of study skins suggests that birds gradually become pa-
ler along an axis from the taiga to the Central Asian steppe, from
the range of blythi to halimodendri (Loskot, 2005; Shirihai et al.,
2001; Vaurie, 1959). Such a clinal pattern could theoretically arise
by quite contrary mechanisms. In a case of isolation by distance,
the extremes of a continuously distributed population may accu-
mulate different traits adapted to different local conditions, while
secondary contact between populations previously adapted to dif-
ferent local conditions would be expected to cause morphological
differences to intergrade. Thus, the status of blythi in relation to
both curruca and halimodendri depends on its evolutionary history,
which is not possible to determine on the basis of the present evi-
dence. We thus consider blythi incertae sedis.
The ranges of minula and margelanica are said to meet in the
Qaidam depression where they are suggested to intergrade (Loskot,
2005; Vaurie, 1959). Our sample from Qaidam is limited to four
individuals, but these were all consistent with minula in cyt bhap-
lotype. Morphologically they were judged to be somewhat inter-
mediate between minula and margelanica, and geneflow can thus
not be rejected by our data.
In conclusion, only althaea appear to fulfil the requirements of
species sui generis according to three of the most widely adopted
species definitions (Cracraft, 1989; Mayr, 1942, 1963; de Queiroz,
2005), primarily based on previous evidence (Korelov, 1972; Los-
kot, 2001a; Stepanyan, 1983). However, recognising this taxon as
a full species separate from a polytypic species containing all the
other taxa would violate modern taxonomic principles. The
remaining taxa show various degrees of conflict with one or more
species definitions. Ongoing gene flow cannot be ruled out be-
tween blythi and curruca; between blythi and halimodendri; and be-
tween minula and margelanica, respectively. These may thus fail
the criterion of reproductive isolation (Mayr, 1942, 1963), although
there are no indications of this in our data. As the morphological
differences are slight, and many individuals cannot be identified
with certainty, on present knowledge most taxa fail the criterion
of diagnosability at the individual level (Cracraft, 1989). Judged
by the mitochondrial data, the six major clades indicated by this
study, and perhaps also clade 2b, all seem to fit the definition of
species according to the ‘‘general lineage concept of species’’ (de
Queiroz, 2005). However, despite the divergences apparently being
surprisingly old, differences have not yet become fixed in the nu-
clear markers, which is most likely the result of the longer coales-
cence times for nuclear markers (Zink and Baraclough, 2008),
although it might indicate continued (nuclear) gene flow. However,
additional data would be required to confirm that the cyt btrees
correspond to population level lineages. Consequently, defining
species limits in this group still presents a delicate challenge,
regardless of species definition.
All except clade 2b correspond to previously named taxa, and
we propose that available names should be applied as proposed
in Figs. 3 and 4. Species limits remain to be determined.
4.3. Dating analysis
The divergence separating the curruca and minula lineage from
the lineage giving rise to the other four taxa is estimated to
approximately 5.7 mya (95% HPD: 4.2–7.2 mya). This coincides
with a series of major tectonic events that gave rise to the Tien
Shan range, the Altai Mountains, the Kunlun Mountains (Sun
et al., 2008) and the Tibetan Plateau (Guo and Wang, 2007). The
divergence between curruca and minula around 3.3 mya (95%
HPD: 2.3–4.4 mya) just predates the shift to a cooler and fluctuat-
ing climate characterising the Pleistocene. The splits between the
althaea,blythi,halimodendri and margelanica lineages, beginning
approximately 2 mya (95% HPD: 1.4–2.5 mya), occured during
Pleistocene climate oscillations.
