Postglacial Colonization of the Tibetan Plateau Inferred from the Matrilineal Genetic Structure of the Endemic Red-Necked Snow F

Molecular Ecology (2005) 14, 1767–1781 doi: 10.1111/j.1365-294X.2005.02528.x

BlackwellPostglacial Publishing, Ltd. colonization of the Tibetan plateau inferred from the matrilineal genetic structure of the endemic red-necked snow finch, Pyrgilauda ruficollis

YAN HUA QU,* PER G. P. ERICSON,† FU MIN LEI* and SHOU HSIEN LI‡ *Institute of Zoology, Chinese Academy of Sciences, 25 Beisihuanxi Road, Haidian District, Beijing 100080, People’s Republic of China, †Department of Vertebrate Zoology, Swedish Museum of Natural History, PO Box 50007, SE-10405 Stockholm, Sweden, ‡Department of Life Sciences, National Taiwan Normal University, Taiwan

Abstract Most phylogeographical studies of postglacial colonization focus on high latitude locations in the Northern Hemisphere. Here, we studied the phylogeographical structure of the red-necked snow finch Pyrgilauda ruficollis, an endemic species of the Tibetan plateau. We analysed 879 base pairs (bp) of the mitochondrial cytochrome b gene and 529 bp of the control region in 41 birds from four regional groups separated by mountain ranges. We detected 34 haplotypes, 31 of which occurred in a single individual and only three of which were shared among sampling sites within regional groups or among regional groups. Haplotype diversity was high (h = 0.94); nucleotide diversity was low (d = 0.00415) and genetic differentiation was virtually non-existent. Analyses of mismatch distributions and geographi- cally nested clades yielded results consistent with contiguous range expansion, and the expansion times were estimated as 0.07–0.19 million years ago (Ma). Our results suggest that P. ruficollis colonized the Tibetan plateau after the extensive glacial period (0.5–0.175 Ma), expanding from the eastern margin towards the inner plateau. Thus, in contrast to many of the postglacial phylogeographical structures known at high latitudes, this colonization occurred without matrilineal population structuring. This might be due to the short glacial cycles typical of the Tibetan plateau, adaptation of P. ruficollis to cold conditions, or refugia and colonized habitat being semicontinuous and thus promoting population mixing. Keywords: genetic structure, Pleistocene glaciations, postglacial colonization, Pyrgilauda ruficollis Received 5 September 2004; revision received 19 November 2004; accepted 7 February 2005

of phylogeographical structures are available for regions at Introduction different latitudes, studies that focus on previously glaciated Postglacial colonization has created a variety of phylogeo- montane areas are rare. Herein we present a study of a species graphical structures in species from different latitudes endemic to the Tibetan plateau. (Rising & Avise 1993; Hewitt 1996; Merila et al. 1997). Previ- The Tibetan plateau occupies an area of 2.5 million km2, or ously glaciated areas in the Arctic and the sub-Arctic regions approximately one-quarter of China, and has an average contain species with low levels of clade divergence, indicat- altitude of 4500 m above sea level (a.s.l.). It is the youngest ing recent colonization followed by population expansion. plateau on Earth; the most recent uplift event occurring In Europe and North America, such areas contain species between 3.6 and 1.7 million years ago (Ma) (Li & Zhou 1998). with intermediate clade divergences, indicating their sur- The uplift caused great climatic changes: grasslands replaced vival during several ice ages. In the tropics, this area contains forests while the climate gradually became drier, colder and species with deeply diverged clades, often within small geo- windier, and glaciers and deserts developed (Wu et al. 2001). graphical areas, indicating their survival there since the The unique geomorphological configuration, the complex land Pliocene (Hewitt 2000, 2004). Whereas many comparisons conditions, the diversified climate, and the unique geological evolution combine to make the Tibetan plateau an area of world- Correspondence: Yan Hua Qu, Fax: 0086 10 62565689; E-mail: wide importance for the evolution of endemic, specialized [email protected] montane species (Cheng 1981; Tang 1996; Macey et al. 1998).

1768 Y. H. QU ET AL.

Glacial cycles in alpine regions have generated varied to avoid sampling relatives (Hansen et al. 1997). Blood or phylogeographical structures that reflect different routes tissue samples were obtained from 43 birds. Groups with of postglacial colonization (Mardulyn 2001; Despres et al. adequate sample sizes were created by pooling birds into 2002; Kropf et al. 2003). The topographical diversity of the four regional groups: QR (Qinghai region, average altitude Tibetan plateau might have created both networks of refugia 4000 m a.s.l.), TR (Tanggulashan region, average altitude during glaciation and complex barriers to subsequent expan- 5500 m a.s.l.), WTR (west Tibet region, average altitude 4800 m sion (Hewitt 2004). In comparison to species that colonized a.s.l.), and ETR (east Tibet region, average altitude 4500 m their present-day ranges from lower latitudes, montane a.s.l.) (Fig. 1b, c and Table 1). species undergoing postglacial colonization would have needed to undertake altitudinal shifts, and might have been DNA extraction, polymerase chain reaction and able to spread more widely across tundra and steppe plains. sequencing In this study, we assessed these possibilities by conducting a phylogeographical study of an alpine bird endemic to Genomic DNA was extracted from blood or tissue samples the Tibetan plateau, the red-necked snow finch Pyrgilauda using the QIAamp DNA Mini Kit (QIAGEN) following manu- ruficollis. facturer’s instructions. Initially, 879 bp of the cytochrome b Pyrgilauda ruficollis is one of the four species of the genus gene was amplified as a single fragment with the primer pair Pyrgilauda (Eck 1996), three of which (Pyrgilauda ruficollis, L14841 (5′-CCATCCAACATCTCAGCATGATGAAA-3′) Pyrgilauda blanfordi and Pyrgilauda davidiana) have similar (Kocher et al. 1989) and H15915 (5′-AACTGCAGTCATCT- ranges in the Tibetan plateau (Qu et al. 2002). This species CCGGTTTACAAGAC-3′) (Edwards & Wilson 1990). The is a year-round resident across the mountain steppe zone thermocycling program consisted of an initial denaturation at altitudes of 3500–5300 m a.s.l., or higher (Fig. 1a), where at 94 °C for 5 min, followed by 40 cycles of 94 °C for 40 s, 49 °C it occurs in alpine meadows and breeds inside pika for 40 s, and 72 °C for 5 min. For the sequencing reactions, (Ochotona spp.) burrows. Pyrgilauda ruficollis makes irregular the following primers were used: L14841, H15915, P5L (5′- altitudinal movements, descending to lower altitudes in CCTTCCTCCACGAAACAGGCTCAAACAACCC-3′) and large flocks during autumn and winter when driven by H658 (5′-TCTTTGATGGAGTAGTAGGGGTGGAATGG-3′) extreme weather conditions (Cramp & Perrins 1994). The (Irestedt et al. 2002), with P5L and H658 as internal primers highest known records of P. ruficollis are from 5300 m a.s.l. on the light and heavy strands, respectively. in the Tanggula Mountains. Several mountain ranges, A 529-bp fragment of the control region was amplified some with peaks over 6500 m a.s.l., occur within the distribu- using the primer pair, F304 (5′-CTTGACACTGATGCAC- tion of P. ruficollis, and these might create barriers to gene TTG-3′) and H1261 (5′-AGGTACCATCTTGGCATCTTC- flow because they are believed to be major zoogeographi- 3′) (Marshall & Baker 1997). The thermocycling program cal barriers associated with evolutionary divergence consisted of an initial denaturation at 94 °C for 5 min, (Mayr 1963; Macey et al. 1998; Bos & Sites 2001; Roslin 2001; followed by 40 cycles of 94 °C for 40 s, 56 °C for 40 s, and Sorenson & Payne 2001). 72 °C for 5 min. The same primers were used for the sequen- Here, we assume that the present-day distribution of cing reactions. P. ruficollis stems directly from postglacial colonization. We The polymerase chain reaction (PCR) products were hypothesize that isolation in different refugia surrounding purified using QIAquickTM PCR purification Kit (QIAGEN), the Tibetan plateau led to phylogeographical divergence in and then sequenced on a Perkin-Elmer 377 semiautomated this species. We also hypothesize that the mountain ranges DNA sequencer (Applied BioSystems), using Perkin-Elmer within the present distribution range constitute barriers to Prism terminator cycle sequencing kits (Applied BioSystems) gene flow that have led to population differentiation. The with AmpliTaq FS polymerase with BigDye terminators. goal of the study is to test these hypotheses by describing the Both strands of each PCR product were sequenced. The phylogeographical and population structures of Pyrgilauda sequencing program consisted of 25 cycles of denaturation ruficollis in the Tibetan plateau and using these structures at 96 °C for 30 s, annealing at 50 °C for 15 s, and extension to infer evidence for population bottlenecks and expansion, at 60 °C for 4 min. and genetic divergence. Multiple sequence fragments were obtained by sequencing with different primers for each gene and individual. While a pair of internal primers (P5L and H658) of the cyto- Materials and methods chrome b was used to sequence approximately half of the gene (about 400 bp and 500 bp, respectively), other pair of Study area and sample collections primers (L14841 and H15915) obtained whole sequence. The birds were collected using mist nets from 10 sites covering No length variation in the control region was found, mak- major parts of the range of Pyrgilauda ruficollis. Each bird ing alignment straightforward. Complete sequences were within a site was taken from a different part of the colony assembled using seqman II (DNASTAR). Sequences were

