The Evolutionary History of Biston Suppressaria (Guenã©E

Systematic Entomology (2016), 41, 732–743 DOI: 10.1111/syen.12184

The evolutionary history of Biston suppressaria (Guenée) (Lepidoptera: Geometridae) related to complex topography and geological history

RUI CHENG1,2, NAN JIANG1, DAYONG XUE1, XINXIN LI1,2, XIAOSHUANG BAN1,2 andHONGXIANG HAN1

1Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, Beijing, China and 2University of Chinese Academy of Sciences, Beijing, China

Abstract. The Mountains of Southwest China (MSC) – the most important mountain system in East Asia – creates physical barriers to migration and isolates organisms in different regions. In this study, we explored the genetic structure and phylogeography of Biston suppressaria, a moth species widely distributed in MSC and south China, based on three mtDNA and three ncDNA loci. MtDNA revealed high genetic differentiation among the geographical populations and divided B. suppressaria into five phylogroups (A–E). With the exception of phylogroups B and E which are sympatric, the other three phylogroups have their own distinct distribution areas. Four population differentiations occurred, and the divergence time was consistent with the three phases of the Qinghai–Tibet Movement and Kunhuang Movement, respectively. In contrast, ncDNA did not reveal any phylogeographical structure. Incomplete lineage sorting is considered the most plausible cause of the discordance between mtDNA and ncDNA genes, after excluding secondary admixture and Wolbachia infection. A complex combination of topography and geological movements of south China and MSC contributes to the complex evolutionary history of B. suppressaria.

Introduction region. High mountains act as geographical barriers, which can separate populations, prevent gene flow and produce genetic The Mountains of Southwest China (MSC) is a major global differentiation leading to high genetic diversity. In addition to biodiversity hotspot and is recognized as the most important complex topography, ecological stability is another important mountain system in East Asia as a consequence of its high levels factor driving high genetic diversity within mountainous regions of endemism and species richness (Li, 1988; Myers et al., 2000; (Carnaval et al., 2009). Previous studies have shown that Pleis- Zhao et al., 2007; Huang et al., 2010; Lei et al., 2014). There are tocene glaciations were restricted to high altitudes (>2000 m) many high mountains in this region and its complex topography in the MSC, but not in nearby central China, which had exten- is one major factor driving high genetic diversity and endemism sive ice cover (Li et al., 1991; Liu et al., 2002; Shi, 2002). For (Sarkinen et al., 2011; Fjeldså et al., 2012). After experiencing species distributed below 2000 m within the MSC, the glacial multiple strong tectonic uplifts of the Qinghai–Tibet Plateau oscillations of the Pleistocene had little effect. These two fac- (QTP) in the Pliocene and Pleistocene, this region has become tors (complex topography and ecological stability) are consistent characterized by a series of north–south parallel mountains with with two hypotheses – the ‘mountain barrier hypothesis’ and altitudinal differences often exceeding 2000 m (a.s.l.) between ‘ecological stability hypothesis’ – which are used to explain the valleys and mountain ridges. The resulting topographical com- high genetic diversity of the MSC (Lei et al., 2014, 2015). plexity has led to great environmental heterogeneity within the The MSC is located on the south-east edge of the QTP and its formation was thus connected with the uplift of the QTP. Correspondence: Hongxiang Han, Key Laboratory of Zoological The Qinghai–Tibet Region had an altitude of approximately Systematics and Evolution, Institute of Zoology, Chinese Academy of 1000 m before the Pliocene (Li et al., 1979) and experienced Sciences, No. 1 Beichen West Road, Chaoyang District, Beijing 100101, China. E-mail: [email protected] a strong uplift caused by the Himalayan Orogeny during the

Pliocene. Corresponding to the strong uplift of the QTP over B. suppressaria; (ii) to explore the historical events responsi- the past 3.6 Ma, three main geological movements took place: ble for the present genetic distribution patterns; and (iii) to test the Qinghai–Tibet Movement (3.6 to 1.4 Ma), the Kunhuang whether the influence of the tectonic uplift on population dif- Movement (1.2 to 0.6 Ma) and the Gonghe Movement (0.15 ferentiation of B. suppressaria is consistent with the ‘mountain to 0 Ma) (Li & Fang, 1998; Li, 1999; Zhao et al., 2011). barrier hypothesis’ and ‘ecological stability hypothesis’. The Qinghai–Tibet Movement itself can be divided into three phases: Phase a or Hengduan Movement (3.4 Ma); Phase b or Pulan Movement (2.47 Ma) and Phase c or Yuanmou Move- Materials and methods ment (1.67 Ma) (Chen, 1992; Zhu et al., 1994; Li et al., 1996; Li & Fang, 1998; Shi et al., 1999). Each movement played an Sampling and sequence data important role in promoting species or population differentiation in many kinds of organism: for instance, the Kunhuang A total of 142 specimens of B. suppressaria were collected Movement was a key event in the population differentiation from 26 sampling sites in China (see Table S1, Supporting Infor- of the Niviventer (Marshall) (Mammalia) and Apodemus draco mation; Fig. 1). Samples for DNA extraction were preserved in (Barrett-Hamilton) (Mammalia) and Biston panterinaria (Bre- 100% ethanol and stored at −20∘C. DNA was extracted using the mer & Grey) (Jing et al., 2007; Fan et al., 2012; Cheng et al., DNeasy Tissue kit (Qiagen, Beijing, China) and vouchers were 2016). deposited at the Museum of IZCAS (the Institute of Zoology, South China, the region to the south of the Qinling–Huaihe Chinese Academy of Sciences, Beijing, China). line, is adjacent to the MSC (Zhang, 1999). These two regions Three mitochondrial genes [cytochrome c oxidase subunit I constitute the Oriental realm of China and close connections (COI), cytochrome b (CYTB), NADH dehydrogenase subunit have been found between them. Many species are distributed in 5gene(ND5)], and three nuclear genes [elongation factor both of these regions and some such as Parus monticolus (Vig- 1a (EF-1a), wingless (wg) and glyceraldehyde-3-phosphate ors) (Aves) and Biston panterinaria have experienced popula- dehydrogenase (GADPH)] were targeted through Polymerase tion differentiation between the two regions (Wang et al., 2013; Chain Reaction (PCR) amplification. PCR conditions were as Cheng et al., 2016). For populations distributed in south China, follows: denaturation at 94∘C for 2 min, followed by 40 cycles at especially species with powerful dispersal ability, there are low 93∘C for 1 min, annealing temperature (51∘CforCOI,47∘Cfor intraspecific genetic variations or no obvious phylogeographical CYTB,46∘C for ND5, 56∘CforEF-1a,58∘Cforwg and 55∘C structure (Zhang et al., 2009, 2012). However, populations dis- for GAPDH) for 45 s, and 72∘C for 1 min, and a final 10 min tributed in the MSC show very high genetic diversity and deeper at 72∘C. One pair of new primers of GAPDH was designed by population differentiation (Ma et al., 2012; Qu et al., 2014). Primer Premier 5.0 based on previous sequences from NCBI. For species living at different altitudes, and with different Sequences of all primers used in this study are listed in Table 1. distribution ranges and dispersal abilities, the influence of Sequences were deposited in GenBank; the accession numbers topography and geological history may be different (Lu et al., are provided in Table S1, Supporting Information. 2012; Zhang et al., 2012). In this study, we selected the moth species, Biston suppressaria, mainly distributed in the lowlands (<1000 m) of the MSC and south China, to investigate the Mitochondrial DNA analyses possible effects of topography and geological history in these two regions on its evolutionary history. This species occupies MtDNA and ncDNA were analysed separately to explore many kinds of habitats, including subtropical and tropical potentially different evolutionary histories of the two genomes. evergreen forest, and deciduous broadleaved arbors (Wu, 1960; For mtDNA, in order to assess how genetic diversity varied Yao et al., 2007). In a previous morphological study, we found across geographical populations, the following summary statis- that B. suppressaria varied in several aspects, especially wing tics were calculated based on the combined mtDNA dataset. pattern, but male genitalia showed little variation (Jiang et al., Haplotype diversity (h) and nucleotide diversity (n) were cal- 2011). Thus, the evolutionary relationship between genes and culated to estimate DNA polymorphism using DnaSP 5.10.01 morphological characteristics warrants closer scrutiny. (Librado & Rozas, 2009). McDonald & Kreitman’s (1991) test In this study, two kinds of genes (mtDNA and ncDNA) with was used to examine the selective neutrality of the mtDNA pro- different evolutionary rates were sequenced and analysed sepa- tein coding fragments. Two additional neutrality tests, Tajima’s rately. Mitochondrial genes are linked, whereas nuclear genes D (Tajima, 1989) and Fu’s FS (Fu, 1997), were used to detect (in theory) are unlinked with respect to mtDNA and other departures from the mutation–drift equilibrium that would be nuclear genes (Hoelzer, 1997). Thus, these two genes potentially indicative of changes in historical demography and natural tell different evolutionary stories. Nuclear genes often, but cer- selection. Analysis of molecular variance (amova) and FST tainly not always, evolve more slowly than mitochondrial genes calculations were performed using Arlequin 3.5 (Excoffier & (Miyata et al., 1982; Sunnucks, 2000). A total of 142 specimens Lischer, 2010) with 10 000 permutations, based only on popula- covering the main distribution regions and various populations tions that contained more than three individuals. of B. suppressaria in the MSC were included. The aims were: Haplotype networks were constructed for combined mtDNA (i) to uncover the population genetic structure and demographic sequence to better visualize the nonbifurcating (multifurcations history in order to infer the origin and colonization patterns of and reticulations) relationships (Posada & Crandall, 2001).