82 U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85

4.4. Demography
We have calculated three indicators that evaluate whether a
population is in demographic equilibrium or undergoing expan-
sion. The three methods utilize different sources of information,
the distribution of mutation frequencies, the haplotype distribu-
tion, and the mismatch distribution. These methods investigate
whether there is a deviation from a null hypothes; a deviation that
may have been caused by several different scenarios, such as the
population having undergone a recent population expansion, a
bottleneck or a selective sweep. We regard selective sweep as
the least likely explanation for these patterns as independent loci
convey the same signals for the three indicators of demographic
equilibrium. Whether the populations have expanded or gone
through a bottleneck is uncertain, but differences among clades
in values of
s
may lend evidence. In cyt b, the values of
s
are similar
for blythi,halimodendri and minula, resulting in similar estimate of
onset of expansion. For the nuclear loci they are often quite differ-
ent when these three clades are compared, indicating different
coalescent times for the clades, thus speaking against a population
bottleneck.
All these taxa occur in regions where Pleistocene glacial cycles
have had significant impact on habitat distributions and local cli-
mate. The region presently inhabited by halimodendri,margelanica
and minula has become increasingly arid since as early as the
Eocene, caused by the uplift of the Himalayan–Tibetan mountain
belt (Graham et al., 2005). The Miocene–Pliocene period was char-
acterised by a predominance of tropical and subtropical climates in
this region. Present day Kazakhstan and Xinjiang became separated
primarily by the rise of the Tian Shan, which started about 7 mya
and diverted moisture-bearing westerly winds (Sun et al., 2004),
which together with the uplift of the Qinghai–Tibetan Plateau dur-
ing the Tertiary and Quaternary that prevented the flow of humid
Indian monsoons from passing through the Himalayas into the ba-
sin, resulted in increasing aridity in the Tarim basin. (Ren, 1980;
Shi et al., 1998; Sun and Liu, 2006; Sun et al., 2008; Zhu et al.,
1980). Already at 5.3 mya large areas were converted to shifting-
sand desert. The climatic shift from near-continuous warm condi-
tion to frequent and chaotic fluctuations afterwards occurred
approximately 2.5 mya (Zhongli et al., 1992). Major glaciations oc-
curred between 1170–800 kya, 720–500 kya, 300–130 kya and 70–
10 kya (Zheng et al., 2002). During interglacial times, meltwater
rivers from glaciers in the Tianshan and Kunlun mountains served
as the dominant source of water in the Tarim basin. At times of
maximum glaciation the climate was colder and drier, and water
supply diminished, causing northern and eastern portions of the
Tarim River to dry up. In eastern Kazakhstan, the present-type for-
est-steppe/semi-desert dominated during warm stages, and peri-
glacial steppe during cold stages (Aubekerov and Gorbunov,
1999). For both blythi clade 1a, halimodendri clade 2a and minula,
the initiation of population expansions are inferred to have
occurred more or less at the same time for all three clades, approx-
imately 90 kya. This indicates that the warmer climate during the
MIS5 (Ogg et al., 2008) interglacial between 130 and 70 kya may
have produced similar favourable conditions for all three
populations.
4.4.1. Clade 1 – blythi
This clade consists of two recently divided subpopulations
(Fig. 3), clade 1a showing signs of having undergone a recent pop-
ulation expansion, while clade 1b may have remained more stable.
All evidence suggest that clade 1a inhabits the Siberian taiga, and a
post-glacial population expansion into this area seems to be a rea-
sonable assumption. However, the onset of the expansion was esti-
mated at 89 kya, which predates the LGM and coincides with the
MIS5 interglacial. If the current ranges of clades 1a and 1b differ,
one possible interpretation of the different signals from the demo-
graphic indicators is that clade 1b has remained in demographic
equilibrium closer to a refugium. Clarifying the breeding range of
clade 1b would therefore be of great interest, as it could indicate
a region where the impact of Pleistocene climate fluctuations has
been less severe.