POSTGLACIAL COLONIZATION OF RED-NECKED SNOW FINCH 1769

Fig. 1 The study area and locations of the four regional groups separated by mountain ranges. (a) The distribution records and range of Pyrgilauda ruficollis. (b) The sampling sites and regional groups. (c) The mountain ranges in the study area. Notes: Bayan Har mountains at 5300 m a.s.l. on average and A’nyemaqen mountains at 5500 m a.s.l. on average separate QR from other regional groups. Tanggula mountains with an average altitude of 6000 m a.s.l. separate TR from other regional groups. Gangdise-Nyaingentanglha mountains at an average altitude of 6000 m a.s.l. separate WTR from ETR.

1770 Y. H. QU ET AL.

Table 1 Details of sampled and sampling sites, and values of nucleotide divergence, diversity and haplotype diversity in regional groups

Regional Sampling UTM Haplotypes Nucleotide Nucleotide Haplotype groups sites coordinates present N divergence (d) diversity (3) diversity (h)

Qr (Qinghai region) Huashixia –1 194 125.22 25, 26, 29, 33, 34 5 0.0151 0.00708 0.959 4 041 084.6 Maduo –1 222 319.88 8 (4), 20, 21, 22, 30 5 4 100 988.26 Tianjun –1 105135.81 8 (4), 27, 28, 31 4 4 271 047.31 Heimahe –1 088 395.23 23 (4), 24 5 4 182 057.91 TR Tanggulashan –1 925 113.55 17, 18, 19 3 0.0046 0.00322 1.000 (Tanggulashan region) 3 895 250.63 ETR (East Tibet region) Bangda –1 434 149.14 5, 6 (2), 7, 32 5 0.0042 0.00322 0.973 3 504 700.64 Changdu –1 420 555.5 1, 2, 3, 4 4 3 614 299.4 WTR (West Dingri –2 512 295.09 8 (4), 14, 15 3 0.0064 0.00331 0.977 Tibet region) 3 536 135.94 Nanmulin –2 272 707.1 8 (4), 9, 10, 11 4 3 599 856.15 Langkazi –2 156 311.52 12, 13, 16 3 3 536 135.94

Three haplotypes (in bold) occurred in more than one bird from different regional groups (haplotype 8) and different sampling sites (haplotypes 6 and 23) (numbers of occurrence in italics). compared visually to the original chromatograms to avoid (r) between observed and expected mismatch distributions reading errors. Complete sequences were aligned by eye. were used as a test statistic and their P value represented All sequences are accessible at GenBank (Accession nos the probability of obtaining a simulated sum of squared AY825286–AY825329, AY961957–AY961980). deviation larger or equal to the one observed. Values of Tajima’s D (Tajima 1989) were calculated from the total number of segregating sites and used to assess Sequence variation and genetic diversity evidence for population expansion, under which negative Numbers of haplotypes (HT ), and values of haplotype values are expected (Aris-Brosou & Excoffier 1996). Esti- diversity (h) (Nei 1987), nucleotide diversity (3; Nei & mation and testing were done by bootstrap resampling Tajima 1981) and nucleotide divergence (d; Nei & Tajima (10 000 replicates) using arlequin 2.0. 1981) were computed using the package, dnasp (version The relationship τ = 2ut (Rogers & Harpending 1992) 4.0; Rozas et al. 2003). was used to estimate a time of expansion (t), where τ is the mode of the mismatch distribution, expressed in units of evolutionary time, and u is the mutation rate for the whole Hierarchical analysis of molecular variance sequence. The value u was calculated using the formula A hierarchical analysis of molecular variance (amova; u = µk, where µ is the mutation rate per nucleotide and k is Excoffier et al. 1992) was implemented using the arlequin the number of nucleotides assayed. A mutation rate of 2% version 2.0 package (Schneider et al. 1997). The F statistics were per million years (Myr) (µ = 2.0 × 10−8) was used, as a standard computed using haplotype frequencies alone, and the signi- evolutionary rate of mitochondrial DNA used in most studies ficance of departures from zero of F statistics and genetic of avian species. A generation time of 1.5 years was used in variance components was tested using 10 000 permutations. all calculations (Summers-Smith 1988).

