Molecular Phylogenetics and Evolution 69 (2013) 717–727

Contents lists available at SciVerse ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier .com/ locate /ympev

A new phylogeny for the genus Picea from plastid, mitochondrial, andnuclear sequences

1055-7903/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.ympev.2013.07.004

⇑ Corresponding author.E-mail address: renner@lmu.de (S.S. Renner).

1 These authors contributed equally to the work.

Jared D. Lockwood a,1, Jelena M. Aleksic b,1, Jiabin Zou c,1, Jing Wang c, Jianquan Liu c, Susanne S. Renner a,⇑a Systematic Botany and Mycology, University of Munich (LMU), Menzinger Strasse 67, 80638 Munich, Germanyb University of Belgrade, Institute of Molecular Genetics and Genetic Engineering, 11010 Belgrade, Serbiac State Key Laboratory of Grassland Agro-ecosystem, School of Life Sciences, Lanzhou University, Lanzhou, Gansu 730000, China

a r t i c l e i n f o

Article history:Received 28 March 2013Revised 21 June 2013Accepted 5 July 2013Available online 18 July 2013

Keywords:Molecular clocksHistorical biogeographyMitochondrial nad intronsStochastic Dollo modelSecondary structure-based alignmentNorth American sprucesFossil calibrations

a b s t r a c t

Studies over the past ten years have shown that the crown groups of most conifer genera are only about15–25 Ma old. The genus Picea (spruces, Pinaceae), with around 35 species, appears to be no exception. Inaddition, molecular studies of co-existing spruce species have demonstrated frequent introgression. Per-haps not surprisingly therefore previous phylogenetic studies of species relationships in Picea, basedmostly on plastid sequences, suffered from poor statistical support. We therefore generated mitochon-drial, nuclear, and further plastid DNA sequences from carefully sourced material, striking a balancebetween alignability with outgroups and phylogenetic signal content. Motif duplications in mitochon-drial introns were treated as characters in a stochastic Dollo model; molecular clock models were cali-brated with fossils; and ancestral ranges were inferred under maximum likelihood. In agreement withprevious findings, Picea diverged from its sister clade 180 million years ago (Ma), and the most recentcommon ancestor of today’s spruces dates to 28 Ma. Different from previous analyses though, we finda large Asian clade, an American clade, and a Eurasian clade. Two expansions occurred from Asia to NorthAmerica and several between Asia and Europe. Chinese P. brachytyla, American P. engelmannii, and Nor-way spruce, P. abies, are not monophyletic, and North America has ten, not eight species. Divergencetimes imply that Pleistocene refugia are unlikely to be the full explanation for the relationships betweenthe European species and their East Asian relatives. Thus, northern Norway spruce may be part of anAsian species complex that diverged from the southern Norway spruce lineage in the Upper Miocene,some 6 Ma, which can explain the deep genetic gap noted in phylogeographic studies of Norway spruce.The large effective population sizes of spruces, and incomplete lineage sorting during speciation, meanthat the interspecific relationships within each of the geographic clades require further studies, especiallybased on genomic information and population genetic data.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

Gymnosperms comprise just over 1000 species in 85 genera,with conifers making up 600 species and 68 genera (Christenhuszet al., 2011). While the gymnosperm fossil record extends back tothe Triassic, research over the past 10 years has made clear that themost recent common ancestors of most gymnosperm genera dateto just 15–25 Ma (Treutlein and Wink, 2002; Won and Renner,2006; Eckert and Hall, 2006; Willyard et al., 2007; Knapp et al.,2007; Qiao et al., 2007; Gernandt et al., 2008; Ickert-Bond et al.,2009; Wei et al., 2010; Lin et al., 2010; Nagalingum et al., 2011;Mao et al., 2010, 2012; Leslie et al., 2012). Northern hemisphereconifer clades (Pinaceae, Cupressoideae) tend to be especially

young, with the largest genera, Abies, Cupressus, Picea, Pinus, andJuniperus, having median node ages under 3.5 Mya (Leslie et al.,2012).

Interspecific relationships in young radiations are usually diffi-cult to resolve due to the incomplete lineage sorting, and an addi-tional challenge in many conifer groups is introgression oforganelle genes following hybridization between co-occurring spe-cies. For spruces (Picea), the focal group here, examples are P. abiesand P. obovata in the Ural Mountains (Krutovskii and Bergmann,1995), P. mariana and P. rubens in eastern North America (Perronand Bousquet, 1997; Perron et al., 2000; Jaramillo-Correa et al.,2003), P. glauca, P. engelmannii, and P. sitchensis in western NorthAmerica (Sutton et al., 1991; Rajora and Dancik, 2000), and P.liki-angensis, P. purpurea, and P. wilsonii from the Qinghai–Tibetan Pla-teau in China (Li et al., 2010; Du et al., 2011; Zou et al., 2012).Nevertheless, phylogenetic studies of the genus Picea have reliedheavily on plastid DNA. The first such study used RFLP data and

718 J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727

inferred a North American origin because three North Americanspecies (P. engelmannii, P. glauca, P. mexicana) branched off basallyin the cladogram (Sigurgeirsson and Szmidt, 1993). A second studyrelied on plastid loci and mitochondrial intron haplotypes (Ranet al., 2006), as did a third, which however, focused more on treeincongruence (Bouillé et al., 2011). In both these latter studies,plastid data yielded a topology in which the western American P.breweriana and P. sitchensis formed a basal grade. This causedRan et al. (2006) to suggest an origin of Picea in North America.They also inferred two migrations from North America to Asia,one from Asia to North America, one from North America to Eur-ope, and one from Asia to Europe. By contrast, a pre-molecularstudy suggested that Picea originated in Asia and expanded toNorth America from there (Wright, 1955). None of the molecularstudies had statistical support for deeper relationships in Picea.

To better resolve relationships among spruces we decided to fo-cus on mitochondrial (mt) introns and, if appropriate, to combinemt data with plastid and nuclear data. Among the most informa-tive regions in conifers are group II introns in the mitochondrialgenes that code for NADH dehydrogenase subunits, specifically thefirst intron of nad5 and the second intron of nad1 (Grivet et al.,1999; Gugerli et al., 2001a, 2001b; Sperisen et al., 2001; Ranet al., 2006; Meng et al., 2007; Bouillé et al., 2011). Using structuralcriteria from the folding of group II introns (Kelchner, 2000), analignment of 17 species of Picea revealed two indels of up to1500 base pairs (bp) and several complex motifs of about 30 bpin domain IV of nad1 intron 2 that were conserved at the level ofspecies groups (Aleksic, 2008; Aleksic and Geburek, 2010, in press).This suggested that these two introns would be phylogeneticallyuseful as long as alignment homology was maintained.

We here study phylogenetic relationships in Picea based onDNA sequences from all but one of its ca. 35 species (Farjón,1990, 2001), with a few problematic species represented by multi-ple individuals. Picea ranges from the Arctic Circle to Mexico andTaiwan (Farjón, 2001; our Fig. 1), with America harboring a sup-posed eight species, Europe (including Turkey and the Caucasus re-gion) to the Ural Mountains three, and Asia c. 23 (Farjón, 2001; theFlora of China lists 16 native species, seven of them endemic; Fuet al., 1999). Of special interest in terms of economic importanceis Norway spruce, Picea abies. Phylogeographic studies of this spe-cies have consistently revealed a genetic gap between its northernand southern European populations, coinciding with the middlePolish disjunction (Schmidt-Vogt, 1974, 1977; Lagercrantz and Ry-man, 1990; Sperisen et al., 2001; Collignon et al., 2002; Tollefsrudet al., 2008; our Fig. 1), which has been attributed to Holocenerange changes one or several millennia ago (Latałowa and vander Knaap, 2006). We thus aimed to infer (i) the age and relation-ships of Norway spruce, (ii) the relationship among the American,Asian, and European spruces, and (iii) possible topological conflictbetween plastid, nuclear, and mitochondrial data in a presumablyyoung conifer radiation (Bouillé and Bousquet, 2005; Crisp andCook, 2011), an assumption we test with strict and relaxed clockmodels, calibrated with various Pinaceae fossils.

2. Materials and methods

2.1. Taxon sampling

We sampled 47 individuals representing all commonly recog-nized species and subspecies (Farjón, 1990, 2001). Not sampled isPicea aurantiaca, an endangered species endemic to West Sichuan,China, which has been treated as a variety of P. asperata (Fu et al.,1999; Eckenwalder, 2009). To represent P.abies, we sampled threeindividuals from the Baltico–Nordic domain, one from the Hercy-no-Carpathian domain, and four from the Alpine domain (domains

are shown in Fig. 1). Farjón (1990) considers the American P. mar-tinezii a synonym of P. chihuahuana; to test this, we included se-quences representing both names, using topotypical material. Wealso included both subspecies of P. engelmannii, the typical subspe-cies and subsp. mexicana (of Schmidt, 1988; Farjón, 1990). The fiveJapanese endemics were all sampled, P. alcoquiana (P. bicolor (Max-im.) Mayr), P. glehnii, P. koyamae, P. maximowiczii, and P. torano (P.polita (Siebold and Zucc.) Carriere). As outgroups, we included rep-resentatives of the other 10 genera of Pinaceae (a family of c. 215species), namely one species each of Abies (fir), Cathaya, Cedrus (ce-dar), Keteleeria, Larix (larch), Nothotsuga (Bristlecone hemlock),Pseudolarix (Golden larch), Pseudotsuga (Douglas fir), and Tsuga(hemlock), and three species from the two subgenera (Pinus andStrobus) of Pinus.