4.4.2. Clade 2 – halimodendri
The cyt bdata indicates possible population expansion in this
clade, but all the support for population expansion comes from
clade 2a. For clade 2a, the cyt bdata strongly suggest either a re-
cent population expansion or a bottleneck. For clade 2b, no devia-
tion from the null hypothesis is suggested. For the nuclear markers,
clades 2a and 2b were analysed combined. For these loci, the
demographic evidence is contradictory, but more in favour of pop-
ulation expansion or bottleneck than a stable population. A possi-
ble interpretation of these results is similar to that for clade 1b:
that clade 2b has remained in demographic equilibrium, while
clade 2a has expanded, perhaps colonising new areas or recovering
after a bottleneck. Also in this case, clarifying the breeding range of
clade 2b would potentially indicate a region less influenced by
Pleistocene climate fluctuations.
4.4.3. Clade 3 – margelanica
The evidence for or against demographic equilibrium are rather
weak for this clade, and somewhat contradictory. Fu’s F
S
suggest
departure from demographic equilibrium, but this is not supported
by R
2
. The raggedness index, r, for the ODC data indicates departure
from the null hypothesis of population expansion. In conclusion
this clade may represent a population that has remained relatively
stable. It should be investigated whether the ranges of margelanica
and clade 2b are similar, and whether there is a correlation be-
tween their demography and historical distribution.
4.4.4. Clade 4 – althaea
The demographic indicators for althaea suggest a more or less
stable population. Neither Fu’s F
S
or R
2
provide strong support for
departure from demographic equilibrium, and the raggedness in-
dex, r, for the myo data also indicates a stable population.
4.4.5. Clade 5 – curruca
The cyt bdata for this clade is ambiguous. Relatively strong sup-
port for departure from demographic equilibrium is provided by
R
2
, at the same time as the raggedness index indicates departure
from population expansion. The nuclear data show weak support
for population expansion. Generally, R
2
may be regarded as the
most reliable indicator for small samples, whereas Fu’s F
S
is better
for large sample sizes. The raggedness is the weakest of these tests
(Ramos-Onsins and Rozas, 2002). In conclusion, the evidence point
toward a recent population expansion or bottleneck, rather than a
stable population. Much of the area currently inhabited by this tax-
on was severely affected by the most recent glacial maximum, and
in recolonising the area both population expansion and recovery
from a population bottleneck are reasonable assumptions.
4.4.6. Clade 6 – minula
For the cyt bdata, Fu’s F
S
or R
2
provide strong support for depar-
ture from demographic equilibrium. There is also some support for
a population expansion or bottleneck in the BRM data, but both
myo and ODC show very weak support for this. This taxon inhabits
an area geographically delimited by mountain ranges, in suitable
surrounding habitats replaced by other taxa in this complex. There
are no indications that it has invaded this area from elsewhere re-
cently, and demographic events may thus have occurred within the
same range. Studies of reptiles and mammals have indicated that
southwestern Tarim Basin functioned as a glacial refugium during
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 83

times of maximum glaciation (Shan et al., 2011; Zhang et al., 2010).
For minula, initiation of population expansion appears to coincide
with the interglacial between 130 and 70 kya.
5. Conclusions
The Lesser Whitethroat complex emerges as a suitable model
both for studies of cryptic speciation and effects of climate fluctu-
ations. Our study does not resolve the inclusiveness and detailed
definition of ranges of the different taxa, and these remain to be
determined, as do species limits. The morphological differentiation
is slight and contradicts phylogenetic evidence to some extent; a
thorough review of the morphological differences of all taxa based
on the findings of this study is needed. Future studies are needed to
evaluate the correlation between demography and historical distri-
bution, amount of niche competition, interaction and selection
against hybrids in contact zones between these taxa, as well as
the roles and relative importance of gene flow and parallel evolu-
tion in shaping the present morphological patterns.