Mismatch distribution analysis Coalescent-based estimation of gene flow Mismatch distributions were calculated using arlequin We used the program migrate version 1.7.6 (Beerli 1997) to 2.0, and their fit to Poisson distributions was assessed estimate maximum-likelihood migration rates among four by Monte Carlo simulations of 1000 random samples. The regional groups. This approach, based on coalescence using sum of squared deviations (SSD) and raggedness indexes Markov chain Monte Carlo (MCMC) searches, takes account

© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 1767–1781 POSTGLACIAL COLONIZATION OF RED-NECKED SNOW FINCH 1771 of unequally effective population sizes and asymmetrical The combined length of these sequences (1405 bp) contained gene flow (Beerli & Felsenstein 1999). Effective population 49 polymorphic sites, 27 of which were parsimony informative. sizes and gene flow rates were estimated from FST values The cytochrome b gene sequences contained 32 polymorphic and were set as initial values. We performed 10 short chains sites, 20 of which were parsimony informative, and 19 of (500 trees used out of 10 000 sampled) and three long chains which were characterized by T↔C transitions, nine by A↔G (5000 trees used out of 100 000 sampled). Adaptive heating transitions, two by both a transition and a transversion with four chains of different temperature (1, 3, 5, 8) was used. (C↔T↔G; A↔G↔C), one by T↔G transversion and one These runs were repeated using the same condition until by G↔C transversion. Seven variable amino acid positions consistent results were obtained. were detected in cytochrome b. The CR sequences contained 17 polymorphic sites, seven of which were parsimony informative, and 11 of which were characterized by T↔C transi- Nested design and nested cladistic analysis of geographical tions, two by A↔G transitions, two by T↔G transversion, distances one by A↔C transversion and one by C↔G transversion. The networking algorithm developed by Templeton et al. These polymorphic sites defined 34 unique haplotypes, (1992) was used to construct the intraspecific maximum 31 of which were observed in a single bird each and only parsimony phylogenetic relationship among haplotypes three of which were shared among different regional groups using the program tcs version 1.18 (Clement et al. 2000). or different sampling sites from same regional groups.

Clade distances (Dc) and nested clade distances (Dn) were Haplotype 8 was shared between the groups QR and WTR, defined based on the geographical locations of samples in and the sampling sites within each group. Two individuals the nesting cladogram, and were estimated as described in in ETR shared haplotype 6, and haplotype 23 was shared Templeton et al. (1995). The differences between interior by four individuals in QR. Within regional groups, haplotype

(ancestral) and tip (recent) clade Dc and Dn distances were diversity values were nearly maximal; nucleotide diversity − − calculated to yield DcI DcT and DnI DnT values, where was the highest in QR, whereas the other three groups had I and T were interior and tip clades, respectively. the similar values; pairwise nucleotide divergence values The null hypothesis of no geographical associations of tip were highest in QR and lowest in ETP (Table 1). The value clades and interior clades was tested by considering that of Tajima’s D test was −1.3254 in the group of all haplo- the dispersion distance of clades was not greater or less types combined (P > 0.1), suggesting that the observed than expected by chance, and comparing observed Dc and nucleotide polymorphism is selectively neutral. Dn values with a distribution of such values, calculated for each 10 000 random permutations of clades against sam- Hierarchical analysis of molecular variance pling locations (Templeton & Sing 1993; Templeton 1995). Permutation tests were conducted separately for each level Virtually all of the total genetic variance was located at of the nested cladogram using geodis version 2.2 (Posada the smallest geographical scale: among individuals within et al. 2000). As soon as significance levels for Dc and Dn sampling sites. The only remaining component that was were determined, inferences about the processes that were significantly greater than zero was located among sites likely to be responsible for observed patterns of clade within regional groups. The component located among the structure were made using the latest inference keys pro- regional groups was not significantly different from zero vided at http://darwin.uvigo.es (updated July 2004). (Table 2).

Results Mismatch distribution analysis The mismatch distributions for the three largest regional Sequence variation and genetic diversity groups and for the group containing all haplotypes com- Full-length DNA sequences of the cytochrome b gene and bined consisted of distinct unimodal curves (Fig. 2). The TR the control region (CR) were obtained for 41 of 43 birds. group was excluded from this analysis because of its low

Table 2 Hierarchical analysis of molecular variance for Pyrgilauda ruficollis

Source of variation Variance component Φ-statistics Variance explained (%) P Fixation indices

Among regions 0.0008 0.17 0.393 FCT = 0.0017 Among samples/within regions 0.0292 7.93 0.002 FSC = 0.0794 Within samples 0.454 91.9 0.0004 FST = 0.0810

Table 3 Mismatch distribution analysis Table 4 Estimates of gene flow (Nem) and theta between regional groups of Pyrgilauda ruficollis QR ETR WTR Whole data set