Fig. 1. Sampling sites of Biston suppressaria. Colours represent different lineages and shape sizes represent different sample sizes. See Table S1 about the acronyms of the sampling sites.

Table 1. Sequence of all primers.

Region Primer pairs (F/R) Sequence (forward and reverse) 5′ → 3′ Source

COI LCO1490 GGTCAACAAATCATAAAGATATTGG Folmer et al. (1994) HCO2198 TAAACTTCAGGGTGACCAAAAAATCA Folmer et al. (1994) CYTB CP1 GATGATGAAATTTTGGATC Sezonlin et al. (2006) TRS TATTTCTTTATTATGTTTTCAAAAC Simon et al. (1994) ND5 LIU-ND5-F TCTTAATCCGTAAGTTCTTAAACTTGGTAT Liu et al. (2016) LIU-ND5-R GTGGGAAATTATTTCATTTAATTCGTTAAG Liu et al. (2016) EF-1a EF1alepF2 ACAAATGCGGTGGTATCGACAA Yamamoto & Sota (2007) EF1aR GATTTACCRGWACGACGRTC Kawakita et al. (2004) wg Wg-a CGGTGAGGCACAGGGTCGTT Cheng et al. (2016) Wg-b GCAACAGGCACCCATTCG Cheng et al. (2016) GAPDH BisGAPDH-F CTTGATGAGATCGATGACTCGGCTA This study BisGAPDH-R TATGTCGTTGAATCCACTGGCGTCT This study

A maximum parsimony method, implemented in TCS 1.23 The maximum clade credibility tree from divergence- (Clement et al., 2000), was used to draw an unrooted network time-rooted phylogenetic analyses of B. suppressaria was evaluating the haplotype relationships with 95% parsimoniously estimated using BEAST based on three mtDNA loci. Although plausible branch connections. the use of a molecular clock as the only way for calibrating the Phylogenetic trees were reconstructed for mtDNA dataset divergence time of phylogenetic trees is controversial, it does using RAxML 7.2.6 (Stamatakis, 2006; Stamatakis et al., provide a method for estimating approximate time when no other 2008) and BEAST 1.8.0 (Drummond & Rambaut, 2007). The calibration information, such as fossil or geological evidence, maximum-likelihood (ML) analysis was inferred in RAxML is available (Maekawa et al., 2001; López-López et al., 2015). with the GTR + GAMMA + I model for each partition. All Thus, in this study, the widely accepted mutation rates for the model parameters were estimated during the ML analysis. A insect mitochondrial COI gene (0.0115–0.0177 site−1 Ma−1; rapid bootstrapping algorithm was applied with 1000 replicates. Brower, 1994; Papadopoulou et al., 2010) was adopted; other

© 2016 The Royal Entomological Society, Systematic Entomology, 41, 732–743 Evolutionary history of Biston suppressaria 735 genes were scaled to the COI rate in BEAST. Each gene was Coalescent estimates of isolation with migration assigned a separate unlinked relaxed clock model in the analysis. Default priors were used. Chains were run for 200 million Gene flow between inferred groups was tested with the pro- generations, with sampling every 2000 generations. Tracer 1.5.0 gram IMa2 (Hey, 2010) based on mtDNA data. This software is was used to verify the posterior distribution and the effective based on an isolation-with-migration model and estimates effec- sample sizes (ESSs) from the Markov Chain Monte Carlo tive population sizes (present and ancestral), splitting times, and (MCMC) output. TreeAnnotator in the BEAST package was population migration rates using Markov chain Monte Carlo used to summarize tree data with ‘mean height’ and discarded (MCMC) simulations. A uniform distribution with a maximum the first 25% of trees as the ‘burn-in’ period, which ended well of three was used for the migration prior value. The HKY model after the stationarity of the chain likelihood values that had was chosen for each locus. The maximum for population size been established. The tree and divergence times are displayed parameter was set to 100, and the maximum time of splitting was in FigTree 1.3.1. set to five. Runs included a 5000 step burn-in followed by at least 100 000 sampled genealogies. Convergence to the stationary distribution of parameter values was confirmed with three separate Nuclear DNA analyses runs with different random seeds. The lowest effective sample size (ESS) among the six parameters was at least 50 in each For ncDNA sequence datasets, the program PHASE was used run. The mutation rates were set based on substitution rates esti- to phase heterozygous nuclear sequences (Stephens et al., 2001). mated by Brower (1994) and Papadopoulou et al. (2010). The Only individuals with resulting phase probabilities greater than IMfig was used to generate a figure based on the result of IMa2. 0.7 were used in analyses (Harrigan et al., 2008; Carling et al.,