A problem in previous studies seems to have been misidentifiedmaterial or contaminated DNA, which would explain certain con-tradictory findings. For example, some nad1 intron sequences la-beled as representing one species differ strikingly from eachother. Thus, the Picea abies nad1 sequences produced by Sperisenet al. (2001) and Tollefsrud et al. (2008) from 369 natural popula-tions across Europe do not match the P. abies nad1 sequences sub-mitted to GenBank by Bouillé et al. (2011; accessions EF440449,EF440450, EF440451). Similarly, sequences in GenBank of P. asper-ata, P. crassifolia, P. jezoensis, P. omorika, and P. wilsonii do notmatch newly generated sequences from documented plant mate-rial. We have excluded all such doubtful sequences.

2.2. Gene sequencing and alignment

We sequenced mitochondrial introns, several relatively con-served plastid gene regions, and a nuclear gene. Note that spruceplastid DNA is paternally inherited (Sutton et al., 1991; Mogensen,1996; Grivet et al., 1999). From the mitochondrial genome, wesequenced nad5 intron 1 and nad1 intron 2; from the chloroplastgenome, the rbcL gene and the matK gene. We also sequenced threeplastid intergenic spacers, viz. trnL-trnF, trnH-psbA, and trnS-trnG(Ran et al., 2006; Meng et al., 2007; Bouillé et al., 2011; Du et al.,2009, 2011). All data matrices have the same taxonomic represen-tation except for the outgroup genera Abies and Keteleeria, in whichsequences from congeneric species were combined to representthe genus (see Table S1 in Supporting Information). From the nu-clear genome, we sequenced the gene encoding 4-coumarate:coenzyme A ligase (4CL) in the lignin biosynthetic pathway. Partial4CL (600 bp) sequences had already been obtained for 400+ indi-viduals from several Chinese species as part of phylogeographicwork (Li et al., 2010), and 2000 bp sequences were available from117 individuals of most Picea (J.Q. Liu and collaborators, unpub-lished data). Of the 117 sequences, 53 contained no ambiguousbase calls, while the remaining contained a few ambiguous basecalls. For P. asperata, P. crassifolia, P. koraiensis, P. meyeri, P. retro-flexa, and P. schrenkiana, sequences with ambiguous base callswere identified as either the paternal or maternal allele by compar-ing them to 4CL sequences from haploid megagametophyte tissue.For the Japanese species P. alcoquiana we only obtained 816 bp of4CL.

Total genomic DNA was extracted from silica-dried needlesusing a commercial plant DNA extraction kit (NucleoSpin, Mache-rey-Nagel, Düren, Germany). Primers used for polymerase chainreactions (PCRs) and sequencing are listed in Table S2. PCR prod-ucts were purified with the ExoSAP or FastAP clean-up kits (Fer-mentas, St. Leon-Rot, Germany), and sequencing relied on BigDye Terminator kits (Applied Biosystems, Foster City, CA, USA)and an ABI 3130 automated sequencer.

A total of 350 sequences were generated and submitted to Gen-Bank. Table S1 provides species names with authorities, voucherinformation, GenBank accession numbers, and the general

P. mexicana

P. rubensP. breweriana

P. sitchensis

P. engelmannii

P. orientalis

P. obovataP. abies

P. morrisonicola

P. abies

P. omorika

P. alcoquianaP. koyamaeP. maximowicziiP. torano

P. crassifolia

P. schrenkiana

P. smithiana

P. spinulosa P. farreri

P. meyeri

P. koraiensis

P. asperata

P. wilsonii

P. retroflexaP. brachytyla

P. likiangensis

P. purpurea

P. neoveitchii

P. pungens

P. glaucaP. mariana

P. chihuahuana

P. jezoensis

P. glehnii

P. martinezii

Picea abiesBaltico-Nordic domain

Picea abiesHercyno-Carpathian and Alpine domains

Fig. 1. Map of the Northern Hemisphere in equal-area projection (Mollweide) showing the distribution of Picea based on Farjón (1990) and eFloras (2008). Black dashed linesrepresent the Tropic of Cancer and the Arctic Circle.

J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727 719

geographic distribution of each species. Alignment was done man-ually, and for the group II mitochondrial introns nad1 and nad5,conserved domains I, II, III, V, and VI and the variable domain IVwere aligned according to structural criteria (Fig. 2). The combinedalignment contained approximately 50% empty cells, mostly fromoutgroup nad1 and 4CL sequences.

2.3. Phylogenetic analyses

Phylogenetic trees were obtained by Bayesian inference (BI) inBEAST v. 1.7.2 (Drummond et al., 2012) and maximum likelihood(ML) in RAxML (Stamatakis, 2006) with the raxmlGUI v. 1.1 (Silv-estro and Michalak, 2011). Best-fitting models of sequence evolu-tion were estimated using jModelTest v. 0.1.1 (Posada, 2008).Based on the Akaike information criterion, the TIM1 + I + C model(I: invariant sites; C: gamma-distributed rate heterogeneity) best

fit the plastid matrix, the TPM3uf + C model the nuclear matrix,and the SYM + I model the mitochondrial matrix. In all cases, themodel with the highest likelihood score was the general timereversible (GTR) model, and we opted for GTR + C with four cate-gories of rate heterogeneity for the final analyses. Trees wererooted on the Abietoideae subfamily of Pinaceae (Abies, Cedrus,Keteleeria, Nothotsuga, Pseudolarix, Tsuga), which is the sister groupof the Pinoideae (Cathaya, Larix, Picea, Pinus, Pseudotsuga; Gernandtet al., 2008).

Bayesian analyses (using BEAST) used a Yule tree prior, severalMonte Carlo Markov chains (MCMC) of 60 million generationseach, with parameters sampled every 10,000 generations. To takeadvantage of informative insertions and deletions (indels) in themitochondrial introns, we coded indels as characters using simpleindel coding (SIC; Simmons and Ochoterena, 2000) as imple-mented in SeqState (Müller, 2005). Simple indel coding was

Con

serv

edD

omai

n S

ectio

nVa

riabl

eSe

ctio

n D

omai

n

I

IV

G1

A

V

G2

B

G3

II III VI

Section G2: Present in all species, motifs present one variant per speciesG2-1a -----------------TTTGGTCGAGCGTTCATTTATCA 23 bpG2-1b TTTGGTCTCCCTTTTGGTTTGGTCGAGCGTTCATTTATCA 40 bp

G2-2a CCCTTTATTTCTT-----CCCCTCAAA---GGG 25 bpG2-2b CCCTTGATTTCTT-----CCCCTCAAAGGGGGG 28 bpG2-2c CCCTTTATTTCATTTCTTCCCCTCAAA---GGG 30 bpG2-2d CCCTTATAGGCTT------CCCTCAAA----GG 23 bpG2-2e CCCTTATAGGCTT-----CCCCTCAAA---GGG 25 bpG2-2f CCCTTTATTTCTT-----CCCCTCAAAGGGGGG 28 bp

Section B: Present in clades II and III, B-a harbors indels and duplications, either B-b or B-c present per species

B-b CCCTGTATTTATTCCCCTCAAAGGG 25 bpB-c CCCTGTATTTATTCCCCTAAAAGGG 25 bp

Section G3: Present in all species

(b)

(a)

269 bp 185 bp 115 bp

414 bp 76 bp 948 bp

36bp

17bp

493 bp 1333 bp

A1 A2 A3 A4 A5 A6 A7 A8A4 A7 B-a B-bB-c

Picea nad1 intron 2 domain IVSection G1: G1-1 present in all species, G1-2 motifs present one variant per species

G1-2a CGCCCTCGGAGGGCGAGCGTTCGTTTATTACCCTCTCCCT----- 40 bpG1-2b CGCCCTCGGAGGGCGAGCGTTCGTTTATTCCCCTCTCCCT----- 40 bpG1-2c CGCCCTCGGAGGGCGAGCGTTCGTTTATTACCCTCTCCCTTTTCT 45 bp

Section A: Present in clade I and P. morrisonicola, harbors minisatellitesA1-a CACCCATATGATGAGTGAGCGTTAACACCCT 31 bpA1-b CACCCATATGATGAGGGAGCGTTAACACCCT 31 bpA1-c CACCCATATGATGAGKGAGCGTTAACACCCT 31 bpA2-a CACCCATGAATGGATGAGCGACTTCGTTCCT 31 bpA2-b CACC-ATGAATGGATGAGCGACTTCGTTCCT 30 bpA3 CTCCCGTTGGTCGAGAGTTCGCTGCCTCACCCT 33 bpA4 CGCCCTTTTTGGGTCGAGTCACTTAACGTACCT 33 bpA5 CGCCCCTTTTGGTCGAGTCACTTAACGTACCT 32 bpA6-a CGCCTTCCTACCTCAGTCGAGTCACTTAACGTACCT 36 bpA6-b CGCCTTCCTACCTCAGGCGAGTCACTTAACGTACCT 36 bpA7 CACCCATCGGATGGATGAGCGACTTCGTACCT 32 bpA8 CCCCCTCCGTTGTCAGGGGAGCGACTTCGTACCT 34 bp

G1-1 G1-2

* *

***••••**• •

*

Fig. 2. (a) Schematic representation of nad1 intron 2 in Picea. Variable (darker gray) and conserved (lighter gray) intron domains and sections with lengths based on theingroup alignment. (b) Motifs identified within both conserved and variable sections of domain IV. Single asterisks (�) mark motifs present as one variant per species in allspecies, double asterisks (��) mark motifs present in single or multiple copies in all species. Single bullets (�) mark motifs present as one variant per species in some species, adouble bullet (��) mark motifs present in single or multiple copies in some species.