Values of 2Nm for each recipient group Parameters Source Theta S 39 14 19 49 θ group (2Neµ) QR TR ETR WTR 0 3.896 0.709 2.527 5.764 θ 53.74 40.149 2826.25 116.403 1 QR 0.01729 — 2.75e-16 0.78e-15 19.37 τ 9.66 5.154 2.789 4.375 TR 0.00201 1.504 — 15.73 1.22e-10 T (KY) 194.286 137.1 72.227 116.4 ETR 0.00835 15.86 1.88e-16 — 10.77 Goodness-of-fit test WTR 0.00384 1.52 2.09e-16 6.37 — SSD 0.0152 0.0045 0.043 0.005 P 0.206 0.976 0.125 0.44 N is the effective population size of females, µ is the mutation rate R 0.03 0.02 0.12 0.0118 e and m is the migration rate. P 0.15 0.97 0.11 0.259 Tajima’s D −0.2322 −0.86556 −1.445 −1.26813 P 0.427 0.2249 0.07 0.1027 loop of ambiguity (loop 1 in Fig. 4). Three major haplotype The parameters of the model of sudden expansion (Rogers & groups (haplogroups) were identified: 4.1, 4.2 and 4.3, all Harpending 1992) are presented as well as goodness-of-fit test to of which contained haplotypes from all four regional groups. the model; SSD, sum of squared deviations; r, raggedness indexes. Some phylogeographical structures within clades were Tajima’s (1989) D test values and their statistical significance are also θ θ detected (Table 5). Three associations between clade and given (S, number of polymorphic sites, 0, pre-expansion and 1, postexpansion population size, τ, time in number of generations, geographical location were significant (1-5, 1-18 and 2-12), elapsed since the sudden expansion episode). all at low nesting levels (one-step to two-step haplogroups). These patterns are consistent with contiguous range expansion. For three-step (3-2) and four-step (4-1 and 4-2) sample size. Both the variance (SSD) and raggedness index haplogroups, results were ambiguous because of inadequate (r) tests suggested that the curves did not significantly differ geographical sampling. Geographical associations were either from the distributions expected from a model of population due to long-distance colonization and past fragmenta- expansion. Tajima’s D statistics showed negative values tion in the scenario where P. ruficollis was absent in the for all three regional groups (Table 3). These results intermediate areas, or due to contiguous range expansion are consistent with the groups having undergone recent in the scenario where P. ruficollis was present in these areas. population expansion. The estimates of expansion time To discriminate the type of movement leading to this ranged from approximately 72 000–190 000 years. pattern, we performed mismatch distribution analysis and calculated SSD and raggedness index for haplogroups 3-2, 4-1 and 4-2. A model of demographic expansion was Coalescent-based estimate of gene flow statistically supported for these clades: P(SSDobs) values Significant levels of historical gene flow were detected for were 0.38, 0.86 and 0.33; raggedness indexes were 0.04, 0.05 4 of the 12 possible source-recipient relationships between and 0.02; and P(Ragobs) values were 0.77, 0.55 and 0.63, pairs of regional groups (Fig. 3 and Table 4). The gene flow respectively. A contingency test detected a significant estimates for the remaining eight possible relationships geographical association of haplotypes contained within were close to zero. The estimates of female effective popu- haplogroups 3-5, 4-2 and the total cladogram. lation sizes scaled by mutation rate were larger for QR and ETR, than TR and WTR (Table 4). Discussion The combination of low nucleotide diversity and high haplo- Nested cladistic analysis of mitochondrial DNA type diversity, and the shape of the mismatch distributions, haplotypes both suggest that Pyrgilauda ruficollis underwent a rapid range The nested clade network of haplotypes of Pyrgilauda expansion following a population bottleneck. Time estimates ruficollis was centred on haplotype 8, which occurred in derived from the mismatch distributions suggest that this each of four sites in two regional groups. Many of the other postglacial colonization occurred at 0.07–0.19 Ma, which is haplotypes were derivable from it by one or two muta- consistent with expansion occurring after the extensive tions, and the network contained several extinct or unsampled glacial period (0.5–0.175 Ma). The nested cladistic analysis haplotypes (Fig. 4). The most diverged pair of haplotypes suggests that the range expansion was contiguous with differed by six substitutions. The parsimony network of gradual movement, with no strong geographical differ- haplotypes was resolved, except for the presence of one entiation for any clade. amova and migrate indicate that

Table 5 The nested cladistic analysis of geographical distances for the mitochondrial DNA haplotypes of Pyrgilauda ruficollis. The haplotype designations are given at the top and are boxed together to reflect the one-step nested design given in Fig. 4

Zero-step One-step Two-step Three-step Four-step Five-step

Haplos Dc Dn Clades Dc Dn Clades Dc Dn Clades Dc Dn Clades Dc Dn Clades Dc Dn

20 1-1 0 319 1-2 0 320 I-T 0 0.21 2–1 319 536 3 1-3 0 451 15 1-4 0 565 60476

70S 525L 14 0 477 1-5 460 483 − I-T 0S 48L I-T 460 25 2-2 496 485 − I-T 176 51 3-1 495S 499 23 1-9 2-16 0 348

16 1-10 2-15 0 854 3-7 496 641L − − I-T 0 505 I-T 1 142S 4-1 552 565 32 1-38 30 1-39 2-7 319 324

34 1-35 2-8 0S 47S 3-4 207 13 31 2-9 0 277 I-T −160 91 22 19 1-29 315 304 11 1-28 548 561 17 I-T 233 256 26 1-31 2-10 0 476

2-11 464 479 3-5 460 458 4-3 391S 486 I-T 464 2 I-T 253 144 24 1-15 21 1-14 2-5 79 85

3-3 78S 640 5-1 28 1-17 2-6 0 66 I-T 79 18 33 1-36 29 1-26 2-14 3-6 0 522 10 1-40 0 257 18 1-41 0 192 4-2 498 507 I-T 0 −65 2-3 220 279 1 1-11 2-4 0 408 I-T 16 −23

8 648L 652L 40410 13 0 348

I-T 648L 272L 1-18 552L 552L 19 1-19 0 423 2 1-20 0 402

25 1-21 0 688 3-2 462S 486 S − 17 1-22 0 176 I-T 404L 123 12 1-27 0 347 2-12 473 472

I-T 552L 144L I-T 326 149

Tests determine whether the within-clade (Dc) or nested clade (Dn) geographical distances are significantly large (L) or significantly small (S) at the 0.05 (*), 0.01 (**) or 0.001 (***) levels. Where interior and tip clades are present, significance is also tested for the average difference between these two types of clades (I-T). Interior clades are in bold. Note that the number of steps indicated refers to the clades within the nested clades.

Fig. 2 Mismatch distributions for the three largest regional groups and for the entire sample. The histograms represent the observed frequencies of pairwise differences among haplotypes and the line shows the curve expected for a population that has expanded. (a) Mismatch distribution in the QR group. (b) Mismatch distribution in the ETR group. (c) Mismatch distribution in the WTR group. (d) Mismatch distribution in the entire sample. no significant genetic divergence exists among regional the Tibetan plateau. Three of the low-level clades provide groups separated by mountain ranges. the clearest evidence that range expansion was contiguous with gradual movement. The inadequate geographical sampling prevents resolution between long-distance colo- Inference of recent population expansion in Pyrgilauda nization, combined with population fragmentation and ruficollis gradual movement as the mechanism to explain range The levels of haplotype diversity in Pyrgilauda ruficollis expansion based on high-level clades. For clarifying this populations are higher than in most other avian species question, further work based on microsatellites will be from previously glaciated areas, whereas the values of continued. nucleotide diversity are lower (Table 6). This relationship between haplotype and nucleotide diversity is also evident Postglacial colonization of P. ruficollis populations from the unimodal mismatch distributions. Such a pattern is frequently attributed to population expansion, which Analyses of the cytochrome b sequences from four species enhances the retention of novel mutations (Watterson 1984; of the genus Pyrgilauda suggest that the speciation of Avise & Walker 1998) and creates an excess of haplotypes P. ruficollis occurred at 1.0 –1.5 Ma (Qu 2003), which is approxi- differing by one or a few mutations (Slatkin & Hudson 1991; mately contemporaneous with the uplift of the Tibetan Rogers & Harpending 1992). plateau. The plateau has since undergone four or five The analyses of matrilineal gene flow and of nested glaciations (Shi 2002; Zheng et al. 2002). The largest glacier clades provide evidence for range expansion at the scale of of the Tibetan plateau occurred in middle Pleistocene (0.5

© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 1767–1781 POSTGLACIAL COLONIZATION OF RED-NECKED SNOW FINCH 1775 . 2004 . 1990 . 2003 et al et al . 2004 . 1997 . 2003 . 2000 . 2000 . 2003 et al et al et al et al et al et al et al Zink Pitra Randi Tiedemann Merila Pavlova Godoy Moum & Arnason 2001 Tegelström Peck & Congdon 2004 Avise in the Tibetan plateau with values from

glaciation glaciations period/or several Quaternary cold periods Riss (2–13.5 kyr) glaciation (20 kyr) glaciations Last Pleistocene glaciation glacial maxima glaciations glaciation glacial cycle (12.5–17.5 kyr) glacial cycle (12.5–17.5 kyr) Pyrgilauda ruficollis 15 kyr Post-Pleistocene 200 kyr Last glacial 12 kyr 5–8 kyr Post-Pleistocene 12–14 kyr Pleistocene 100 kyr Last Pleistocene Expansion time10 kyr GlaciationsPleistocene Final 7.5–187 kyrPleistocene Final Reference northern groups Peninsula and European mainland other regions European and southern European southwestern Asia, northeastern Asia and southeastern Asia Western Eastern and tropic structure determine subdivision Phylogeographical structure Indo-Pacific and southwest Pacific Indo-Pacific ) h Haplotype diversity ( ) 3 0.00023–0.003 0.86 Two clades: southern and 0.0032 0.17 Two clades: Iberian Nucleotide diversity ( 0.0025–0.0047 0.76 Three clades: Europe, 0.0042–0.0066 0.68–0.89 Lack of geographical 3 , ND , tRNA’s b b b cyt cyt mtDNA, CR 0.008 0.76 Two clades: Sicily and cyt mtDNA, CR 0.0292 0.932 Two clades: Western, and mtDNA, CR 0.0007–0.0012mtDNA, CR 0.53 0.029 Three clades: Atlantic, 0.82–0.90 Two clades: Atlantic and Eurasia mtDNA, Eurasia mtDNA CR, Mid-latitude temperate zone Northern Europe mtDNA, CR 0.004–0.033Europe 0.6–1Europe mtDNA, CR Isolation by distance 0.00134 mtDNA CR, 10 kyr 0.612 Weichsel Two clades: northern Eurasia 0.0014–0.0022 0.86 Two clades: Eastern and Temperate and tropical regions Atlantic Ocean mtDNA, CR 0.0093–0.0198Sweden 0.81–0.97 Two clades mtDNA RFLP Pleistocene Too small sample size to Atlantic, Pacific, and Indian oceans Atlantic, Pacific, and Indian oceans Comparison of nucleotide diversity, haplotype phylogeographical pattern and expansion time estimates Yellow wagtail Motacilla flava Bluethroat Luscinia svecica Rock Patridge Alectoris graeca Eider duck Somateria mollissima Greenfinch Carduelis chloris Great bustard Otis tarda Table 6 birds from other glaciated areas. See References for details Species Study areas Marker Citrine wagtail Motacilla citreola Bearded vulture Gypaetus barbatus Razobill Alca torda Common guillemot Uria aalge Pied flycatch Ficedula hypoleuca Sooty tern Sterna fuscata Sooty tern Sterna fuscata

© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 1767–1781 1776 Y. H. QU ET AL. . 2004 et al . 1993 . 2000 . 2000 et al et al et al Present study Wenink Mila Fry & Zink 1998 Barrowclough Milot extensive (175 kyr) Pleistocene glaciations glacial maxima glaciations glaciations glaciation 70–190 kyr Pleisotcene Expansion time Glaciations Reference 90–350 kyr Later half 12.5 kyr Pleistocene 12 kyr Pleistocene 240–700 kyr Pleistocene 10 kyr Last Pleistocene Without phylogeographical divergence Phylogeographical structure Five clades: Alaska, west coast of North America, Gulf of Mexico, western Europe and Taymyr Peninsula phylogeographical structure between paraphyly and reciprocal monophyly northeastern, western and southern Mexico subdivision. Three clades: BCO, ALA and eastern set ) h : 0.952 : 0.47 b b Haplotype diversity ( 0.94 cr: 0.904 cyt cr: 0.45 cyt ) 3 : 0.00629 : b b Nucleotide diversity ( 0.00415 cr: 0.00372 cyt 0.0009–0.0022 cr: 0.0003–0.0066 cyt 0.005 0.776 Two clades: USA and b b b CR, cyt mtDNA, CR 0–0.0075mtDNA, CR, cyt 0.41 Three clades: Tibetan plateau mtDNA, North America mtDNA, CR 0.00095–0.008Western North 0.77America Arctic tundra of the Northern Intermediate stage of Hemisphere North America mtDNA, cyt North America mtDNA, CR 0.0172 0.38 East-west population

continued Species Study areas Marker Blanks indicate that data are unavailable. Red-necked snow finch Pyrgilauda ruficollis Song sparrow Melospiza melodia Blue grouse Dendragapus obsurus Dunlin Calidris alpina Table 6 Mac Gillivray’s warbler Oporornis tolmiei Yellow warble Dendroica petechia

Fig. 3 Gene flow connections among regional groups. The shadow area represented the range of ice cover during the maximum Pleistocene glaciation (data from 1: 4 000 000 digital elevation data developed by Institute of Geography, Chinese Academy of Sciences).

Myr). The extensive glacial advance continued until 0.17 margin of the Tibetan plateau, which was free of ice cover, Ma, after the penultimate glaciation (0.3–0.13 Ma) (Zhuo whereas ETR is located on the southeastern margin, which et al. 1998; Zhang et al. 2000; Shi 2002; Zheng et al. 2002). was covered by ice (Fig. 3; Li 1986; Shi et al. 1990; Shi 1996). Our estimates of time since population expansion are thus Some ice-free areas might have existed around QR and consistent with range expansion that occurred after the ETR and these could have provided suitable refugia extensive glacial period. (Cheng 1981; Tang 1996). Pikas (Ochotona spp.) migrated to The different estimates of time since expansion suggest ETR during the late Pleistocene (Li 1986). Abandoned bur- that these populations are still in genetic disequilibria. This rows of pikas provide the only nest site used by P. ruficollis inference is consistent with our analysis of matrilineal (Dementiev & Gladkov 1954; Cheng 1981; Fu 1998). It is gene flow, which identified QR and ETR as the source of therefore plausible that P. ruficollis existed in refugia gene flow to other regions and identified these groups as around the eastern margin of the Tibetan plateau during potential refugia during glacial advance. The larger esti- the glacial advance. From there, populations could have mates of female effective population size for these groups expanded towards the inner plateau after the retreat of also suggested that they were important historical sources the glaciers. of migrants. Geological evidence suggests that during the maximum Population structure in P. ruficollis glacial advance, an ice sheet covered an area five to seven times larger than it does today (Shi et al. 1990; Wu et al. We found that genetic differentiation among sampling sites 2001; Zheng et al. 2002). At that time, ice cover would have and regional groups was extremely low and that most of been permanent in the highest altitude and central regions the variance occurred within sampling sites. Consequently, of the Tibetan plateau (Shi et al. 1990; Shi 1996), and the we found no evidence that mountain ranges created barriers frequency of glaciers in the east was less than that in the to gene flow. This conclusion was supported by the nested west (Zhang et al. 2000). QR is located in the northeastern cladistic analysis, in which most of the haplotypes and