2010). The genetic diversity, Tajima’s D and Fu’s FS test were calculated using DNASP. For each ncDNA gene, the Wolbachia assays phylogenetic analyses were reconstructed using RAXML. Wolbachia infection is a potential factor resulting in the discordance between mtDNA and ncDNA gene (Kodandara- Demographic history maiah et al., 2013). Mitochondrial haplotypes can hitchhike along with the bacteria, leading to drastic changes in pop- Signatures of population demographic changes were tested ulation structure of the mtDNA gene (Johnstone & Hurst, for five phylogroups based on combined mtDNA genes, respec- 1996; Hurst & Jiggins, 2005; Charlat et al., 2009). To tively. Bayesian skyline plots (BSP) implemented in BEAST test the Wolbachia infection among five phylogroups of were used to estimate population size changes through time. For B. suppressaria, two genes (wsp and ftsZ) were amplified using each BSP analysis, the substitution model was selected using primers wsp81F (TGGTCCAATAAGTGATGAAGAAAC) jModeltest. Samples were drawn every 1000 steps for 50 million or wsp178F (AAAGAAGACTGCGGATAC) – wsp691R steps under an uncorrelated lognormal relaxed clock model. The (AAAAATTAAACGCTACTCCA) and ftsZ_F1 (ATYATG- mutation rate was set to 0.0115–0.0177 site−1 Ma−1. Demo- GARCATATAAARGATAG) – ftsZ_R1 (TCRAGYAATGGAT- graphic plots were visualized in Tracer 1.5.0, with a burn-in TRGATA) (Zhou et al., 1998; Baldo et al., 2006). PCR protocols of 20%. used for mtDNA and ncDNA loci followed the conditions of six genes above mentioned, but with annealing temperature set at 50–55∘C. All reaction sets included positive and negative Ancestral area reconstruction controls (dH2O).

Statistical dispersal-vicariance analysis (S-DIVA) and Bayesian binary MCMC (BBM) implemented in RASP 3.02 Results (Yu et al., 2015) were employed to reconstruct the possible ancestral distribution areas of B. suppressaria. The mitochon- Mitochondrial DNA analyses drial ultrametric tree inferred by BEAST was used as input to the program. According to the distribution range of the five phy- A total of 2522-bp fragment of mtDNA (COI: 628 bp; CYTB: logroups, the distribution range of B. suppressaria was divided 1005 bp; ND5: 884 bp) was obtained, which contained 502 into five main areas (A, the Gaoligong Mountains region; B, polymorphic sites, 377 of which were parsimony informative. part region of south China; C, the Southeastern Tibet region; These polymorphic sites defined 32 unique haplotypes, 16 D, south Yunnan and E, south China). These two analyses use of which were found only in single individuals. The most the DIVA method in a statistical context, calculating the prob- abundant haplotype was shared by 36 individuals of phylogroup ability of ancestral areas over a Bayesian posterior distribution E distributed in south China. There is no haplotype shared of tree topologies. The estimation of ancestral area marginal in all populations. Within the five phylogroups, haplotype probabilities, taking into account phylogenetic uncertainty, has diversity (h) values ranged from 0.549 to 0.819, and nucleotide been suggested to reduce uncertainty in the biogeographical diversity (𝜋) from 0.00056 to 0.00237 (Table 2). Values for reconstruction (Nylander et al., 2008). haplotype diversity of phylogroup C were highest, followed in

Table 2. Summary of genetic diversity includes sample size (N), the Demographic history number of haplotypes (nh), nucleotide diversity (𝜋), haplotype diversity (h), Fu’s FS and Tajima’s D. The five phylogroups had negative and positive, but not significant, Tajima’s D and Fu’s F values (Table 2). The BSP 𝜋 S Phylogroup N nh hFu’sFS Tajima’s D analyses rejected population stability in the E phylogroup, but A 22 5 0.00056 0.576 −1.251 −1.39547 could not reject population stability in the phylogroups A, B, C B 8 5 0.00145 0.786 −0.895* −1.7232* and D (Fig. 5). The E phylogroup has continued to expand since C 15 6 0.00146 0.819 0.823 0.33405 0.081 Ma. D 18 5 0.00147 0.549 −1.466 −1.55308 E 74 24 0.00237 0.785 −23.185* −2.06793

*P < 0.05. Ancestral area reconstruction

The results of two methods of S-DIVA and BBM were turn by the B, E, A and D phylogroups. Values for nucleotide consistent (Fig. 6). The results inferred that the origin of diversity of the E phylogroup were highest, followed in turn B. suppressaria was in one of the A, B or C regions, but the by phylogroups D, C, B and A. The McDonald–Kreitman test precise ancestral area was not revealed. The ancestral area of the showed no significant deviation from neutrality for the three C, D and E regions was southeastern Tibet, currently occupied > mtDNA coding segments (P 0.05, with Fisher’s exact test). by phylogroup C. The AMOVA analyses based on the mtDNA dataset revealed genetic differentiation among all populations of B. suppressaria (F = 0.549). Most variation was accounted for by variation ST Coalescent estimates of isolation with migration within groups (45.10%), followed by variation among the five groups (35.11%) and variation among populations within groups The gene flow between groups was estimated under the (19.79%). isolation-with-migration model for mtDNA data. The result is The reconstructed phylogenetic trees and haplotype network shown at Figure S1, Supporting Information. The analysis of of the mtDNA dataset revealed a distinct geographical structure IM2 showed that no gene flow was found between the sympatric in B. suppressaria, and grouped the samples into five major B and E phylogroups. The common ancestor of the phylogroups phylogroups (Fig. 2): phylogroup A including samples from the B and E existed about 3.998 Ma. This date was consistent with Gaoligong Mountains; B including samples from south China; the first divergence time [3.2 Ma (2.22–4.40 Ma, 95% HPD)] C including southeastern Tibet; D including samples from the between (A + B) and (C + D + E) estimated by BEAST. tropical area of southern Yunnan; and E including samples from south China (Fig. 1). B and E phylogroups are sympatric. Divergence time dating (Fig. 3) indicated that divergence Wolbachia assays within B. suppressaria happened four times. The first divergence occurred between phylogroups (A and B) and (C, D and E) Two gene segments and annealing temperature from 50 to at approximately 3.2 Ma [2.22–4.40 Ma, 95% highest posterior 55∘CwereusedtotestforWolbachia infection repeatedly. No density (HPD)]. The second divergence occurred between A and Wolbachia infection was detected in any of the samples from five B phylogroups at approximately 2.49 Ma (1.72–3.50 Ma, 95% phylogroups of B. suppressaria. It is possible that the Wolbachia HPD). The third divergence occurred between phylogroup C and infection frequency of our sampled B. suppressaria is zero or phylogroups (D + E) at approximately 1.64 Ma (1.10–2.29 Ma, very low. 95% HPD). The fourth divergence occurred between phylogroup D and phylogroup E at approximately 0.50 Ma. The first three dates were roughly congruent with the three phasesof Discussion the Qinghai–Tibet Movement, respectively. The last divergence time was roughly congruent with the Kunhuang Movement. Evolutionary history of Biston suppressaria based on mtDNA data