720 J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727

performed after removal of autapomorphic insertions, potentiallyhomoplastic microsatellites, and minisatellites. The binary SIC ma-trix was treated as a separate data partition in BEAST, and a sto-chastic Dollo model was applied to it. This model wasappropriate for the motifs because it is implausible that a motif,once lost, would be regained in the exact original form. Log fileswere analyzed in Tracer v. 1.5 (Rambaut and Drummond, 2007)to assess convergence and to confirm that the effective samplesizes for all parameters were larger than 200, indicating that sta-tionarity had been reached. TreeAnnotator (part of the BEAST pack-age) was used to discard 10% of the saved trees as burn-in and toproduce a maximum clade credibility tree from the remainingtrees. For ML analyses, we first transformed the DNA matrix intoa five-state matrix in which nucleotides were represented by char-acter states 0 to 3, and gaps were treated as a fifth character state.In a second approach, we removed all sites containing gaps but notpoint mutations, and ran the analysis with both a gap-less DNAmatrix and a binary indel matrix, created using SIC.

For Bayesian analyses, posterior probabilities P0.98 were con-sidered good support. For ML analyses, we carried out 100 boot-strap replicates under the same model as tree searches andconsidered P75% good support. Final trees were viewed and edi-ted in FigTree v. 1.3.1 (http://tree.bio.ed.ac.uk).

2.4. Molecular clock dating

For the molecular clock dating, we used the nuclear 4CL exon ma-trix plus the chloroplast genes matK and rbcL. We took out nine Piceaaccessions (P. abies Norway, P. abies Russia, P. abies Austria 1, P. abiesAustria 2, P. abies Serbia 2, P. abies Germany, P. crassifolia, P. koraiensisand P. retroflexa) to reduce the number of very short branches, whichcause problems for molecular-clock modeling and do not add infor-mation. We ran strict clocks and uncorrelated lognormal relaxedclocks, using the same tree prior, substitution model and burn-inas used for tree inference. We also ran analyses with empty

alignments (‘prior-only’ option) and compared the resulting posteriordivergence times to assess the influence of the prior settings

Relative divergence times were translated into absolute timeusing fossil calibrations. The Pinaceae family has an extensive fossilrecord (for a recent review: Klymiuk and Stockey, 2012: theirFig. 5). Fossils used in our study or used in previous Pinaceae datingbut that we consider difficult to assign are listed in Table S3 alongwith detailed justifications for why certain fossils were used as cal-ibrations or instead as outside evidence to provide independent val-idating arguments in the discussion of results. The first calibrationused was a Pseudolarix fossil from the Tsagaan Tsab Formation ofsouthern Mongolia (Krassilov, 1982; Table S3), a locality radiomet-rically dated to the Late Jurassic (156 ± 0.76 Ma; Keller and Hendrix,1997). This provided a minimum age for the stem of Pseudolarix; wechose a gamma prior distribution offset to 156 Myr, with a shapeparameter of 4.0, allowing 90% of the ages to fall between 163.8and 157.4 Myr. Keller and Hendrix’s dating is not for the preciselocation of Krassilov’s fossil (Table S3), but a Pinaceae cone datedto the Lower Kimmeridgian (ca. 155 Ma; Table S3) indicates thatthe Pseudolarix node must be at least that old.

Second, permineralized wood of Pinus (form genus Pinuxylon)recovered from the Aachen Formation of Belgium (Meijer, 2000), da-ted to the Santonian (86.3–83.6 Ma; Cohen et al., 2012), provided aminimum age for the two subgenera of Pinus (cf. Willyard et al.,2007: Pinus; Gernandt et al., 2008: Pinaceae; Lin et al., 2010: Pinaceae;He et al., 2012: Pinus). We chose a gamma prior distribution offset by85 Myr, with a shape parameter of 4.0, which allowed 90% of the agesto fall between 92.8 and 86.4 Myr. The Pseudolarix and Pinuxylon cal-ibrations were tested against each other by applying each in isolation,thereby testing whether they supported each other’s assignment.Next, we applied a strict clock and the mean Pinaceae substitutionrates for matK and rbcL inferred by Gernandt et al. (2008), otherwiseusing the same settings as in the relaxed-clock BEAST runs. Third, wecompared the inferred ages with ages inferred in other studies andwith the fossil record, using fossils not used as constraints.

J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727 721

2.5. Biogeographic analyses

Ancestral area reconstruction (AAR) relied on two approaches:Trait reconstruction under the Markov k-state 1-parameter model(Mk1, a generalization of the Jukes-Cantor model; Lewis (2001)) inMesquite v. 2.75 (Maddison and Maddison, 2011) and the dis-persal-extinction-cladogenesis (DEC) model in Lagrange (likeli-hood analysis of geographic range evolution; Ree and Smith,2008). The AAR analyses were run on the chronogram obtained un-der the fossil-calibrated relaxed clock. The species ranges of the 37Picea species (Farjón, 1990; our Fig. 1) were categorized as follows:(1) western North America including Mexico; (2) eastern NorthAmerica; (3) Europe to the Ural Mountains (including Turkey andthe Caucasus region), (4) Asia; and (5) ambiguous. These 5 catego-ries (areas) provided a balance between having more than one ofthe 37 species in each area and subdividing the overall range ofthe genus into biogeographically meaningful units. Two wide-spread boreal American species (P. glauca and P. mariana) werecoded as ‘ambiguous’ in the Mesquite analysis but assigned to re-gions 1 and 2 in Lagrange. The outgroup ancestral areas were codedas follows: Pinus (110–115 species distributed in North America,Europe, Asia and northern Africa): Europe and Asia (Eckert andHall, 2006) in Lagrange but ambiguous in Mesquite; Cathaya (onesubtropical endemic Chinese species): Once coded as Asian, onceas ambiguous because of North American and Eurasian fossils(Liu and Basinger, 2000); Larix (14 species in North America, Asia,and Europe): Asia (Wei and Wang, 2003); Pseudotsuga (five speciesin Asia and North America): Western North America (Wei et al.,2010); Pseudolarix (one species in subtropical China): Once codedas Asian, once as ambiguous because of fossils in North Americaand Europe (LePage and Basinger, 1995; our Table S3); Abies(�48 species in North America, Europe, and Asia): Ambiguous(Xiang et al., 2009); Cedrus (four species in the circum-Mediterra-nean and western Himalayas): Asia (Qiao et al., 2007); Keteleeria(three species in tropical Asia): Asia; Nothotsuga (one species intropical Asia): Asia; Tsuga (nine species in North America andAsia): Ambiguous (Havill et al., 2008)

Python input scripts for Lagrange were generated using an on-line tool (http://www.reelab.net/lagrange/configurator/index),with the maximum number of ancestral areas constrained totwo. In one run, connection probabilities between geographic areasthrough time (Table S4) were those developed by Moore andDonoghue (2007: Fig. 7) and used also by Havill et al. (2008:Fig. 4), a model that involves 19 time slices. This model is basedon Cenozoic paleoclimates and Mesozoic continental positionsand includes such barriers to dispersal as the emergence and re-treat of the Late Cretaceous Western Interior Seaway separatingeastern and western North America, and the Turgai Strait separat-ing Europe from Asia from the Jurassic into the Oligocene. Themodel also accounts for changing connectivity during the Neogene,which is likely to have affected crown group Picea (Section 3). Inother runs, we explored simpler models that either included 11time slices or no time slice. We also explored the effects of differ-ent outgroup area coding (above).

3. Results

3.1. Sequence characterization and utility

Picea nad1 intron 2 sequences ranged from 1912 nucleotides inP.jezoensis to 3175 in P. morrisonicola, with the range chiefly due totwo large indels flanking a 45-bp conserved region in domain IV ofall spruce species (Fig. 2, sections A and B). Five indels in this do-main described by Sperisen et al. (2001) in P. abies were readilyfound in other species. We also found minisatellite motifs in

Section A (our Fig. 2a and b: A1–A7) that resemble 32-bp and34-bp repeats previously described. All but one Picea have eithersection A or section B in domain IV of nad1. The exception isP. morrisonicola in which the presence of both sections accountsfor its unusually long intron sequence. When aligned with Abies,Cedrus, Cathaya, and Pinus, the nad1 intron matrix contained5677 positions. Removal of all autapomorphic insertions andmicrosatellites resulted in an alignment of 4020 nucleotide charac-ters with 144 separately coded informative indels.

Picea nad5 intron 1 sequences ranged from 1167 to 1198 nucle-otides. Within the highly variable region of this intron (Jaramillo-Correa et al., 2003; Ran et al., 2006) six mitotypes were identified,one of which is new (Fig. 3 inset, mitotype J). Two indels of five andten bp were detected upstream of this region and an autapomor-phic 11-bp duplication unique to P. mariana downstream from it.Alignment of ingroup and outgroup sequences yielded a matrixof 1908 bp. A 16-bp region of the Picea intron sequences couldnot be unambiguously aligned and was excluded from analysis.When autapomorphic insertions, microsatellites, and the ambigu-ous region were removed, the matrix consisted of 1206 nucleotidepositions plus 50 informative indels.

Alternative treatments (Section 2) of the indel characters fromnad1 and nad5 failed to significantly improve topology or statisticalsupport, and completely omitting indel characters yielded topolo-gies that only differed in the positions of a few species at weaklysupported branches near the tips of the tree. Excluding the mini-satellite region in nad1 intron 2 from the matrix demonstrated thatthese minisatellites are informative while not creating spuriousrelationships.