Fig. 4 Nested clades of Pyrgilauda ruficollis haplotypes. The zeros refer to unobserved haplotypes intermediate between observed haplotypes. One-step clades are indicated by ‘1-#’; two-step clades by ‘2-#’, three-step clades by ‘3-#’ and four-step clades by ‘4-#’, where # is the number assigned to the clades within each level. haplogroups showed no strong geographical divergence. Comparison of the phylogeographical structure of Extensive gene flow after postglacial colonization can P. ruficollis with that of avian species from other explain this pattern of low divergence. At equilibrium, glaciated areas gene flow of the order of a few individuals per generation is theoretically sufficient to prevent the genetic divergence We compared the phylogeographical structure of P. ruficollis (Hartl & Clark 1989). Nonetheless, our results suggest that with that of avian species from other glaciated areas P. ruficollis has undergone recent range expansion and is still (Table 6). This comparison is preliminary and needs to be in genetic disequilibrium. We therefore favour extensive confirmed using larger data sets. The phylogeographical gene flow occurred during a range expansion and an insuffi- structure of P. ruficollis is most similar to that of Arctic cient time for genetic differentiation as the best working birds, with some clear differences, and is very different hypothesis to explain the virtual lack of matrilineal genetic from the marked phylogeographical structure typical of structure in P. ruficollis. European, North American and tropical birds.

Whereas Arctic birds have shallow clades with significant structure of an endemic bird in the Tibetan plateau. Our genetic divergence, P. ruficollis has weak phylogeographical results suggest that this bird experienced rapid population structure and no trace of genetic divergence. The possible expansion after the retreat of extensive glaciers. However, reason might be the sampling at different geographical unlike the strong phylogeographical structures that are scale. While our sampling in the Tibetan plateau was at well known in temperate birds, postglacial colonization smaller geographical scale, most studies on Arctic avian has led to very weak phylogeographical structure and no species were done on the circum-Arctic range, which wider detectable genetic divergence. Potential explanations for geographical scale allows population subdivision and this result are short glacial cycles, adaptation to cold con- subspecies speciation (Wenink et al. 1993; Holder et al. 1999, ditions, and semicontinuous habitats and refugia. Overall, 2000). this study provided new evidence for the role of post- The influence of the ice ages on climate was long term glacial colonization in shaping the phylogeographical and continuous in temperate region (Hewitt 1996, 2004). structure of endemic species of the Tibetan plateau. Most temperate species experienced the postglacial expansion after the retreat of the Last Glacial Maximum (23–18 Acknowledgements bp). Phylogeographical studies reveal the survival of deep lineages, often in several glacial refugia, indicating survival Our sincere thanks to the following people for their help in obtain- of populations in these southern refugia over many ice ages. ing samples for this study: Zuo Hua Yin, Gang Wang, Hong Feng With repeated range changes, survival populations may Zhao, Qi Sen Yang and Jian li Lu. We thank Martin Irestedt, Pia Eldenäs and Elisabeth Köster for help in laboratory work. Authors pass through many such adaptations and reorganizations, thank Yohannes Elizabeth for the useful comments. The editors of which allowed their lineages to diverge and accumulate Molecular Ecology and three anonymous referees provided valu- genetic differences (Hewitt 1996, 2000, 2004). able comments on the manuscripts. The research was supported In contrast, the Tibetan plateau has been less affected by by NSFC 30170126, 30270182, as well as by the Swedish Research ice sheets than its highly glaciated neighbouring regions Council (grant no. 621- 2001-2773 to P.E.). during last two glacial cycles (Sharma & Ower 1996; Zheng et al. 2002). When the plateau uplifted to 4500 m a.s.l., the References cold conditions restricted glacier growth. The largest glacier development in the Tibetan plateau occurred during the Aris-Brosous S, Excoffier L (1996) The impact of population expansion middle Pleistocene (0.5 Myr). Glacial retreat has occurred and mutation rate heterogeneity DNA sequence polymorphism. 13 since 0.17 Ma (Zhang et al. 2000; Shi 2002; Zheng et al. 2002). Molecular Biology and Evolution, , 494–504. Avise JC, Walker C (1998) Pleistocene phylogeographic effects on Our results suggest that P. ruficollis populations expanded avian populations and the speciation process. Proceedings of following this retreat, and that the glacial period might have the Royal Society of London. Series B, Biological Sciences, 265, 457– been too short for genetic divergence among refugia to arise. 463. Moreover, P. ruficollis is an endemic species of the Avise JC, Nelson WS, Bowen BW, Walker D (2000) Phylogeography Tibetan plateau and is well adapted to this extreme en- of colonially nesting seabirds, with special reference to global vironment. Its high elevation meadow and steppe habitats matrilineal patterns in the sooty tern (Sterna fuscata). Molecular 9 were probably less fragmented during the glacial advance Ecology, , 1783–1792. Barrowclough GF, Groth JG, Mertz LA, Gutierrez RJ (2004) than at present. Comprehensive pollen analyses indicate Phylogeographic structure, gene flow and species status in blue that both alpine meadow and steppe extended in the east grouse (Dendragapus obscurus). Molecular Ecology, 13, 1911–1922. part of the Tibetan plateau during the ice ages (Kong & Du Beerli P (1997) migrate 0.7: documentation and program, part of LAMARC 1980; Ke & Sun 1992; Liu et al. 2002). It is plausible that the [WWW document]. URL http://evolution.genetics.washington. populations of this cold-tolerant bird were widespread in Edu/lamarc.html [accessed 1997 to January 2005]. these habitats, and that its distribution range was displaced Beerli P, Felsenstein J (1999) Maximum-likelihood estimation of eastward during glaciation. If so, haplotype composition migration rates and effective population numbers in two populations using a coalescent approach. Genetics, 152, 763–773. could have remained homogeneous during range shifts Bos DH, Sites JW Jr (2001) Phylogeography and conservation and demographic fluctuations. By contrast, the considerable genetics of the Columbia spotted frog (Rana luteiventris; levels of haplotype divergence in many birds in previously Amphibia, Ranidae). Molecular Ecology, 10, 1499–1513. glaciated regions of Europe and North America are generated Cheng TH (1981) A study of the fauna and evolution of terrestrial by isolation in different refugia during several glacial cycles vertebrates of the Tibet plateau. Memoirs of Beijing Natural History (Hewitt 1996, 2000, 2004). Museum, 9, 1–17. Clement M, Posada D, Crandall KA (2000) tcs: a computer program to estimate gene genealogies. Molecular Ecology, 9, 1657–1659. Conclusion Cramp S, Perrins CM (1994) The Birds of the Western Palearctic. Oxford University Press, Oxford & New York. The goal of this study was to test hypotheses about the Dementiev G, Gladkov NA (1954) Birds of the Soviet Union. Israel effects of postglacial colonization on the phylogeographical Program for Scientific Translations Jerusalem.