Nuclear DNA analyses The mtDNA results suggest that Biston suppressaria in China consists of five populations: the phylogroups A, B, C, Dand Three ncDNA fragments (EF-1a: 972 bp; wg: 735 bp; E. Three phylogroups have disjunct distributions: phylogroup A GAPDH: 647 bp) were obtained. For EF-1a, wg and GAPDH, includes individuals from the Gaoligong Mountains; phylogroup 59, 36 and 72 polymorphic sites were present, respectively. C includes individuals from southeastern Tibet (represented by The genetic diversity of the five phylogroups is shown at Table Mêdog); and phylogroup D includes individuals from south S2a–c, Supporting Information. The phylogenetic analyses of Yunnan (represented by Xishuangbanna). The phylogroups B each ncDNA gene (Fig. 4) revealed similar results: all phy- and E co-occur at a few locations in south China: phylogroup logroups were mixed and did not reveal any correspondence B is the relatively older group and the sympatric phylogroup E between geographical origins and genetically defined groups. is the relatively younger group. The analysis of ancestral area

Fig. 2. (a) Median joining network based on mitochondrial DNA (mtDNA) for Biston suppressaria. Each circle represents a haplotype and the size of the circle is proportional to that haplotype’s frequency. Colours denote lineage membership and are the same as in Fig. 1. (b) Maximum-likelihood tree based on combined mtDNA genes for B. suppressaria. Values on the left of nodes indicate bootstrap supports for major clades. (c) Respective adults of five phylogroups. reconstruction failed to reveal the specific ancestral area, but did The role of topography and geological history of the Mountains confirm that the C region (southeastern Tibet) is the ancestral of Southwest China (MSC) and south China area of the D (south Yunnan) and E (south China) regions. Because there is no other distribution region accessing south In the face of Pleistocene environmental fluctuations, the three China except south Yunnan, south Yunnan is the most likely path phylogroups (A, C and D) distributed in the MSC as well as its into south China. If this is the case, the young E phylogroup adjacent regions maintained stability (Fig. 5). This is consistent represents a secondary dispersal of the C lineage. with the ‘ecological stability hypothesis’, which posits that the Four gradual population differentiation events are indi- low altitudes of the MSC may have provided relatively stable cated in the evolution of B. suppressaria at approximately environmental conditions for species (Qu et al., 2014; Lei et al., 3.20, 2.49, 1.64 and 0.55 Ma, respectively. The former three 2015). Similar results have also been found in other organisms, differentiation events occurred during phases a, b and c such as birds (Lei et al., 2007), plants (Wu et al., 2005), other (respectively) of the Qinghai–Tibet Movement (Li et al., moths (Cheng et al., 2016) and aphids (Huang et al., 2010). 1979; Li, 1999). The last differentiation event between phy- Results for phylogroups B and E distributed in south China are logroups D and E occurred during the Kunhuang Movement different. The older phylogroup B has maintained stability. The (Cui et al., 1997, 1998). younger phylogroup E has maintained a continuous expansion

phylogroups. Apparently, the physical barrier isolating these two phylogroups was the Yungui Plateau. The third divergence occurred between phylogroups C and (D + E) with a physical barrier situated roughly in the Kachin Hills. The fourth divergence occurred between phylogroups D and E with the physical barrier of the mountains of northern Yunnan (a branch of the Hengduan Mountains). In addition, the divergence times of the four population differentiations are consistent with the times of phases a, b and c of the Qinghai–Tibet Movement and the Kunhuang Movement, respectively. These movements were the most intense and frequent uplift events induced by the uplift of the Qinghai–Tibet Plateau during the Pleistocene (Li & Fang, 1998; Li, 1999; Zhao et al., 2011) and can be linked with multiple differentiation events of endemic organisms in the MSC, including mammals, amphibians, reptiles and insects (Jing et al., 2007; Jin et al., 2008; Ci et al., 2009; Fan et al., 2012; Li et al., 2012; Tan Fig. 3. The maximum clade credibility tree from divergence- et al., 2012). Based on distribution range, physical barriers time-rooted phylogenetic analysis of Biston suppressaria basedona and divergence time, we believe that the complex topography mitochondrial DNA (mtDNA) dataset. Posterior probabilities of clade of the MSC resulting from tectonic uplift created physical support are shown at nodes above branches. Estimates of divergence barriers to migration and isolated B. suppressaria populations time with 95% confidence intervals are shown at nodes as purple bars in different regions and mountain systems. Species living at low and numbers below branches. altitudes, such as B. suppressaria, appear to be more liable to fragmentation of ancestral populations (Lu et al., 2012). This is since 0.081 Ma (including the Late Glacial Maximum and Last consistent with the ‘mountain barrier hypothesis’, which was Interglacial). A possible reason for this may be that the two proposed to explain the importance of mountain barriers in groups are in different phases of population growth: the older lineage diversification of multiple intraspecific genetic lineages phylogroup B may have suffered from genetic bottleneck and in the MSC (Lei et al., 2015). be in population equilibrium, whereas the younger phylogroup South China, the region to the south of the Qinling–Huaihe E may be in a period of rapid population growth with relatively line, is adjacent to the MSC (Zhang, 1999). These two regions small survival pressure (Anderson, 1971; May, 1975). have close connections but also some differences. Compared Of the four population differentiations in B. suppressaria with the groups distributed in the MSC, the phylogroups of revealed by mtDNA data, the first divergence occurred between B. suppressaria in south China have low genetic diversity. This the phylogroups (A + B) and (C + D + E), and southeastern character may be linked with the relatively smooth topography Tibet (the distribution region of phylogroup C) is the ances- of south China, which has little effect on species with powerful tral area of the latter (Fig. 6). The physical barrier isolating dispersal ability, such as B. suppressaria.However,theMSCis phylogroups (A + B) and C is probably the Baxoila–Gaoligong not closed and still has some routes available to south China, Mountains and the Bomi–Zayü region, which is also the as evidenced by the successful recolonisation of south China boundary of the Himalayan and Southwest Subregion (Zhang, by phylogroup C of this study from southeastern Tibet, again 1999). The second divergence occurred between the A and B through south Yunnan. This type of secondary dispersal is