Picea nuclear 4CL gene sequences ranged from 1710 nucleotides(P. glehnii) to 1796 (P. likiangensis var. rubescens), and complete se-quences contained four introns, five exons, and a 3 0 untranslatedregion. An alignment with outgroup genera had a length of 2610nucleotides, of which 1245 were coding.

Plastid matrices consisted of 833 aligned nucleotides of the cod-ing region of matK, 129 bases at the 5 0 end of the trnK intron, and676 nucleotides of rbcL. Picea trnL-trnF sequences ranged from 698to 709 bases, with only short indels of 5 or 6 bases. Ingroup se-quences of the trnH-psbA intergenic spacer were 601 bp in all sam-pled individuals except P. glauca and P. engelmannii, which share asynapomorphic 5-bp insertion. An alignment of the Picea trnS-trnGspacer sequences was 626 nucleotides long.

There were two statistically supported topological incongruenc-es among the data partitions (>87% ML bootstrap support). One in-volved the placement of P. maximowiczii with P. wilsonii, supportedin the nuclear tree but not the plastid and mitochondrial trees. Theother involved the Japanese species P. alcoquiana (with only anincomplete nuclear 4CL sequence) and the Serbian P. omorika. Withthe incomplete nuclear sequence added, these two species werepulled into an otherwise North American clade, while the remain-ing topology stayed unchanged.

3.2. Evolutionary relationships in Picea

A Bayesian tree shows a monophyletic Picea with three geo-graphically coherent clades (labeled I, II, III in Fig. 3). Maximumlikelihood yielded the same major clades, except that P. morrison-icola was placed as sister to clade I, and the western North Amer-ican P. breweriana, which in the Bayesian tree is a member ofclade III, is sister to all other species with weak support (tree notshown). With the exception of P. morrisonicola and P. breweriana,the three clades fit with the distribution of the nad5 mitotypes,which contributed 50 of the total 194 informative indels. Haplo-type D, which lacks insertions in the variable region (Fig. 3 inset),occurs in P. breweriana, Taiwanese P. morrisonicola, and theEurasian clade I, while the remaining North American spruces

0.0090

Picea torano

Abies

Pseudolarix amabilis

Picea breweriana

Picea chihuahuana

Picea likiangensis var. linzhiensis

Picea martinezii

Cedrus atlantica

Picea obovata

Picea morrisonicola

Picea glehnii

Picea farreri

Picea koyamae

Picea glauca

Picea mexicana

Picea abies Serbia 1

Cathaya argyrophylla

Keteleeria Larix gmelinii

Picea alcoquiana

Picea retroflexa

Picea jezoensis

Picea rubens

Picea koraiensis

Picea abies Russia

Picea engelmannii

Picea abies Serbia 2

Picea spinulosa

Picea wilsonii

Picea abies Norway

Picea omorika

Tsuga canadensis

Picea pungens

Picea sitchensis

Picea abies Austria 2

Picea likiangensis var. likiangensis

Picea asperata

Picea meyeri

Picea smithiana

Picea abies

Picea brachytyla Gansu, China

Picea purpurea

Picea neoveitchii

Picea crassifolia

Picea orientalis

Picea likiangensis var. hirtella

Picea abies Germany

Pseudotsuga menziesii

Picea maximowiczii

Picea brachytyla Yunnan, China

Picea abies Austria 1

Picea likiangensis var. rubescens

Nothotsuga longibracteata

Picea mariana

Picea schrenkiana

0.98

1

1

1

1

1

1

1

Pinus pumilaPinus cembra

Pinus sylvestris

11

1

1

1

1

1

1

1

1

0.98

0.98

1

0.98

1

1

1

1

1

0.98

1

DDDDDDDDDDDDDDDDDDDDDD

EEEEEEEEEJEEEDD

Pinoideae

Abietoideae

1

Abietoideae

1

Pinoideae

1

Clade IAsia + 3 European species

Clade IINorth America

Clade IIIAsia +

1 North American species

*

BBAAACCAA

* *

*

*

*

*

**

*

*

Picea abies Serbia 1Picea abies Serbia 2

Picea abies Austria 2Picea abies Germany

Picea abies Austria 1

*

*y

Pi j i*

P. abiesHercyno-Carpathianand Alpine domains

Picea abies RussiaPicea abies Norway

Picea abies** P. abies

Baltico-Nordic domain

ACTTACTT--------------------------GAGTGGCAACTTACTTTACTT---------------------GAGTGGCAACTTACTT---------------------GACTTGAGTGGCAACTTACTT----------GCTT-------GACTTGAGTGGCAACTTACTT-----GACTTGCTT-------GACTTGAGTGGCAACTTACTTTACTT-----GCTTTACTTGAGACTTGAGTGGCA

ACTTACTT--------------------------AAGTGGCAACTTACTT---------------------------AG--GCAACTTACTT----------GCTT------------GAGTGGCAACTTACTT--------------------------GAGTGGCA

DEACBJ

LarixPseudotsugaTsugaAbies, Cathaya,Cedrus, Keteleeria,Nothotsuga, Pinus,Pseudolarix

Piceanad5 intron 1

mitotypes

Fig. 3. Bayesian consensus tree based on 9457 plastid, mitochondrial and nuclear nucleotides including coded indels. Numbers at nodes refer to posterior probability (PP)values P0.98. The asterisks mark nodes that have high support (P0.98 PP) if P. koyamae and P. torano, which lack nad1 and 4CL, are removed from the data matrix; supportfor the remaining nodes remained unaffected. Letters on the right margin represent nad5 intron 1 mitotypes (see inset). Inset shows the six nad5 intron 1 mitotypes detectedin Picea and the homologous sequences in outgroup genera.

722 J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727

(P. chihuahuana, P. engelmannii, P. glauca,P. mariana, P. martinezii,P. mexicana, P. pungens, P. rubens, P. sitchensis) have mitotypes A,B, or C, and the chiefly Asian clade (III) has mitotypes D, E, andJ. All members of clade I lack section B of nad1, while membersof clades II and III except P. morrisonicola lack section A.

Clade I, centered in Asia with three European species, includesthe five Japanese endemics, P. alcoquiana, P. glehnii, P. koyamae, P.maximowiczii, and P. torano, which did not originate from a singleradiation. Norway spruce, P. abies, is not monophyletic because itsBaltico–Nordic populations (Fig. 1) are more closely related toAsian species of the P. asperata complex (P. asperata, P. crassifolia,P. koraiensis, P. koyamae, P. meyeri, P. obovata, and P. retroflexa)than to southern Hercyno-Carpathian and Alpine populations ofNorway spruce. In clade II, from North America, the western P.martinezii is the first to diverge, followed by the nine remainingspecies. The widespread P. mariana and the eastern P. rubens bothpossess mitotype C; the widespread P. glauca and the western P.engelmannii both possess mitotype B. The endemic P. mexicanaand the more widespread P. pungens are sister species; the othertwo Mexican species (P. chihuahuana, P. martinezii) are not closelyrelated. Clade III contains another instance of species non-mono-phyly because P. brachytyla material from Gansu, China is more

closely related to P. spinulosa than it is to P. brachytyla from Yun-nan (Fig. 3).

3.3. Divergence times and substitution rates in Pinaceae and Picea

A chronogram from a relaxed clock calibrated with Pinuxylonand Pseudolarix fossils is shown in Fig. 4. Results from BEAST runswith an empty alignment revealed no contradictions among theprior constraints and showed that the posterior age distributionswere substantially influenced by the signal in the data (Fig. S1).With the Pinuxylon and Pseudolarix constraints (gray column in Ta-ble 1, which also shows the 95% confidence intervals for inferredages), all other Pinaceae divergence events are inferred to timespredating their earliest reliable fossil record (Table S3). For exam-ple, the oldest conclusive evidence of Larix (Middle to Late Eocene,which would be 44–34 Ma; LePage and Basinger, 1991) post-datesthe inferred time of its divergence from Pseudotsuga, 75 Ma.

The Pinaceae crown group emerged in the Upper Triassic, andthe divergence between Picea and Pinus occurred in the LowerJurassic (c. 180 Ma; but inferring these deep splits was not the focusof our study). The Picea crown group emerged in the mid-Oligocene,at 28 Ma (95% CI: 37–21 Ma). Most Picea speciation events date to

Tabl

e1

Div

erge

nce

tim

esan

dsu

bsti

tuti

onra

tes

unde

rst

rict

and

rela

xed

cloc

km

odel

sus

ing

alte

rnat

ive

foss

ilca

libra

tion

poin

tsor

prev

ious

lyin

ferr

edPi

nace

aesu

bsti

tuti

onra

tes

( Ger

nand

tet

al.,

2008

).N

ode

ages

are

med

ian

valu

esin

mill

ions

ofye

ars

wit

h95

%co

nfide

nce

inte

rval

ssh

own

belo

w.N

ucle

otid

esu

bsti

tuti

onra

tes

are

insu

bsti

tuti

ons/

site

/mill

ion

year

s.M

arke

din

pale

gray

are

the

resu

lts

obta

ined

unde

rth

epr

efer

red

calib

rati

onsc

hem

e.

J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727 723

the Miocene. The North American clade (II) first diversified 20(27–14) Ma, the Eurasian clade (I) 23 (32–15) Ma. The Hercyno-Carpathian and Alpine (southern European) lineage of P. abies splitfrom the Baltico–Nordic (northern) lineage 6 Ma.