Acknowledgments
We are most grateful to the following persons and institutions
for samples: Robert Pry
ˆs-Jones and Mark Adams and the Natural
History Museum, Tring, UK; Magdalena Bus
´-Kicman and Wieslaw
Bogdanowicz and the Museum and Institute of Zoology, Polish
Academy of Sciences, Warszawa, Poland; Ulf Johansson and the
Swedish Museum of Natural History, Stockholm, Sweden; Jon
Fjeldså and Jan Bolding Kristensen and the Zoological Museum of
the University of Copenhagen, Copenhagen, Denmark; Silke Fregin
and Martin Haase and Vogelwarte Hiddensee, Zoological Institute
and Museum, Ernst Moritz Arndt University of Greifswald, Greifs-
wald, Germany. Ma Ming, Ying Hak King, Barry Williams and John
Allcock assisted us in the field. We are also grateful to Björn Ander-
son, Annika Forsten, Edward Gavrilov, Magnus Gelang, Darren Ir-
win, Knud Jønsson, Peter Kennerley, Antero Lindholm, Ross
McGregor, Ulf Ottosson, Martin Stervander, Andrew Lassey, An-
drew Grieve, Roger Riddington, D. I. M. Wallace, Patrik Wildjang,
Stephen Votier and Reuven Yosef for providing samples. Hadoram
Shirihai and Nigel Redman at A&C Black generously allowed us to
use the map used in Figs. 1 and 2. We are most grateful to Jornvall
Foundation for financial support and the Chinese Academy of Sci-
ences Visiting Professorship for Senior International Scientists
(No. 2011T2S04) (both to P.A.). The Swedish Research Council pro-
vided financial support (Grant No. 621-2006-3194 to U.O.).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ympev.2012.12.
023.
References
Akaike, H., 1973. Information theory as an extension of the maximum likelihood
principle. In: Petrov, B.N., Csaki, F. (Eds.), Second International Symposium on
Information Theory. Akademiai Kiado, Budapest.
Allen, E.S., Omland, K.E., 2003. Novel intron phylogeny supports plumage
convergence in orioles (Icterus). Auk 120, 961–969.
Arctander, P., 1995. Comparison of a mitochondrial gene and a corresponding
nuclear pseudogene. Proc. R. Soc. Lond. B Biol. Sci. 262, 13–19.
Aubekerov, B., Gorbunov, A., 1999. Quaternary permafrost and mountain glaciation
in Kazakhstan. Permafrost Periglac. Process. 10, 65–80.
Borge, T., Webster, M.T., Andersson, G., Saetre., G.P., 2005. Contrasting patterns of
polymorphism and divergence on the Z chromosome and autosomes in two
Ficedula flycatcher species. Genetics 171, 1861–1873.
Cracraft, J., 1989. Speciation and its ontology: the empirical consequences of
alternative species concepts for understanding patterns and processes of
differentiation. In: Otte, D., Endler, J.A. (Eds.), Speciation and its
Consequences. Sinauer Associates, Sunderland, MA, pp. 28–59.
Cramp, S., 1992. The Birds of the Western Palearctic, vol. 7. Oxford University Press,
Oxford.
de Queiroz, K., 2005. A unified concept of species and its consequences for the
future of taxonomy. Proc. Calif. Acad. Sci. 56, 196–215.
del Hoyo, J., Elliot, A., Christie, D.A. (Eds.), 2006. Handbook of the Birds of the World:
Old World Flycatchers to Old World Warblers, vol. 11. Lynx Edicions, Barcelona.
Dementiev, G.P., Gladkov, N.A. (Eds.), 1954. Birds of the Soviet Union. Israel Program
for Scientific Translations, Jerusalem, English translation 1968.
Dickinson, E.C. (Ed.), 2003. The Howard and Moore Complete Checklist of the Birds
of the World, third ed. Christopher Helm, London.
Drummond, A.J., Suchard, M.A., 2012. BEAST. Version 1.7.0. <http://beast.bio.
ed.ac.uk>.
Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics
and dating with confidence. PLoS Biol. 4, e88.
Drummond, A.J., Suchard, M.A., Xie, D., Rambaut, A., 2012a. Bayesian phylogenetics
with BEAUti and the BEAST 1.7 Mol. Biol. Evol. 29, 1969–1973.
Drummond, A.J., Rambaut, A., Suchard, M.A., Xie, W., 2012b. BEAUti. Version 1.7.0.