Despres L, Loriot S, Gaudeul M (2002) Geographic pattern of Li WH, Zhou XM (1998) Ecosystems of Qinghai-Xizang (Tibetan) Plateau genetic variation in the European globeflower Trollius europaeus and Approach for their Sustainable Management. Guangdong L. (Ranunculaceae) inferred from amplified fragment length Science and Technology Press, Guangzhou. polymorphism. Molecular Ecology, 11, 2337–2347. Liu J, Yu G, Chen X (2002) Paleoclimate simulation of 21 ka for the Eck S (1996) Die palaearktischen Vôgel — Geospezies und Bio- Tibetan plateau and Eastern Asia. Climate Dynamic, 19, 575–583. spezies. Zoologische Abhandlungen des Staatlichen Museums Fur Macey JR, Schulte JA, Larson A et al. (1998) Phylogenetic relation- Tierkunde Dresden, 49 (Suppl.) 46. ships of toads in the Bufo bufo species group from the eastern Edwards SV, Wilson AC (1990) Phylogenetically informative length escarpment of the Tibetan plateau: a case of vicariance and dis- polymorphism and sequence variability in mitochondrial DNA persal. Molecular Phylogenetics and Evolution, 9, 80–87. of Australian songbirds (Pomatostomus). Genetics, 126, 695–711. Mardulyn P (2001) Phylogeography of the Vosges Mountains popu- Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular lations of Gonioctena pallida (Coleoptera: Chrysomelidae): a nested variance inferred from metric distance among DNA haplotypes: clade analysis of mitochondrial DNA haplotypes. Molecular application to human mitochondrial DNA restriction data. Ecology, 10, 1751–1763. Genetics, 131, 479–491. Marshall HD, Baker AJ (1997) Structural conservation and vari- Fry AJ, Zink RM (1998) Geographic analysis of nucleotide diver- ation in the mitochondrial control region of fringilline finches sity and song sparrow (Aves: Emberizidae) population history. (Fringilla spp.) and the greenfinch (Carduelis chloris). Molecular Molecular Ecology, 7, 1303–1313. Biology and Evolution, 14, 173–184. Fu TS (1998) Fauna Sinica. Aves. Science Press, Beijing. Mayr E (1963) Animal Species and Evolution. Harvard University Godoy FA, Negro JJ, Hiraldo F, Donnazar JA (2004) Phylogeo- Press, Cambridge, Massachusetts. graphy, genetic structure and diversity in the endangered bearded Merila J, Bjorklund M, Baker AJ (1997) Historical demography and vulture (Gypaetus barbatus, L.) as revealed by mitochondrial present day population structure of the greenfinch, Carduelis DNA. Molecular Ecology, 13, 371–390. chloris-an analysis of mtDNA control region sequence. Evolu- Hansen MM, Nielsen EE, Mensberg KLD (1997) The problem of tion, 51, 946–956. sampling families rather than populations: relatedness among Mila B, Girman DJ, Kimura M, Smith TB (2000) Genetic evidence individuals in samples of juvenile brown trout Salmo trutta L. for the effect of a postglacial population expansion on the phylo- Molecular Ecology, 6, 469–474. geography of a North American songbird. Proceedings of the Royal Hartl DL, Clark AG (1989) Principles of Population Genetics, 3rd edn. Society of London. Series B, Biological Sciences, 267, 1033–1040. Sinauer Associates, Sunderland, Massachusetts. Milot E, Gibbs HL, Hobson K (2000) Phylogeography and genetic Hewitt GM (1996) Some genetic consequences of ice ages and their structure of northern populations of the yellow warbler (Dend- role in divergence and speciation. Biological Journal of the Linnean roica petechia). Molecular Ecology, 9, 667–681. Society, 58, 247–276. Moum T, Arnason E (2001) Genetic diversity and population history Hewitt GM (2000) The genetic legacy of the Quaternary ice ages. of two related seabird species based on mitochondrial DNA Nature, 405, 907–913. control region sequences. Molecular Ecology, 10, 2463–2478. Hewitt GM (2004) Genetic consequences of climatic oscillations in Nei M (1987) Molecular Evolutionary Genetics. Columbia University the Quaternary. Philosophical Transactions of the Royal Society of Press, New York. London. Series B, Biological Sciences, 359, 183–195. Nei M, Tajima F (1981) DNA polymorphism detectable by restric- Holder K, Montgomerie R, Friesen VL (1999) A test of the glacial tion endonuclease. Genetics, 97, 145–163. refugium hypothesis using patterns of mitochondrial and nuclear Pavlova A, Zink RM, Drovetski SV, Red’Kin Y, Rohwer S (2003) DNA sequence variation in rock ptarmigan (Lagopus mutus). Phylogeographic patterns in Motacilla flava and Motacilla citreola: Evolution, 53, 1936–1950. species limits and population history. Auk, 120, 744–458. Holder K, Montogomerie R, Friesen VL (2000) Glacial vicariance Peck DR, Congdon BC (2004) Reconciling historical processes and and historical biogeography of rock ptarmigan (Lagopus mutus) population structure in the sooty tern Sterna fuscata. Journal of in the Bering region. Molecular Ecology, 9, 1265–1278. Avian Biology, 35, 327–335. Irestedt M, Fjeldså J, Johansson US, Ericson PGP (2002) Systematic Pitra C, Lieckfeldt D, Alonso JC (2000) Population subdivision in relationships and biogeography of the Tracheophone suboscines Europe’s great bustard inferred from mitochondrial and nuclear (Aves: Passeriformes). Molecular Phylogenetics and Evolution, 23, DNA sequence variation. Molecular Ecology, 9, 1165–1170. 499–512. Posada D, Crandall KA, Templeton AR (2000) geodos: a program Ke MH, Sun JZ (1992) Paleoclimate and paleoenvironment of the for the cladistic nested analysis of the geographical distribution last glacial stage in Salawusu area of Inner Mongolia. Acta of genetic haplotypes. Molecular Ecology, 9, 487–488. Botanica Sinica, 34, 717–719. (In Chinese with English abstract). Qu YH (2003) Studies on phylogeography and distribution patterns of Kocher TD, Thomas WK, Meyer A et al. (1989) Dynamics of Montifringilla species. A Dissertation submitted to the Faculty of mitochondrial DNA evolution in animals: amplification and Institute of Zoology, Chinese Academy of Sciences, in press. sequencing with conserved primers. Proceedings of the National Qu YH, Lei FM, Yin ZH, Fabrizio-De Ritis S (2002) Distribution Academy of Sciences of the United States of America, 86, 6196–6200. patterns of snow finches (genus Montifringilla) in the Tibetan Kong ZC, Du NQ (1980) Vegetation and climate changes in the plateau of China. Avocetta, 26, 11–18. past 30 000–10 000 years in Beijing. Acta Botanica Sinica, 22, 331– Randi E, Tabarroni C, Rimondi S, Lucchini Y, Sfougaris A (2003) 338. (In Chinese with English abstract). Phylogeography of the rock partridge (Alectoris graeca). Molecu- Kropf M, Kandereit JW, Comes HP (2003) Differential cycles of lar Ecology, 12, 2201–2214. range contraction and expansion in European high mountain Rising JD, Avise JC (1993) An application of genealogical concord- plants during the late Quaternary: insight from Pritzelago alpina ance principles to the taxonomy and evolutionary history of the (L.) O. Kuntze (Brassicaceae). Molecular Ecology, 12, 931–949. sharp-tailed sparrow (Ammodramus caudacutus). Auk, 110, 844– Li JJ (1986) Glaciers of Xizang. Science Press, Beijing. 856.