Relaxed clock models with the preferred fossil calibrationscheme yielded a matK substitution rate (subs/site/Mya) of1.92 � 10�4 and an rbcL rate of 1.08 � 10�4 (Table 1, rates ob-tained with either one of the two calibration fossils are alsoshown). We also tested the matK and rbcL substitution rates in-ferred by Gernandt et al. (2008; our Table 1).

3.4. Ancestral area reconstruction

Results of the two AAR approaches, i.e., reconstruction underthe Mk1 model in Mesquite and reconstruction under the DECmodel in Lagrange, are shown in Table S5. A DEC model with19 time slices (Section 2) had a global likelihood of 71.11, andone with 11 time slices 71.27, and without time slices 70.39.We therefore preferred the simplest model. Likelihood values ofbiogeographic scenarios for ingroup nodes of interest were notsignificantly affected by the number of time slices or ambiguityin the coding of outgroup ancestral areas (Section 2; Table S5).

Using the topology obtained in the Bayesian runs, the most re-cent common ancestor of today’s species of Picea is inferred tohave lived in Asia; using the ML topology in which P. brewerianais sister to all other species yields an ambiguous reconstructionfor the root as Asia/North America. Using either tree, the stemlineage of the North American clade (II) apparently entered NorthAmerica before 20 Ma (Fig. 4, Table 1). Under the Mk1 model, itcame from Asia (likelihood of 0.86), while under the 1-time-sliceDEC model it lived in Asia and western North America (likelihoodof 0.92). The stem lineage of the Asian/North American clade (III),under the Mk1 model, may have occurred in Asia (likelihood of0.87), while under the 1-time-slice DEC model it occurred in Asiaand North America (likelihood of 0.99). The crown node of clade I(the Eurasian clade) under both models existed in Asia (likeli-hoods of 0.97 under Mk1 and 0.90 under DEC). The ancestral areaof the clade comprising P. omorika, endemic in Serbia, and P. ori-entalis endemic in the Caucasus region (Figs. 1 and 3), is inferredas Asia (likelihood of 0.83).

4. Discussion

By using signal from large motifs in mitochondrial introns weobtained a well-resolved phylogeny of Picea in which most nodeshave strong statistical support. Different from previous molecularphylogenies of the genus, we find three geographically coherentmajor groups, an American clade with nine species, a predomi-nantly Asian clade, and a Eurasian clade. Surprisingly, our dataalso suggest the paraphyly of Norway spruce. We now discussthese results in turn.

4.1. Mitochondrial indels under a stochastic Dollo model

Group II introns in the mitochondrial NADH genes, specificallythe first intron of nad5 and the second intron of nad1, have longbeen useful in conifer phylogenetics (Grivet et al., 1999; Gugerliet al., 2001a, 2001b; Sperisen et al., 2001; Ran et al., 2006; Menget al., 2007; Bouillé et al., 2011). However, aligning length muta-tions in variable domains of introns or spacers is not straightfor-ward (Kelchner, 2000). Previous work by Aleksic (2008) hadrevealed two indels of up to 1500 bp and several complex motifsof about 30 bp in domain IV of nad1 intron 2 that were conservedat the level of species groups, suggesting that introns would bephylogenetically useful. We applied simple indel coding, which

724 J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727

yielded a binary matrix on which we then used the stochastic Dollomodel, appropriate for characters that are unlikely to be regainedonce lost, which likely is true of the long motifs coded here. Theimpact of indel characters was assessed in analyses with and with-out coded mitochondrial indels. Results showed that the indelcharacters support species groups also seen without coded indels,albeit without high statistical support.

Population genetic studies of spruces have used some of thesame mt introns (albeit aligned without taking advantage of sec-ondary structure) to infer progenitor–derivative relationships. Forexample, nad1 and nad5 (plus sequence-tagged-site markers ofexpressed genes) were used to infer an ancestor–descendant rela-tionship between P. mariana and P. rubens because the latterspecies has a subset of the diversity found in the former (Perronet al., 2000; Jaramillo-Correa and Bousquet, 2003). In our tree dia-gram (Fig. 2), they are sister species and uniquely share nad5 mito-type C.

4.2. Recent origin and diversification of Picea

Spruces provide yet another example of a relatively youngdiversification atop an unbranched stem lineage. Such stemlineages are thought to reflect major extinctions (Harvey et al.,1994). In gymnosperms, extinctions occurred when the global tem-peratures declined sharply at the end of the Eocene, followed by re-radiations during the late Oligocene and early Miocene warming(Crisp and Cook, 2011; Nagalingum et al., 2011; Leslie et al.,2012; Mao et al., 2012). The three regional radiations in spruces(clades I, II, III) all fall in this warmer period, between 25 and20 Ma (Zachos et al., 2001).

Regardless of calibration method and clock model, the most re-cent common ancestor of living Picea species was dated to about28 Ma (37.2–20.6 Ma, Table 1). Previous studies dated the Piceacrown group to c. 30 Ma (using two species selected to span thedeepest divergence; Lin et al., 2010) or to 13–20 Ma (with threespecies, Bouillé and Bousquet, 2005). Leslie et al. (2012: Fig. S4C)included most species and inferred an age of c. 23 Ma. The age ob-tained by Crisp and Cook (2011) was 5.8 Ma, which seems tooyoung. Coalescence dating that assumed an average generationtime in Picea of 50 years and effective population sizes between100,000 and 200,000 individuals yielded divergence times thatare younger than those obtained here. Thus, an analysis of 12–16nuclear loci dated the split between P. likiangensis and P. schrenki-ana to 10.9 Ma (Li et al. 2010), while we estimated 19.7 Ma for thisnode. The same study dated the split between P. purpurea and P.schrenkiana to 3.2 Ma, while we dated this split to 11.5 Ma. Piceageneration times are poorly known and may be 25 or 50 years(Chen et al., 2010). The different effective population sizes andmutation rates of the nuclear and organellar genomes also meanthat coalescence ages estimated from nuclear data are severaltimes higher than those inferred from plastid data (Li et al.,2012), a point that may be insufficiently accommodated in ourmolecular clocks. The Pinaceae plastid substitution rates of 1.0 to1.9 � 10-4 subs/site/Mya years obtained here are identical to plas-tid rates inferred for other woody angiosperms, for example, le-gumes (Wojciechowski, 2005; Scherson et al., 2008), but areslower than the 1.0 to 2.8 subs/site/Mya rates inferred by Gernandtet al. in a Pinus-focused study (2008).

Wright’s (1955) observation that morphological variation in allof Picea is comparable to that found in species groups of Pinus fitswith its young age. Cones of the Early Cretaceous P.burtonii (Fig. 4inset), the Oligocene P.diettertiana (Miller, 1970), and extantspecies show little change during the last 136 Ma of the genus(Klymiuk and Stockey, 2012). While this stasis may facilitate thegeneric placement of fossils, it reduces the phylogenetic utility ofmorphological characters in living Picea. Thus, Farjón’s taxonomic

treatment (1990) recognized P. crassifolia as a distinct species,while Eckenwalder (2009) treats this entity as a variety of P.schrenkiana (in clade III in Fig. 3). In our tree it is a member ofthe P. asperata complex (clade I in Fig. 3), rejecting the latter place-ment. Conversely, Eckenwalder’s (2009) recognition of P. mexicanaas a species is supported (Fig. 3).

4.3. Biogeography of the major clades of Picea

Using the preferred topology shown in Fig. 4, in which P. brewe-riana is placed in clade III, we inferred that the immediate ancestorof the living species of Picea lived in Asia as first proposed byWright (1955); with the ML topology in which P. breweriana is sis-ter to all other species the root is reconstructed as Asia/NorthAmerica. Three previous molecular studies (Sigurgeirsson andSzmidt, 1993; Ran et al., 2006; Bouillé et al., 2011) instead inferredan evolution of Picea in North America, based on the finding that inplastid gene trees western North American species were the first todiverge. Ran et al. (2006) also inferred a back-dispersal from Asia toNorth America (ancestor of P. chihuahuana) and one from NorthAmerica to Europe (ancestor of P. omorika). This last event theythemselves considered ‘‘nearly impossible’’ (Ran et al., 2006:414). Our reconstruction implies two entries from Asia into NorthAmerica, likely through Beringia between 25 and 20 Ma, and noback-dispersal.

One of the two entries involves the stem lineage of the NorthAmerican clade (II), the other the ancestor of clade III, which com-prises only Asian species and the North American P. breweriana.The latter species is genetically and morphologically highly distinct(Schmidt-Vogt, 1977; Sigurgeirsson and Szmidt, 1993; Weng andJackson, 2000; Ledig et al., 2004; Ran et al., 2006; Bouillé et al.,2011). With Bayesian inference it formed an early-branching line-age in clade III, albeit not as sister to all other spruces as found bySigurgeirsson and Szmidt (1993), Ran et al. (2006), and Bouilléet al. (2011). With maximum likelihood, however, P. brewerianawas placed as sister to all other spruces, but with low statisticalsupport (tree not shown). That P. breweriana might be part of asouthern Asian clade would fit with it being ‘‘most similar to somesouth Chinese species [and it likely being] an offshoot of an earliermigration than the one that give rise to the other spruces of north-west America’’ (Wright, 1955: 328).