Fig. 4. Maximum-likelihood tree based on each nuclear DNA (ncDNA) gene for Biston suppressaria. Colours indicate the distributions of individuals in relation to their geographical grouping. Orange, phyloroup A; light blue, phylogroup B; purple, phylogroup C; dark blue, phylogroup D; green, phylogroup E.

Fig. 5. Bayesian skyline plots for five phylogroups of Biston suppressaria. The x axis is in units of millions of years ago. The y axis is equal to the effective population size. Estimates of means are joined by a solid line, whereas the upper and lower lines delineate the 95% HPD limits. E means a natural constant and is approximately equal to 2.718. LGM, Last Glacial Maximum; LIG, Last Interglacial. different from that of plants and birds, driven by wind and and instead show a widely shared nuclear polymorphism. Such the differential fitness of philopatric and dispersing individuals discordance has been observed in several organisms (Mao et al., (Bélichon et al., 1996; Vander Wall et al., 2005). Altogether, a 2010; Kindler et al., 2012; Qu et al., 2012). Reasons adduced complex combination of topography and geological movements for this kind of discordance between mtDNA and ncDNA genes of south China and MSC has contributed to the complex include incomplete lineage sorting of ancestral polymorphisms, evolutionary history of B. suppressaria. recent admixture or Wolbachia infection (Ballard et al., 2002; Funk & Omland, 2003; Ballard & Whitlock, 2004; Hurst & Jig- Incongruence between mitochondrial (mtDNA) and nuclear gins, 2005; Charlat et al., 2009; Kodandaramaiah et al., 2013). (ncDNA) genealogies In our study, apart from the sympatric distribution of phylogroups B and E in south China, the phylogroups are allopatric In contrast to mtDNA, the analyses of the ncDNA dataset and there are evident geographical barriers between them do not reveal any geographic structuring in B. suppressaria (Fig. 1). The analysis of gene flow reveals that no gene flow

differences between phylogroups B and E. For this reason it seems possible that the mechanism, whatever it may be, which results in reproductive isolation between the young phylogroup E and the old phylogroup B existed before the latter group spread into south China, as suggested above. The discordance between mtDNA and ncDNA genes may result in uncertainty and difficulty for taxonomy. According to the divergence of COI (1.8–10.1%) between five phylogroups, five subspecies are well supported (Hebert et al., 2003, 2004; Hausmann et al., 2011). All five phylogroups have stable morphological characteristics, with transverse lines, the absence of the median yellow band and the sparser black dots on the wings (Fig. 2). However, it seems insufficient to establish subspecies based on the male genitalia and phylogeny of ncDNA. We need to expand the different data, methods and perspectives to draw a reasonable conclusion of the validity of the subspecies in the future.

Supporting Information

Additional Supporting Information may be found in the online version of this article under the DOI reference: Fig. 6. Biogeographical ancestral optimization performed with statisti- 10.1111/syen.12184 cal dispersal-vicariance analysis (S-DIVA) and Bayesian binary MCMC (BBM). A circle with more than one colour indicates more than one pos- Table S1. GenBank accession numbers. sibility with the same probability. Black means other unknown region. Table S2. Genetic diversity of each ncDNA gene. Biogeographical regions: A, the Gaoligong Mountains region; B, part Figure S1. Summary of the IMa2 results. region of south China; C, the Southeastern Tibet region; D, south Yun- nan; E, south China. Acknowledgements exists between the sympatric phylogroups B and E. In addition, We gratefully thank all collectors for their assistance in collect- the result of Wolbachia assays shows that none of the five phy- ing the specimens used in this study. We thank Sir Anthony logroups in our study are infected. The low or null Wolbachia Galsworthy, The Natural History Museum, London, U.K., infection frequencies have little effect on discordance between for giving valuable comments on the research and correcting mtDNA and ncDNA genes. Against this background, incom- the English of this manuscript. This work was supported by plete lineage sorting seems to be the most likely reason for the the National Science Foundation of China (Nos. 31272288, discordance between mtDNA and ncDNA genes. As populations 31372176, 31572301), the National Science Fund for Fostering recently diverge from ancestral lineages, the transition to mono- Talents in Basic Research (No. NSFC-J1210002) and a grant phyly requires 8–12 Ne generations (Ne, the effective population from the Key Laboratory of the Zoological Systematics and Evo- size) (Hudson, 1990; Chesser & Baker, 1996; Hoelzer, 1997). lution of the Chinese Academy of Sciences (No. O529YX5105). The Ne of maternally inherited mtDNA is four times smaller than that of biparentally inherited ncDNA, which suggest that the former need shorter times for lineage sorting of ancestral References polymorphisms than ncDNA (Palumbi et al., 2001; Sefc et al., Anderson, W.W. (1971) Genetic equilibrium and population growth 2005; Qu et al., 2012). Such discordance between mtDNA and under density-regulated selection. American Naturalist, 105, ncDNA may result in the deep divergence in mtDNA but admix- 489–498. ture in ncDNA. Baldo, L., Hotopp, J.C.D., Jolley, K.A. et al. (2006) Multilocus sequence The lack of gene flow between phylogroups B and typing system for the endosymbiont Wolbachia pipientis. Applied and EofB. suppressaria suggests that they are significantly Environmental Microbiology, 72, 7098–7110. reproductively isolated, suggesting incipient speciation (Mayr, Ballard, J.W. & Whitlock, M.C. (2004) The incomplete natural history 1942, 1963, 1970; Via, 1999). When sympatric speciation of mitochondria. Molecular Ecology, 13, 729–744. occurs, it is likely to involve a rapid evolution of divergence Ballard, J.W., Chernoff, B. & James, A.C. (2002) Divergence of mitochondrial DNA is not corroborated by nuclear DNA, morphology, or in response to selection in different ecological environments, behavior in Drosophila simulans. Evolution, 56, 527–545. such as host choice or temperature adaptability (Bush, 1994; Bélichon, S., Clobert, J. & Massot, M. (1996) Are there differences in Schluter, 1998). Based on previous literature and our obser- fitness components between philopatric and dispersing individuals? vations of B. suppressaria, we found no evidence for such Acta Oecologica, 17, 503e517.