Rogers AR, Harpending H (1992) Population growth makes ping or DNA sequencing V. Analysis of case control sampling waves in the distribution of pairwise genetic differences. Mole- designs: Alzheimer’s disease and the apoprotein E locus. Genetics, cular Biology and Evolution, 9, 552–569. 140, 403–409. Roslin T (2001) Spatial population structure in a patchily distrib- Templeton AR, Sing CF (1993) A cladistic analysis of phenotypic uted beetle. Molecular Ecology, 10, 823–837. associations with haplotypes inferred from restriction endo- Rozas J, Sánchez-De I, Barrio JC, Messeguer X, Rozas R (2003) nuclease mapping IV. Nested analyses with cladogram DnaSP, DNA polymorphism analyses by the coalescent and uncertainly and recombination. Genetics, 134, 659–669. other methods. Bioinformatics, 19, 2496–2497. Templeton AR, Crandall KA, Sing CF (1992) A cladistic analysis of Schneider S, Kueffer JM, Poessli D (1997) ARLEQUIN, Version 1.1: a phenotypic associations with haplotypes inferred from restriction software for population genetic data analysis. Genetics and Biometry endonuclease mapping and DNA sequences data III. Cladogram Laboratory, Department of Anthropology University of Geneva, estimation. Genetics, 132, 619–633. Switzerland. Templeton AR, Routman E, Phillips CA (1995) Separating popu- Sharma MC, Owen LA (1996) Quaternary glacial history of NW lation structure from population history: a cladistic analysis of Garhwal, central Himalayas. Quaternary Sciences Reviews, 15, geographic distribution of mitochondrial DNA haplotypes in the 335–365. tiger salamander, Ambystoma tigrinum. Genetics, 140, 767–782. Shi YF (1996) Timing and height of Qinghai–Xizang plateau uplift- Tiedemann R, Paulus KB, Scheer M, Von Kistoweki KG, Skirnisson K ing into the cryosphere and its impact on the surrounding areas. (2004) Mitochondrial DNA and microsatellite variation in the In: Study of Evolutionary Process, Environmental Change and Ecological eider duck (Somateria mollissima) indicate stepwise postglacial System of Qinghai–Xizang (Tibet) Plateau (ed. Expert committee of the colonization of Europe and limited current long-distance dis- Tibet project), pp. 136–146. Science Press, Beijing. (In Chinese). persal. Molecular Ecology, 13, 1481–1494. Shi YF (2002) Characteristics of late Quaternary monsoonal glaciation Watterson GA (1984) Allele frequencies after a bottleneck. Theoret- on the Tibetan plateau and in East Asia. Quaternary Internation, ical Population Biology, 26, 387–407. 97–98, 79–91. Wenink PW, Baker AJ, Tilanus MGJ (1993) Hypervariable control- ShiYF, Zheng BX, Li SJ (1990) Last glaciation and maximum glaci- region sequences reveal global population structuring in a long- ation in Qinghai–Xizang (Tibet) plateau. Journal of Glaciology and distance migrant shorebird, the Dunlin (Calidris alpina). Proceedings Geocryology, 12, 1–15. (In Chinese with English abstract). of the National Academy of Sciences of the United States of America, Slatkin M, Hudson RR (1991) Pairwise comparisons of mito- 90, 94–98. chondrial DNA sequences in stable and exponentially growing Wu YQ, Cui ZJ, Liu GN (2001) Quaternary geomorphological populations. Genetics, 123, 603–613. evolution of the Kunlun Pass areas and uplift of the Qinghai– Sorenson MD, Payne RB (2001) A single ancient origin of brood Xizang (Tibet) plateau. Geomorphology, 36, 203–216. parasitism in African finches: implications for host-parasite Zhang D, Fengquan L, Jianmin B (2000) Eco-environmental effects coevolution. Evolution, 55, 2550–2567. of the Qinghai–Tibet plateau uplift during the Quaternary in Summers-Smith JD (1988) The Sparrows, A Study of the Genus Passer. China. Environmental Geology, 39, 1352–1358. Town Head House, Calton, Wasterhouses, Staffordshire, England. Zheng B, Xu Q, Shen Y (2002) The relationship between climate Tajima F (1989) Statistical method for testing the neutral mutation change and Quaternary glacial cycles on the Qinghai–Tibetan hypothesis by DNA polymorphism. Genetics, 123, 585–595. plateau: review and speculation. Quaternary Internation, 97–98, Tang CZ (1996) Birds of the Hengduan Mountains Region. Sciences 93–101. Press, Beijing. (In Chinese). Zhuo Z, Baoyin Y, Petit-Maire N (1998) Paleoenvironments in Tegelström H, Gelter HP, Jaarola M (1990) Variation in pied China during the Last Glacial Maximum and the Holocene flycatcher (Ficedula hypoleuca) mitochondrial DNA. Auk, 107, optimum. Episodes, 21, 152–158. 730–736. Zink RM, Drovetski SV, Questiau S, Fadeev IV (2003) Recent Templeton AR (1995) A cladistic analysis of phenotypic associations evolutionary history of the bluethroat (Luscinia svecica) across with haplotypes inferred from restriction endonuclease map- Eurasia. Molecular Ecology, 12, 3069–3075.