The regions of highest Picea diversity are the western Cordilleraof the Rocky Mountains and the Qinghai–Tibetan Plateau (QTP).Western North America (including Mexico) has more distinctspruces than recognized before this study. Thus, P. chihuahuanaand P. martinezii are neither close to Asian species (Sigurgeirssonand Szmidt, 1993; Ran et al., 2006; Bouillé et al., 2011) nor closelyrelated to P. glauca [contra Schmidt-Vogt (1977) and Farjón(1990)]. Mexican spruce (P. mexicana, Fig. 3) is not a subspeciesof P. engelmannii [contra Schmidt (1988) and Farjón (1990,2001)], but instead a sister to the western North American P. pun-gens. For the QTP region, studies at the intra- and interspecific lev-els have indicated high levels of population differentiation,probably attributable to the topography of the QTP, where deepvalleys and high mountain peaks act as barriers to gene flow (Liet al., 2010). Clade III (Fig. 3), which comprises nine to ten speciesin the QTP region, diversified around 19.8 Ma. Several QTP speciesshare mt and cp haplotypes (Du et al., 2011; Zou et al., 2012), per-haps because it takes 18.75 Mya for 95% of organelle genes to be-come species-specific, assuming a generation time of 25 yearsand an effective population size of 150,000 (Chen et al., 2010).Assuming a generation time of 50 years, it would take 37.5 Mya.

An unexpected finding is the grouping of the Serbian endemic P.omorika, the Caucasian P. orientalis, and the two Japanese endemicsP. alcoquiana and P. maximowiczii, a clade that has 1.0 PP (Fig. 3).Since the 4CL sequence of P. alcoquiana is incomplete, this clade

050100150200

Picea martinezii Picea likiangensis var. linzhiensis

Picea alcoquiana

Picea likiangensis var. hirtella

Picea torano

Picea omorika

Picea koyamae

Picea glauca

Cathaya argyrophylla

Picea schrenkiana

Picea likiangensis var. rubescens

Picea brachytyla Gansu, China

Picea rubens Picea mexicana

Picea asperata

Pinus cembra

Picea meyeri

Picea chihuahuana

Picea spinulosa

Picea jezoensis

Picea orientalis

Pseudolarix amabilis

Picea farreri

Picea abies Norway

Larix gmelinii

Cedrus atlantica

Picea maximowiczii

Picea breweriana

Pseudotsuga menziesii

Picea smithiana

Picea wilsonii Picea purpurea

Tsuga canadensis

Pinus sylvestris

Keteleeria

Pinus pumila

Picea neoveitchii

Nothotsuga longibracteata

Picea mariana

Picea abies Serbia

Picea glehnii

Picea morrisonicola

Abies

Picea brachytyla Yunnan, China

Picea engelmannii Picea sitchensis Picea pungens

Picea obovata

Paleogene NeogeneCretaceous

Oli.Eoc.

Jurassic

Pal.UpperUpperUpper

Triassic

Mio. PO

Picea likiangensis var.

LowerLower Middle

Picea burtonii136 Ma

Epoch

Period

Ma

AsiaEurope

EasternNorth America

WesternNorth America

Clade I

Clade II

Clade III

Picea alcoquiana

Picea torano

Picea omorika

Picea koyamae

Picea asperataPicea meyeri

Picea jezoensis

Picea orientalis

Picea abies Norway

Picea maximowiczii

Picea abies Serbia

Picea glehnii

Picea obovata

Clade IP

P

P

P

PP

P

P

P

P

P

P

P

Picea likiangensis var. linzhiensis

Picea likiangensis var. hirtella

Picea schrenkiana

Picea likiangensis var. rubescens

Picea brachytyla Gansu, ChinaPicea spinulosa Picea farreri

Picea breweriana

Picea smithiana

Picea wilsonii Picea purpurea

Picea neoveitchii

Picea morrisonicola

Picea brachytyla Yunnan, China

Picea likiangensis var.

Clade III

P

P

P

P

PPP

P

P

PP

P

P

P

P

Clade IIPicea martinezii

Picea glauca

Picea rubens Picea mexicana

Picea chihuahuanaPicea mariana

Picea engelmannii Picea sitchensis Picea pungens

P

P

PP

PP

PPP

likiangensis

Fig. 4. Chronogram of Pinaceae based on two fossil constraints (stars) and ancestral area reconstruction (AAR) without constrained connection probabilities. Gray barsindicate the 95% highest posterior probability of the node age estimates. Pal. = Paleocene, Eoc. = Eocene, Oli. = Oligocene, PO = Pliocene. Inset image shows Picea burtonii fromthe Early Cretaceous (c. 136 Ma) of Vancouver Island (Klymiuk and Stockey, 2012).

J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727 725

needs further investigation. In the chronogram, it is 16.5 Mya old(Fig. 4), the time of the mid-Miocene climatic optimum (17–15 Ma; Zachos et al., 2001). All four species currently occur at highelevations (700–2200 m) where annual precipitation ranges from1000 to 2500 mm and winters are cold, frequently with snow(Farjón, 1990). Since Picea is generally adapted to boreal andmontane habitats, the increasing seasonality and aridity that aroseduring the late Miocene (10–5 Ma) may have led to widespreadextinction of Picea in the mid latitudes of Eurasia. This greaterseasonality and aridity also led to the fragmentation of a once con-tinuous ruminant ungulate fauna ranging from Iberia to centralAsia during the early Miocene (23–15 Ma; Fortelius et al., 2002).

4.4. Species non-monophyly: Convergence, incomplete lineage sorting,or interspecific introgressions?

At least three species, Norway spruce, P. abies, the ChineseP. brachytyla, and the American P. engelmannii are not monophy-letic in their current circumscription (Farjón, 1990, 2001). This

may have several explanations. First, different populations of thesespecies could have originated from different ancestors but bemorphologically so convergent that they look like a single species.This may the case for P. mexicana, which closely resemblesP. engelmannii. Second, because some (but not all) spruce speciesoccurring in the same geographic region diversified recently,incomplete lineage sorting and hybridization might cause conflictbetween nuclear and organellar gene trees. The distributionalranges of spruce species were greatly affected by the last glacialperiod from approximately 110,000–10,000 years ago (Tollefsrudet al., 2008; Parducci et al., 2012), and range retreats and expan-sions may have caused secondary contact of previously isolatedincipient species and led to introgression (Du et al., 2011). Thismay have occurred in P. abies and P. brachytyla.

Picea abies is a highly polymorphic species, and no consensus onits circumscription has been reached. Thus, P. obovata, which ispart of the P. asperata complex (P. abies, P. asperata, P. crassifolia,P. koraiensis, P. koyamae, P. meyeri, and P. retroflexa; Fig. 1), hasbeen considered a ‘‘minor species’’ near P. abies (Farjón, 1990:

726 J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727

229), an ‘‘eastern form of P. abies’’ (Schmidt-Vogt, 1974: 195, 1977:21), or a distinct species (Farjón, 2001; Eckenwalder, 2009). Studiesof P. abies nuclear markers and organelle DNA have consistently re-vealed a clear separation between the northern (Baltico–Nordic;Figs. 1 and 3) and southern (Hercyno-Carpathian, Alpine) popula-tions (e.g., Lagercrantz and Ryman 1990; Sperisen et al., 2001; Col-lignon et al., 2002; Tollefsrud et al., 2008). In our data, thedistinctness of the southern populations is seen in seven non-homoplastic point mutations, in indel motifs in nad1 (Sperisenet al., 2001; our Fig. 2).

So far, studies of P. abies phylogeography (Tollefsrud et al.,2008; Parducci et al., 2012) have either been rooted with P. obov-ata, which in cladistic terms is part of the ingroup, or also withNorth American P. glauca, which is too distant to infer the directionof population expansion. Proper rooting would require inclusion ofthe Asian species P. glehnii or P. jezoensis (Fig. 2). If one were to di-vide P. abies into two species, northern P. abies (including all namespublished in the P. asperata complex) would remain P. abies sensustricto because Linnaeus described Scandinavian material. Regard-less of naming issues, a possible non-monophyly of P. abies meansthat the history of this important species, the genome of which hasjust become available (Nystedt et al., 2013), has not been fullyunderstood. Our dating shows that the northern and southern lin-eages of P. abies separated 6 Ma, distinctly prior to the ice ages(2.5 Ma). Pleistocene refugia are thus unlikely to be the explana-tion for the relationships between the European species (P. abies,P. omorika, P. orientalis) and their East Asian relatives revealed inthis study. However, the interspecific introgressions documentedin Picea (Section 1), incomplete lineage sorting in large outbreedingpopulations, and differences in the effective sizes and mutationrates of the sampled DNA regions, mean that our findings concern-ing species relationships and divergence times need further confir-mation, especially based on larger genomic data.

Acknowledgments

We thank U. Pietzarka at the Tharandt Botanical Garden, B. Ja-quish at the Kalamalka Research Station, and M. Tollefsrud at theNorwegian Forest and Landscape Institute for samples; M. Silberfor assistance in the laboratory; B. LePage for discussion of fossils;B. Moore for a matrix of past continental connection probabilities;and G. Holman, A. Klein, and C. Campbell for information on Piceaaccessions sampled in earlier studies. Financial support came fromthe Ecology, Evolution, and Systematics program of Munich Uni-versity, the Ministry of Education and Science of the Republic ofSerbia (Research Grant 173030 to JA), and the National Natural Sci-ence Foundation of China (Research Grant 30930072 to LJQ) andthe Key Project of International Collaboration Program, the Minis-try of Science and Technology of China (Research Grant2010DFB63500 to LJQ).

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ympev.2013.07.004.