Brower, A.V.Z. (1994) Rapid morphological radiation and convergence Hausmann, A., Haszprunar, G. & Hebert, P.D. (2011) DNA barcoding among races of the butterfly Heliconius erato inferred from patterns of the geometrid fauna of Bavaria (Lepidoptera): successes, surprises, mitochondrial DNA evolution. Proceedings of the National Academy and questions. PLoS ONE, 6, e17134. of Sciences of the United States of America, 91, 6491–6495. Hebert, P.D.,Cywinska, A. & Ball, S.L. (2003) Biological identifications Bush, G.L. (1994) Sympatric speciation in animals: new wine in old through DNA barcodes. Proceedings of the Royal Society B: Biolog- bottles. Trends in Ecology & Evolution, 9, 285–288. ical Sciences, 270, 313–321. Carling, M.D., Lovette, I.J. & Brumfield, R.T. (2010) Historical diver- Hebert, P.D., Penton, E.H., Burns, J.M., Janzen, D.H. & Hallwachs, W. gence and gene flow: coalescent analyses of mitochondrial, autosomal (2004) Ten species in one, DNA barcoding reveals cryptic species in and sex-linked loci in Passerina buntings. Evolution, 64, 1762–1772. the neotropical skipper butterfly Astraptes fulgerator. Proceedings of Carnaval, A.C., Hickerson, M.J., Haddad, C.F.B., Rodrigues, M.T. & the National Academy of Sciences of the United States of America, Moritz, C. (2009) Stability predicts genetic diversity in the Brazilian 101, 14812–14817. Atlantic forest hotspot. Science, 323, 785–789. Hey, J. (2010) Isolation with migration models for more than two Charlat, S., Duplouy, A., Hornett, E.A. et al. (2009) The joint evolution- populations. Molecular Biology and Evolution, 27, 905–920. ary histories of Wolbachia and mitochondria in Hypolimnas bolina. Hoelzer, G.A. (1997) Inferring phylogenies from mtDNA variation: BMC Evolutionary Biology, 9, 64. mitochondrial-genes trees versus nuclear-gene trees revisited. Evolu- Chen, F.B. (1992) Hengduan event: an important tectonic event of the tion, 51, 622–626. late Cenozoic in Eastern Asia. Mountain Research, 10, 195–202. Huang, X.L., Qiao, G.X. & Lei, F.M. (2010) Use of parsimony analysis Cheng, R., Jiang, N., Yang, X.S., Xue, D.Y. & Han, H.X. (2016) The to identify areas of endemism of Chinese birds: implications for influence of geological movements on the population differentiation conservation and biogeography. International Journal of Molecular of Biston panterinaria (Lepidoptera: Geometridae). Journal of Bio- Sciences, 11, 2097–2108. geography, 43, 691–702. Hudson, R.R. (1990) Gene genealogies and the coalescent process. Chesser, R.K. & Baker, R.J. (1996) Effective sizes and dynamics Oxford Surveys in Evolutionary Biology, 7, 1–44. of uniparentally and diparentally inherited genes. Genetics, 144, Hurst, G.D. & Jiggins, F.M. (2005) Problems with mitochondrial DNA 1225–1236. as a marker in population, phylogeographic and phylogenetic studies: Ci, H.X., Lin, G.H., Cai, Z.Y., Tang, L.Z., Su, J.P. & Liu, J.Q. (2009) the effects of inherited symbionts. Proceedings of the Royal Society Population history of the plateau pika endemic to the Qinghai-Tibetan B: Biological Sciences, 272, 1525–1534. Plateau based on mtDNA sequence data. Journal of Zoology, 279, Jiang, N., Xue, D.Y. & Han, H.X. (2011) A review of Biston Leach, 396–403. 1815 (Lepidoptera, Geometridae, Ennominae) from China, with Clement, M., Posada, D. & Crandall, K.A. (2000) TCS: a com- description of one new species. ZooKeys, 139, 45–96. puter program to estimate gene genealogies. Molecular Ecology, 9, Jin, Y.T., Brown, R.P.& Liu, N.F. (2008) Cladogenesis and phylogeogra- 1657–1659. phy of the lizard Phrynocephalus vlangalii (Agamidae) on the Tibetan Cui, Z.J., Wu, Y.Q. & Liu, G.S. (1997) Findings and character- plateau. Molecular Ecology, 17, 1971–1982. istics of Kun-Huang movement. Chinese Science Bulletin, 42, Jing, M.D., Yu, H.T., Wu, S.H., Wang, W. & Zheng, X.G. (2007) 1986–1989. Phylogenetic relationships in genus Niviventer (Rodentia: Muridae) Cui, Z.J., Wu, Y.Q., Liu, G.N., Ge, D.K., Pang, Q.Q. & Xu, Q.H. (1998) in China inferred from complete mitochondrial cytochrome b gene. On Kunlun-Yellow River tectonic movement. Science in China: Molecular Phylogenetics and Evolution, 44, 521–529. Series D, 28, 53–59. Johnstone, R.A. & Hurst, G.D. (1996) Maternally inherited male-killing Drummond, A.J. & Rambaut, A. (2007) BEAST: Bayesian evolutionary microorganisms may confound interpretation of mitochondrial analysis by sampling trees. BMC Evolutionary Biology, 7, 214. DNA variability. Biological Journal of the Linnean Society, Excoffier, L. & Lischer, H.E.L. (2010) Arlequin suite ver 3.5:a 58, 453–470. new series of programs to perform population genetics analy- Kawakita, A., Takimura, A., Terachi, T., Sota, T. & Kato, M. (2004) ses under Linux and Windows. Molecular Ecology Resources, 10, Cospeciation analysis of an obligate pollination mutualism: have 564–567. Glochidion trees (Euphorbiaceae) and pollinating Epicephala Fan, Z.X., Liu, S.Y., Liu, Y., Liao, L.H., Zhang, X.Y. & Yue, B.S. (2012) moths (Gracillariidae) diversified in parallel? Evolution, 58, Phylogeography of the South China field mouse (Apodemus draco) 2201–2214. on the southeastern Tibetan Plateau reveals high genetic diversity and Kindler, E., Arlettaz, R. & Heckel, G. (2012) Deep phylogeographic glacial refugia. PLoS ONE, 7, e38184. divergence and cytonuclear discordance in the grasshopper Oedaleus Fjeldså, J., Bowie, R.C.K. & Rahbek, C. (2012) The role of mountain decorus. Molecular Phylogenetics and Evolution, 65, 695–704. ranges in the diversification of birds. Annual Review of Ecology, Kodandaramaiah, U., Simonsen, T.J., Bromilow, S., Wahlberg, N. & Evolution, and Systematics, 43, 249–256. Sperling, F. (2013) Deceptive single-locus taxonomy and phylogeog- Folmer, O., Black, M., Hoeh, W., Lutz, R. & Vrijenhoek, R. (1994) DNA raphy: Wolbachia-associated divergence in mitochondrial DNA is not primers for amplification of mitochondrial cytochrome C oxidase reflected in morphology and nuclear markers in a butterfly species. subunit I from diverse metazoan invertebrates. Molecular Marine Ecology and Evolution, 3, 5167–5176. Biology and Biotechnology, 3, 294–299. Lei, F.M., Wei, G.A., Zhao, H.F., Yin, Z.H. & Lu, J.L. (2007) China sub- Fu, Y. (1997) Statistical tests of neutrality of mutations against popu- regional avian endemism and biodiversity conservation. Biodiversity lation growth, hitchhiking and background selection. Genetics, 147, and Conservation, 16, 1119–1130. 915. Lei, F.M., Qu, Y.H. & Song, G. (2014) Species diversification and Funk, D.J. & Omland, K.E. (2003) Species-level paraphyly and poly- phylogeographical patterns of birds in response to the uplift of the phyly: frequency, causes, and consequences, with insights from ani- Qinghai-Tibet Plateau and Quaternary glaciations. Current Zoology, mal mitochondrial DNA. Annual Review of Ecology, Evolution, and 60, 149–161. Systematics, 34, 397–423. Lei, F.M., Qu, Y.H., Song, G., Alstrom, P. & Fjeldså, J. (2015) The Harrigan, R.J., Mazza, M.E. & Sorenson, M.D. (2008) Computation potential drivers in forming avian biodiversity hotpots in the east vs. cloning: evaluation of two methods for haplotype determination. Himalaya Mountains of southwest China. Integrative Zoology, 10, Molecular Ecology Resources, 8, 1239–1248. 171–181.