<http://beast.bio.ed.ac.uk>.
Fregin, S., Haase, M., Olsson, U., Alström, P., 2012. Pitfalls in comparisons of genetic
distances: a case study of the avian family Acrocephalidae. Mol. Phylogenet.
Evol. 62, 319–328.
Friesen, V.L., Congdon, B.C., Kidd, M.G., Birt, T.P., 1999. Polymerase chain reaction
(PCR) primers for the amplification of five nuclear introns in vertebrates. Mol.
Ecol. 8, 2147–2149.
Fu, Y.-X., 1997. Statistical tests of neutrality against population growth, hitchhiking
and background selection. Genetics 147, 915–925.
Graham, S.A., Chamberlain, C.P., Yue, Y., Ritts, B.D., Hanson, A.D., Horton, T.W.,
Waldbauer, J.R., Poage, M.A., Feng, X., 2005. Stable isotope records of Cenozoic
climate and topography, Tibetan plateau and Tarim basin. American Journal of
Science Online. 305 (2), 101–118.
Gu, X., Fu, Y.-X., Li, W.-H., 1995. Maximum likelihood estimation of the
heterogeneity of substitution rate among nucleotide sites. Mol. Biol. Evol. 12,
546–557.
Guo, X.G., Wang, Y.Z., 2007. Partitioned Bayesian analyses, dispersal-vicariance
analysis, and the biogeography of Chinese toad-headed lizards (Agamidae:
Phrynocephalus): a re-evaluation. Mol. Phylogenet. Evol. 45, 643–662.
Harpending, H.C., 1994. Signature of ancient population growth in a low-resolution
mitochondrial DNA mismatch distribution. Hum. Biol. 66, 591–600.
Hasegawa, M., Kishino, K., Yano, T., 1985. Dating the human-ape splitting by a
molecular clock of mitochondrial DNA. J. Mol. Evol. 22, 160–174.
Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny.
Bioinformatics. 17, 754–755.
Huelsenbeck, J.P., Ronquist, F., 2005. MRBAYES version 3.1.2. Available from: URL
http://mrbayes.scs.fsu.edu/download.php.
ICZN, 1976. International commission on zoological nomenclature announcement
A.N. (S) 98. Bull. Zool. Nom. 32 (2), 103.
Irestedt, M., Ohlson, J.I., Zuccon, D., Källersjö, M., Ericson, P.G.P., 2006. Nuclear DNA
from old collections of avian study skins reveals the evolutionary history of the
Old World suboscines (Aves, Passeriformes). Zool. Scr. 35, 567–580.
Kass, R.E., Raftery, A.E., 1995. Bayes factors. J. Am. Stat. Assoc. 90, 773–795.
Korelov, M.N., 1972. Genus Sylvia (in Russian). In: Korelov, M.N., Kovshar, A.F. (Eds.),
Birds of Kazakhstan (Ptitsy Kazakhstana), vol. 4. Alma-Ata, pp. 153–205.
Librado, P., Rozas, J., 2009. DnaSp v5: a software for comprehensive analysis of DNA
polymorphism data. Bioinformatics 25, 1451–1452.
Loskot, V.M., 2001a. Taxonomic revision of the Hume’s Whitethroat Sylvia althaea
Hume, 1878. Avian. Ecol. Behav. 6, 41–42.
Loskot, V.M., 2001b. Name-bearing types of subspecies of Sylvia curruca (L.) and S.
althaea Hume in the collection of the Zoological Institute, St. Petersburg (Aves:
Sylviidae). Zoosyst. Rossica 10 (1), 219–229.
Loskot, V.M., 2005. Morphological variation and taxonomic revision of five south-
eastern subspecies of Lesser Whitethroat Sylvia curruca (L.) (Aves: Sylviidae).
Zool. Mededelingen, Leiden 79–3, 157–165.
Martens, J., Steil, B., 1997. Reviergesänge und Speziesdifferenzierung in der
Klappergrasmücken-Gruppe Sylvia [curruca]. J. Ornithol. 138, 1–23.