References

Aleksic, J.M., 2008. Genetic structure of natural populations of Serbian spruce [Piceaomorika (Panc.) Purk.]. PhD thesis, University of Natural Resources and AppliedLife Sciences, Vienna, Austria. <https://zidapps.boku.ac.at/abstracts/download.php?dataset_id=6870andproperty_id=107>.

Aleksic, J.M., Geburek, T., 2010. Mitochondrial DNA reveals complex geneticstructuring in a stenoendemic conifer Picea omorika [(Panc.) Purk.] caused byits long persistence within the refugial Balkan region. Plant Syst. Evol. 285, 1–11.

Aleksic, J.M., Geburek, T., in press. Quaternary population dynamics of an endemicconifer, Picea omorika, in its last refugium in the Balkans, and conservationimplications. Conservation Genetics.

Bouillé, M., Bousquet, J., 2005. Trans-species shared polymorphisms at orthologousnuclear gene loci among distant species in the conifer Picea (Pinaceae):implications for the long-term maintenance of genetic diversity in trees. Am.J. Bot. 91, 63–73.

Bouillé, M., Senneville, S., Bousquet, J., 2011. Discordant mtDNA and cpDNAphylogenies indicate geographic speciation and reticulation as driving factorsfor the diversification of the genus Picea. Tree Genet. Genomes 7, 469–484.

Chen, J., Källman, T., Gyllenstrand, N., Lascoux, M., 2010. New insights on thespeciation history and nucleotide diversity of three boreal spruce species and aTertiary relict. Heredity 104, 3–14.

Christenhusz, M.J.M., Reveal, J.L., Farjon, A., Gardner, M.F., Mill, R.R., Chase, M.W.,2011. A new classification and linear sequence of extant gymnosperms.Phytotaxa 19, 5570.

Cohen, K.M., Finney, S., Gibbard, P.L., 2012. International chronostratigraphic chart.<http://www.stratigraphy.org/ics%20chart/ChronostratChart2012.pdf>.

Collignon, A.-M., Van de Sype, H., Favre, J.-M., 2002. Geographical variation inrandom amplified polymorphic DNA and quantitative traits in Norway spruce.Can. J. Forest Res. 32, 266–282.

Crisp, M.D., Cook, L.G., 2011. Cenozoic extinctions account for the low diversity ofextant gymnosperms compared with angiosperms. New Phytol. 192, 997–1009.

Drummond, A.J., Suchard, M.A., Xie, D., Rambaut, A., 2012. Bayesian phylogeneticswith BEAUti and the BEAST 1.7. Mol. Biol. and Evol. first published online:February 25, 2012, doi: 10.1093/molbev/mss075.

Du, F.K., Petit, R.J., Liu, J.Q., 2009. More introgression with less gene flow:chloroplast vs. mitochondrial DNA in the Picea asperata complex in China, andcomparison with other conifers. Mol. Ecol. 18, 1396–1407.

Du, F.K., Peng, X.L., Liu, J.Q., Lascoux, M., Hu, F.S., Petit, R.J., 2011. Direction andextent of organelle DNA introgression between two spruce species in theQinghai-Tibetan Plateau. New Phytol. 192, 1024–1033.

Eckenwalder, J.E., 2009. Conifers of the World: The Complete Reference, first ed.Timber Press, Portland.

Eckert, A.J., Hall, B.D., 2006. Phylogeny, historical biogeography, and patterns ofdiversification for Pinus (Pinaceae): phylogenetic tests of fossil-basedhypotheses. Mol. Phylogenet. Evol. 40, 166–182.

eFloras, 2008. Missouri Botanical Garden, St. Louis, MO and Harvard UniversityHerbaria, Cambridge, MA. Published on the Internet <http://www.efloras.org>[accessed 18 June 2012].

Farjón, A., 1990. Pinaceae: Drawings and Descriptions of theGeneraAbies,Cedrus,Pseudolarix,Keteleeria,Nothotsuga,Tsuga,Cathaya,Pseudotsuga,Larix and Picea. Koeltz Scientific Books, Königstein.

Farjón, A., 2001. World Checklist and Bibliography of Conifers, second ed. RoyalBotanic Gardens, Kew, London.

Fortelius, M., Eronen, J., Jernvall, J., Liu, L., Pushkina, D., Rinne, J., Tesakov, A.,Vislobokova, I., Zhang, Z., Zhou, L., 2002. Fossil mammals resolve regionalpatterns of Eurasian climate change over 20 million years. Evol. Ecol. Res. 4,1005–1016.

Fu, L., Li, N., Mill, R.R., 1999. Picea. In: Wu, Z.-Y., Raven, P.H. (Eds.), Flora of China (4).Science Press, Beijing and Missouri Botanical Garden Press, St. Louis, pp. 25–32.

Gernandt, D.S., Magallón, S., López, G.G., Flores, O.Z., Willyard, A., Liston, A., 2008.Use of simultaneous analyses to guide fossil-based calibrations of Pinaceaephylogeny. Int. J. Plant Sci. 169, 1086–1099.

Grivet, D., Jeandroz, S., Favre, J.M., 1999. Nad1 b/c intron polymorphism revealsmaternal inheritance of the mitochondrial genome in Picea abies. Theor. Appl.Genet. 99, 346–349.

Gugerli, F., Senn, J., Anzidei, M., Madaghiele, A., Büchler, U., Sperisen, C., Vendramin,G.G., 2001a. Chloroplast microsatellites and mitochondrial nad1 intron 2sequences indicate congruent phylogenetic relationships among Swiss stonepine (Pinus cembra), Siberian stone pine (Pinus sibirica), and Siberian dwarf pine(Pinus pumila). Mol. Ecol. 10, 1489–1497.

Gugerli, F., Sperisen, C., Büchler, U., Brunner, I., Brodbeck, S., Palmer, J.D., Qiu, Y.L.,2001b. The evolutionary split of Pinaceae from other conifers: evidence from anintron loss and a multigene phylogeny. Mol. Phylogenet. Evol. 21, 167–175.

Harvey, P.H., May, R.M., Nee, S., 1994. Phylogenies without fossils. Evolution 48,523–529.

Havill, N.P., Campbell, C.S., Vining, T.F., LePage, B., Bayer, R.J., Donoghue, M.J., 2008.Phylogeny and biogeography of Tsuga (Pinaceae) inferred from nuclearribosomal ITS and chloroplast DNA sequence data. Syst. Bot. 33, 478–489.

He, T., Pausas, J.G., Belcher, C.M., Schwilk, D.W., Lamont, B.B., 2012. Fire-adaptedtraits of Pinus arose in the fiery Cretaceous. New Phytol. 194, 751–759.

Ickert-Bond, S.M., Rydin, C., Renner, S.S., 2009. A fossil-calibrated relaxed clock forEphedra indicates an Oligocene age for the divergence of Asian and New Worldclades, and Miocene dispersal into South America. J. Syst. Evol. 47, 444–456.

Jaramillo-Correa, J.P., Bousquet, J., 2003. New evidence from mitochondrial DNA of aprogenitor-derivative species relationship between black spruce and red spruce(Pinaceae). Am. J. Bot. 90, 1801–1806.

Jaramillo-Correa, J.P., Bousquet, J., Beaulieu, J., Isabel, N., Perron, M., Bouillé, M.,2003. Cross-species amplification of mitochondrial DNA sequence-tagged-sitemarkers in conifers: the nature of polymorphism and variation within andamong species in Picea. Theor. Appl. Genet. 106, 1353–1367.

Kelchner, S.A., 2000. The evolution of noncoding chloroplast DNA and its applicationin plant systematics. Ann. MO Bot. Gard. 87, 482–498.

Keller, A.M., Hendrix, M.S., 1997. Paleoclimatologic analysis of a Late Jurassicpetrified forest, southeastern Mongolia. Palaios 12, 282–291.

J.D. Lockwood et al. / Molecular Phylogenetics and Evolution 69 (2013) 717–727 727

Klymiuk, A.A., Stockey, R.A., 2012. A Lower Cretaceous (Valanginian) seed coneprovides the earliest fossil record for Picea (Pinaceae). Am. J. Bot. 99, 1069–1082.

Knapp, M., Mudaliar, R., Havell, D., Wagstaff, S.J., Lockhart, P.J., 2007. The drowningof New Zealand and the problem of Agathis. Syst. Biol. 56, 862–870.

Krassilov, V., 1982. Early Cretaceous flora of Mongolia. Palaeontogr. Abt. B. 181, 1–43.

Krutovskii, K.V., Bergmann, F., 1995. Introgressive hybridization and phylogeneticrelationships between Norway, Picea abies (L.) Karst., and Siberian, P. obovataLedeb., spruce species studied by isozyme loci. Heredity 74, 464–480.

Lagercrantz, U., Ryman, N., 1990. Genetic structure of Norway spruce (Picea abies):concordance of morphological and allozymic variation. Evolution 44, 38–53.

Latałowa, M., van der Knaap, W.O., 2006. Late Quaternary expansion of Norwayspruce Picea abies (L.) Karst. in Europe according to pollen data. Quaternary Sci.Rev. 25, 2780–2805.

Ledig, F.T., Hodgkiss, P.D., Krutovskii, K.V., Neale, D.B., Eguiluz-Piedra, T., 2004.Relationships among the spruces (Picea, Pinaceae) of southwestern NorthAmerica. Syst. Bot. 29, 275–295.

LePage, B.A., Basinger, J.F., 1991. A new species of Larix (Pinaceae) from the earlyTertiary of Axel Heiberg Island, Arctic Canada. Rev. Palaeobot. Palynol. 70, 89–111.