Li, D.J. (1988) The primary characteristics of flora in the Hengduan Nylander, J.A., Olsson, U., Alström, P. & Sanmartín, I. (2008) Account- Mountainous regions. Mountain Research, 6, 147–152. ing for phylogenetic uncertainty in biogeography: a Bayesian Li, J.J. (1999) Studies on the geomorphological evolution of the approach to dispersal-vicariance analysis of the thrushes (Aves: Qinghai-Xizang (Tibetan) Plateau and Asian monsoon. Marine Geol- Turdus). Systematic Biology, 57, 257–268. ogy & Quaternary Geology, 19, 1–11. Palumbi, S.R., Cipriano, F. & Hare, M.P. (2001) Predicting nuclear gene Li, J.J. & Fang, X.M. (1998) The uplift of Qinghai-Xizang (Tibetan) coalescence from mitochondrial data: the three time rule. Evolution, Plateau and the studies of environmental changes. Chinese Science 55, 859–868. Bulletin, 43, 1569–1574. Papadopoulou, A., Anastasiou, I. & Vogler, A.P. (2010) Revisiting Li, J.J., Wen, S.X., Zhang, Q.S., Wang, F.B., Zheng, B.X. & Li, B.Y. the insect mitochondrial molecular clock: the mid-Aegean trench (1979) The discussion about age, amplitude and form of the uplift of calibration. Molecular Biology and Evolution, 27, 1659–1672. Qinghai-Tibet Plateau. Science in China, 6, 608–616. Posada, D. & Crandall, K.A. (2001) Intraspecific gene genealogies: Li, Z.W., Chen, Z.R. & Wang, M.L. (1991) Classification and correla- trees grafting into networks. Trends in Ecology and Evolution, tion of the Quaternary glacial epoch in the Hengduan (Transverse) 16, 37–45. Mountains. Geological Reviews, 37, 125–132. Qu, Y., Zhang, R., Quan, Q., Song, G., Li, S.H. & Lei, F. (2012) Incom- Li, J.J., Fang, X.M. & Ma, H.Z. (1996) The geomorphic evolution of the plete lineage sorting or secondary admixture: disentangling historical upstream of Yellow River and the uplift of Qinghai-Xizang (Tibetan) divergence from recent gene flow in the Vinous-throated parrotbill Plateau in the late Cenozoic. Science in China, Series D, 26, 316–322. (Paradoxornis webbianus). Molecular Ecology, 21, 6117–6133. Li, Z.J., Yu, G.H., Rao, D.Q. & Yang, J. (2012) Phylogeography Qu, Y.H., Song, G., Gao, B., Quan, Q., Ericson, P.G.P. & Lei, F.M. and demographic history of Babina pleuraden (Anura, Ranidae) in (2014) The influence of geological events on the endemism of East southwestern China. PLoS ONE, 7, e34013. Asian birds studied through comparative phylogeography. Journal of Librado, P. & Rozas, J. (2009) DnaSP v5: a software for comprehensive Biogeography, 42, 179–192. analysis of DNA polymorphism data. Bioinformatics, 25, 1451. Sarkinen, T., Pennington, R.T., Lavin, M., Simon, M.F. & Hughes, C.E. Liu, J., Yu, G. & Chen, X. (2002) Paleoclimatic simulation of 21 ka for (2011) Evolutionary islands in the Andes: persistence and isolation the Qinghai-Tibetan plateau and eastern Asia. Climate Dynamics, 19, explain high endemism in Andean dry tropical forests. Journal of 575–583. Biogeography, 39, 884–900. Liu, S.X., Jiang, N., Xue, D.Y. et al. (2016) Evolutionary history of Schluter, D. (1998) Ecological causes of speciation. Endless Forms: Apocheima cinerarius (Lepidoptera: Geometridae), a female flight- Species and Speciation (ed. by D.J. Howard and S.H. Berlocher), pp. less moth in northern China. Zoologica Scripta, 45, 160–174. 114–129. Oxford University Press, Oxford. López-López, A., Abdul Aziz, A. & Galián, J. (2015) Molecular Sefc, K.M., Payne, R.B. & Sorenson, M.D. (2005) Genetic continu- phylogeny and divergence time estimation of Cosmodela (Coleoptera: ity of brood-parasitic indigobird species. Molecular Ecology, 14, Carabidae: Cicindelinae) tiger beetle species from Southeast Asia. 1407–1419. Zoologica Scripta, 44, 437–445. Sezonlin, M., Dupas, S., LeRü, B. et al. (2006) Phylogeography and pop- Lu, B., Zheng, Y., Murphy, R.W. & Zeng, X. (2012) Coalescence pat- ulation genetics of the maize stalk borer Busseola fusca (Lepidoptera, terns of endemic Tibetan species of stream salamanders (Hynobiidae: Noctuidae) in sub-Saharan Africa. Molecular Ecology, 15, 407–420. Batrachuperus). Molecular Ecology, 21, 3308–3324. Shi, Y.F. (2002) Characteristics of late Quaternary monsoonal glaciation Ma, C., Yang, P., Jiang, F., Chapuis, M.P., Shali, Y., Sword, G.A. & on the Qinghai-Tibetan plateau and in East Asia. Quaternary Interna- Kang, L. (2012) Mitochondrial genomes reveal the global phylogeog- tional, 97–98, 79–91. raphy and dispersal routes of the migratory locust. Molecular Ecol- Shi, Y.F., Li, J.J., Li, B.Y. et al. (1999) Uplift of the Qinghai-Xizang ogy, 21, 4344–4358. (Tibetan) Plateau and east Asia environmental change during late Maekawa, K., Kon, M., Araya, K. & Matsumoto, T. (2001) Phylogeny Cenozoic. Acta Geographica Sinica, 54, 10–20. and biogeography of wood-feeding cockroaches, genus Salganea Stål Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H. & Flook, P. (Blaberidae: Panesthiinae), in southeast Asia based on mitochondrial (1994) Evolution, weighting and phylogenetic utility of mitochondrial DNA sequences. Journal of Molecular Evolution, 53, 651–659. gene sequences and a compilation of conserved polymerase chain Mao, X., Zhang, J., Zhang, S. & Rossiter, S.J. (2010) Historical reaction primers. Annals of the Entomological Society of America, 87, male-mediated introgression in horseshoe bats revealed by multilocus 651–701. DNA sequence data. Molecular Ecology, 19, 1352–1366. Stamatakis, A. (2006) RAxML-VI-HPC, maximum likelihood-based May, R.M. (1975) Biological populations obeying difference equations: phylogenetic analyses with thousands of taxa and mixed models. stable points, stable cycles, and chaos. Journal of Theoretical Biology, Bioinformatics, 22, 2688–2690. 51, 511–524. Stamatakis, A., Hoover, P. & Rougemont, J. (2008) A rapid bootstrap Mayr, E. (1942) Systematics and the Origin of Species. Columbia algorithm for the RAxML web servers. Systematic Biology, 57, University Press, New York, New York. 758–771. Mayr, E. (1963) Animal Species and Evolution. Harvard University Stephens, M., Smith, N.J. & Donnelly, P. (2001) A new statistical Press, Cambridge, Massachusetts. method for haplotype reconstruction from population data. The Mayr, E. (1970) Population, Species, and Evolution. Harvard University American Journal of Human Genetics, 68, 978–989. Press, Cambridge, Massachusetts. Sunnucks, P. (2000) Efficient genetic markers for population biology. McDonald, J.H. & Kreitman, M. (1991) Adaptive protein evolution at Trends in Ecology & Evolution, 15, 199–203. the Adh locus in Drosophila. Nature, 351, 652–654. Tajima, F. (1989) Statistical method for testing the neutral mutation Miyata, T., Hayashida, H., Kikuno, R., Hasegawa, M., Kobayashi, M. hypothesis by DNA polymorphism. Genetics, 123, 585. & Koike, K. (1982) Molecular clock of silent substitution: at least Tan, S., Zou, D., Tang, L. et al. (2012) Molecular evidence for the sub- six-fold preponderance of silent changes in mitochondrial genes over specific differentiation of blue sheepPseudois ( nayaur) and poly- those in nuclear genes. Journal of Molecular Evolution, 19, 28–35. phyletic origin of dwarf blue sheep (Pseudois schaeferi). Genetica, Myers, N., Mittermeier, R.A., Mittermeier, C., da Fonseca, G.A.B. 140, 159–167. & Kent, J. (2000) Biodiversity hotspots for conservation priorities. Vander Wall, S.B., Kuhn, K.M. & Beck, M.J. (2005) Seed removal, seed Nature, 403, 853–858. predation, and secondary dispersal. Ecology, 86, 801–806.