Mayr, E., 1942. Systematics and the Origin of Species. Harvard University Press,
Cambridge, MA.
Mayr, E., 1963. Animal Species and Evolution. Harvard University Press, Cambridge,
MA.
Mayr, E., Cottrell, G.W. (Eds.), 1986. Checklist to the Birds of the World, vol. XI.
Museum of Comparative Zoology, Cambridge, MA.
Nei, M., 1987. Molecular Evolutionary Genetics. Columbia University Press, New
York.
Newton, M.A., Raftery, A.E., 1994. Approximate Bayesian inference with the
weighted likelihood bootstrap. J. Roy. Statist. Soc. B 56, 3–48.
Nylander, J.A.A., 2004. MrModeltest v2: Program Distributed by the Author.
Evolutionary Biology Centre, Uppsala University, <www.ebc.uu.se/systzoo/
staff/nylander.html>.
Ogg, J.M., Ogg, G., Gradstein, F.M., 2008. The Concise Geologic Time Scale.
Cambridge Univ. Press.
Olsson, U., Alström, P., Ericson, P.G., Sundberg, P., 2005. Non-monophyletic taxa and
cryptic species—evidence from a molecular phylogeny of leaf-warblers
(Phylloscopus, Aves). Mol. Phylogenet. Evol. 36, 261–276.
Olsson, U., Alström, P., Svensson, L., Aliabadian, M., Sundberg, P., 2010. The Lanius
excubitor (Aves, Passeriformes) conundrum – taxonomic dilemma when
84 U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85

molecular and non-molecular data tell different stories. Mol. Phylogenet. Evol.
55, 347–357.
Portenko, L.A., 1960. The birds of the USSR. Part IV. Acad. Sci. USSR 69, 1–413 (In
Russian).
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution.
Bioinformatics 14, 817–818.
Rambaut, A., 2009. FigTree. Version 1.3.1. <http://tree.bio.ed.ac.uk/software/
figtree>.
Rambaut, A., Drummond, A.J., 2009. Tracer. Version 1.5. <http://beast.bio.ed.ac.uk>.
Rambaut, A., Drummond, A.J., 2012. TreeAnnotator. Version 1.7.0. <http://
beast.bio.ed.ac.uk>.
Ramos-Onsins, S.E., Rozas, J., 2002. Statistical properties of new neutrality tests
against population growth. Mol. Biol. Evol. 19, 2092–2100.
Ren, M., 1980. An Outline of the Physical Geography in China. Commercial
Publisher, Beijing.
Ripley, S.D., 1982. A synopsis of the birds of India and Pakistan together with those
of Nepal, Bhutan, Bangladesh and Sri Lanka: i-xxvi, 1-653.— Bombay.
Rogers, A.R., Harpending, H.C., 1992. Population growth makes waves in the
distribution of pairwise genetic differences. Mol. Biol. Evol. 9, 552–569.
Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference
under mixed models. Bioinformatics 19, 1572–1574.
Sæther, B.-E., Lande, R., Engen, S., Weimerskirch, H., Lillegård, M., Altwegg, R.,
Becker, P.H., Bregnballe, T., Brommer, J.E., McCleery, R.H., Merilä, J., Nyholm, E.,
Rendell, W., Robertson, R.R., Tryjanowski, P., Visser, M.E., 2005. Generation time
and temporal scaling of bird population dynamics. Nature 436, 99–102.
Shan, W.-J., Liu, J., Murphy, R.W., Halik, M., Zhang, Y.-P., 2011. Genetic consequences
of postglacial colonization by the endemic Yarkand hare (Lepus yarkandensis)of
the arid Tarim Basin. Chinese Sci. Bull. 2011 (56), 1370–1382.
Shi, Y., Li, J., Li, B., 1998. Uplift and Environmental Changes of Qinghai-Xizang
(Tibetan) Plateau in the Late Cenozoic. Guangdong Science & Technology Press,
Guangzhou.
Shirihai, H., Gargallo, G., Helbig, A., Harris, A., Cottridge, D., Roselaar, D., 2001. Sylvia
Warblers: Identification, Taxonomy and Phylogeny of the Genus Sylvia.
Christopher Helm, A & C Black, London.
Slatkin, M., Hudson, R.R., 1991. Pairwise comparisons of mitochondrial DNA
sequences in stable and exponentially growing populations. Genetics 129,
555–562.
Snigirewski, S., 1929. Übersicht der Formen von Sylvia curruca. J. Orn 77 (2), 252–
261.
Sorenson, M.D., Quinn, T.W., 1998. Numts: a challenge for avian systematics and
population biology. Auk 115, 214–221.
Stepanyan, L.S., 1983. Superspecies and sibling species in avifauna of the USSR
(Nadvidy i vidy-dvoyniki v avifaune SSSR). Moscow, pp. 1–194 (in Russian).
Stolzmann, J. 1897. Oiseaux de la Ferghana. Bull. Soc. Imp. des Naturalistes de
Moscou. New ser., vol. 11, pp. 70–72.
Sun, J.M., Liu, T.S., 2006. The age of the Taklimakan Desert. Science 312, 1621.
Sun, J., Zhu, R., Bowler, J., 2004. Timing of the Tianshan Mountains uplift constrained
by magnetostratigraphic analysis of molasse deposits. Earth Planet. Sci. Lett.
219, 239–253.
Sun, J.M., Zhang, L.Y., Deng, C.L., Zhu, R., 2008. Evidence for enhanced aridity in the
Tarim Basin of China since 5.3Ma. Quat. Sci. Rev. 27, 1012–1023.
Svensson, L., 1992. Identification Guide to European Passerines, fourth ed.
Stockholm.
Tajima, F., 1989. Statistical method for testing the neutral mutation hypothesis by
DNA polymorphism. Genetics 123, 585–595.
Ticehurst, C.B., Whistler, H., 1933. What is Curruca affinis Blyth? Ibis 75, 554–556.
Vaurie, C., 1954. Systematic notes on Palearctic birds, No 11. Sylviinae: the genus
Sylvia. Am. Mus. Novit. 1692, 1–17.
Vaurie, C., 1959. The Birds of the Palearctic Fauna: Order Passeriformes. H.F. and G.
Witherby, London.
Weir, J.T., Schluter, D., 2008. Calibrating the avian molecular clock. Mol. Ecol. 17,
2321–2328.
Williamson, K. 1976. Identification for Ringers 3: The Genus Sylvia, Second
Reprinted ed. British Trust for Ornithology, Tring.
Yang, Z., 1994. Maximum likelihood phylogenetic estimation from DNA sequences
with variable rates over sites: approximate methods. J. Mol. Evol. 39, 306–314.
Zhang, Q., Xia, L., He, J., Wu, Y., Fu, J., Yang, Q., 2010. Comparison of phylogeographic
structure and population history of two Phrynocephalus species in the Tarim
Basin and adjacent areas. Mol. Phylogenet. Evol. 57, 1091–1104.
Zheng, B., Xu, Q., Shen, Y., 2002. The relationship between climate change and
Quaternary glacial cycles on the Qinghai-Tibetan Plateau: review and
speculation. Quatern. Int. 97–98, 93–101.
Zhongli, D., Rutter, N., Han, J., Liu, T., 1992. A coupled environmental system formed
at about 2.5 Ma in East Asia. Palaeogeography, Palaeoclimatology, Palaeoecology.
94 (1–4), 223–242.
Zhu, Z., Wu Z., Liu, S., 1980. An Outline of Chinese Deserts. Science Press, Beijing.
Zink, R.M., Baraclough, G.F., 2008. Mitochondrial DNA under siege in avian
phylogeography. Mol. Ecol. 17, 2107–2121.
U. Olsson et al. / Molecular Phylogenetics and Evolution 67 (2013) 72–85 85