Via, S. (1999) Reproductive isolation between sympatric races of pea Zhang, R.Z. (1999) Zoogeography of China. Science Press, Beijing. aphids. I. Gene flow restriction and habitat choice. Evolution, 53, Zhang, D., Yan, L., Ji, Y., Hewitt, G.M. & Huang, Z. (2009) Unexpected 1446–1457. relationships of substructured populations in Chinese Locusta migra- Wang, W., McKay, B.D., Dai, C. et al. (2013) Glacial expan- toria. BMC Evolutionary Biology, 9, 144. sion and diversification of an East Asian montane bird, the Zhang, R., Song, G., Qu, Y. et al. (2012) Comparative phylogeography green-backed tit (Parus monticolus). Journal of Biogeography, 40, of two widespread magpies: Importance of habitat preference and 1156–1169. breeding behavior on genetic structure in China. Molecular Phylo- Wu, C.C. (1960) The research reports of Biston suppressaria. Forestry genetics and Evolution, 65, 562–572. Science, 2, 158–160. Zhao, H.F., Lei, F.M., Qu, Y.H. & Lu, J.L. (2007) Conservation priority Wu, Z., Sun, H., Zhou, Z., Peng, H. & Li, D.Z. (2005) Origin and based on avian subspecific differentiation of endemic species. Acta differentiation of endemism in the flora of China. Acta Botanica Zoologica Sinica, 53, 378–382. Yunnanica, 27, 577. Zhao, J.D., Shi, Y.F. & Wang, J. (2011) Comparison between Quaternary Yamamoto, S. & Sota, T. (2007) Phylogeny of the Geometridae and glaciations in China and the marine oxygen isotope stage (MIS):an the evolution of winter moths inferred from a simultaneous analysis improved schema. Acta Geographica Sinica, 66, 867–884. of mitochondrial and nuclear genes. Molecular Phylogenetics and Zhou, W., Rousset, F. & O’Neil, S. (1998) Phylogeny and PCR-based Evolution, 44, 711–723. classification of Wolbachia strains using wsp gene sequences. Pro- Yao, Z., He, P.L., Su, B. & Huang, H.Y.(2007) Biological characteristics ceedings of the Royal Society B: Biological Sciences, 265, 509–515. and control tactics of eucalypt Buzura suppressatia. Eucalypt Science Zhu, Z.Y., Ding, Z.L. & Han, J.T. (1994) Activations of neotectonics and & Technology, 24, 41–43. drastic changes in climate. Quaternary Sciences, 1, 56–66. Yu, Y., Harris, A.J., Blair, C. & He, X. (2015) RASP (Reconstruct Ancestral State in Phylogenies): a tool for historical biogeography. Accepted 11 March 2016 Molecular Phylogenetics and Evolution, 87, 46–49. First published online 3 May 2016