LePage, B.A., Basinger, J.F., 1995. History of the genus Pseudolarix Gordon (Pinaceae).Int. J. Plant Sci. 156, 910–950.

Leslie, A.B., Beaulieu, J.M., Rai, H.S., Crane, P.R., Donoghue, M.J., Mathews, S., 2012.Hemisphere-scale differences in conifer evolutionary dynamics. Proc. Natl.Acad. Sci. USA 109, 1–5.

Lewis, P.O., 2001. A likelihood approach to estimating phylogeny from discretemorphological character data. Syst. Biol. 50, 913–925.

Li, Y., Stocks, M., Hemmilä, S., Källman, T., Zhu, H., Zhou, Y., Chen, J., Liu, J., Lascoux,M., 2010. Demographic histories of four spruce (Picea) species of the Qinghai-Tibetan Plateau and neighboring areas inferred from multiple nuclear loci. Mol.Biol. Evol. 27, 1001–1014.

Li, Z., Zou, J., Mao, K., Lin, K., Li, H., Liu, J., Källman, T., Lascoux, M., 2012. Populationgenetic evidence for complex evolutionary histories of four high altitudejuniper species in the Qinghai-Tibetan Plateau. Evolution 66, 831–845.

Lin, C.-P., Huang, J.-P., Wu, C.-S., Hsu, C.-Y., Chaw, S.-M., 2010. Comparativechloroplast genomics reveals the evolution of Pinaceae genera and subfamilies.Genome Biol. Evol. 2, 504–517.

Liu, Y.S., Basinger, J.F., 2000. Fossil Cathaya (Pinaceae) pollen from the CanadianHigh Arctic. Int. J. Plant Sci. 161, 829–847.

Maddison, W.P., Maddison, D.R. 2011. Mesquite: A Modular System for EvolutionaryAnalysis. Version 2.75 <http://mesquiteproject.org>.

Mao, K., Hao, G., Liu, J., Adams, R.P., Milne, R.I., 2010. Diversification andbiogeography of Juniperus (Cupressaceae): variable diversification rates andmultiple intercontinental dispersals. New Phytol. 188, 254–272.

Mao, K., Milne, R.I., Zhang, L., Peng, Y., Liu, J., Thomas, P., Mill, R.R., Renner, S.S., 2012.Distribution of living Cupressaceae reflects the breakup of Pangea. Proc. Natl.Acad. Sci. USA 109, 7793–7798.

Meijer, J., 2000. Fossil woods from the Late Cretaceous Aachen Formation. Rev.Palaeobot. Palyno. 112, 297–336.

Meng, L., Yang, R., Abbott, R.J., Miehe, G., Hu, T., Liu, J., 2007. Mitochondrial andchloroplast phylogeography of Picea crassifolia Kom. (Pinaceae) in the Qinghai-Tibetan Plateau and adjacent highlands. Mol. Ecol. 16, 4128–4137.

Miller, C.N., 1970. Picea diettertiana, a new species of petrified cones from theOligocene of western Montana. Am. J. Bot. 57, 579–585.

Mogensen, H.L., 1996. The hows and whys of cytoplasmic inheritance in seed plants.Am. J. Bot. 83, 383–404.

Moore, B.R., Donoghue, M.J., 2007. Correlates of diversification in the plant cladeDipsacales: geographic movement and evolutionary innovations. Am. Nat. 170(Suppl. 2), S28–55.

Müller, K., 2005. SeqState: primer design and sequence statistics for phylogeneticDNA datasets. Appl. Bioinform. 4, 65–69.

Nagalingum, N.S., Marshall, C.R., Quental, T.B., Rai, H.S., Little, D.P., Mathews, S.,2011. Recent synchronous radiation of a living fossil. Science 334, 796–799.

Nystedt, B., Street, N.R., Wetterbom, A., Zuccolo, A., Lin, Y.-C., Scofield, D.G., Vezzi, F.,Delhomme, N., Giacomello, S., Alexeyenko, A., Vicedomini, R., Sahlin, K.,Sherwood, E., Elfstrand, M., Gramzow, L., Holmberg, K., Hällman, J., Keech, O.,Klasson, L., Koriabine, M., Kucukoglu, M., Käller, M., Luthman, J., Lysholm, F.,Niittylä, T., Olson, A., Rilakovic, N., Ritland, C., Rossello, J.A., Sena, J., Svensson, T.,Talavera-López, C., Theißen, G., Tuominen, H., Vanneste, K., Z-Q, Wu., Zhang, B.,Zerbe, P., Arvestad, L., Bhalerao, R., Bohlmann, J., Bousquet, J., Garcia Gil, R.,Hvidsten, T.R., de Jong, P., MacKay, J., Morgante, M., Ritland, K., Sundberg, B.,Thompson, S.L., Van de Peer, Y., Andersson, B., Nilsson, O., Ingvarsson, P.K.,Lundeberg, J., Jansson, S., 2013. The Norway spruce genome sequence andconifer genome evolution. Nature. http://dx.doi.org/10.1038/nature12211.

Parducci, L., Jørgensen, T., Tollefsrud, M.M., Elverland, E., Alm, T., Fontana, S.L.,Bennett, K.D., Haile, J., Matetovici, I., Suyama, Y., Edwards, M.E., Andersen, K.,Rasmussen, M., Boessenkool, S., Coissac, E., Brochmann, C., Taberlet, P.,Houmark-Nielsen, M., Larsen, N.K., Orlando, L., Gilbert, M.T.P., Kjær, K.H.,

Alsos, Willerslev, E., . Glacial survival of boreal trees in northern Scandinavia.Science 335, 1083–1086.

Perron, M., Bousquet, J., 1997. Natural hybridization between black spruce and redspruce. Mol. Ecol. 6, 725–734.

Perron, M., Perry, D.J., Andalo, C., Bousquet, J., 2000. Evidence from sequence-tagged-site markers of a recent progenitor-derivative species pair in conifers.Proc. Nat. Acad. Sci. USA 97, 11331–11336.

Posada, D., 2008. JModelTest: phylogenetic model averaging. Mol. Biol. Evol. 25,1253–1256.

Qiao, C.-Y., Ran, J.-H., Li, Y., Wang, X.-Q., 2007. Phylogeny and biogeography ofCedrus (Pinaceae) inferred from sequences of seven paternal chloroplast andmaternal mitochondrial DNA regions. Ann. Bot. 100, 573–580.

Rajora, O.P., Dancik, B.P., 2000. Population genetic variation, structure, andevolution in Engelmann spruce, white spruce, and their natural hybridcomplex in Alberta. Can. J. Bot. 78, 768–780.

Rambaut, A., Drummond, A.J. 2007. Tracer v1.4. <http://beast.bio.ed.ac.uk/Tracer>.Ran, J.H., Wei, X.X., Wang, X.Q., 2006. Molecular phylogeny and biogeography of

Picea (Pinaceae): implications for phylogeographical studies using cytoplasmichaplotypes. Mol. Phylogenet. Evol. 41, 405–419.

Ree, R.H., Smith, S.A., 2008. Maximum likelihood inference of geographic rangeevolution by dispersal, local extinction, and cladogenesis. Syst. Biol. 57, 4–14.

Scherson, R.A., Vidal, R., Sanderson, M.J., 2008. Phylogeny, biogeography, and ratesof diversification of New World Astragalus (Leguminosae) with an emphasis onSouth American radiations. Am. J. Bot. 95, 1030–1039.

Schmidt, P.A., 1988. Taxonomisch-nomenklatorische Notiz zur Gattung Picea A.Dietr.. Haussknechtia 4, 37–38.

Schmidt-Vogt, H., 1974. Das natürliche Verbreitungsgebiet der Fichte (Picea abies[L.] Karst.) in Eurasien. Allg. Forst Jagdztg 145, 185–197.

Schmidt-Vogt, H., 1977. Die Fichte, Band I. Verlag Paul Parey, Hamburg and Berlin.Sigurgeirsson, A., Szmidt, A., 1993. Phylogenetic and biogeographic implications of

chloroplast DNA variation in Picea. Nord. J. Bot. 13, 233–246.Silvestro, D., Michalak, I., 2011. raxmlGUI: a graphical front-end for RAxML. Org.

Divers. Evol..Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence based

phylogenetic analyses. Syst. Biol. 49, 369–381.Sperisen, C., Büchler, U., Gugerli, F., Mátyás, G., Geburek, T., Vendramin, G.G., 2001.

Tandem repeats in plant mitochondrial genomes: application to the analysis ofpopulation differentiation in the conifer Norway spruce. Mol. Ecol. 10, 257–263.

Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogeneticanalyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–2690.

Sutton, B.C.S., Flanagan, D.J., Gawley, J.R., Newton, C.H., Lester, D.T., El- Kassaby, Y.A.,1991. Inheritance of chloroplast and mitochondrial DNA in Picea andcomposition of hybrids from introgression zones. Theor. Appl. Genet. 82, 242–248.

Tollefsrud, M.M., Kissling, R., Gugerli, F., Johnsen, Ø., Skrøppa, T., Cheddadi, R., vander Knaap, W.O., Latałowa, M., Terhürne-Berson, R., Litt, T., Geburek, T.,Brochmann, C., Sperisen, C., 2008. Genetic consequences of glacial survivaland postglacial colonization in Norway spruce: combined analysis ofmitochondrial DNA and fossil pollen. Mol. Ecol. 17, 4134–4150.

Treutlein, J., Wink, M., 2002. Molecular phylogeny of cycads inferred from rbcLsequences. Naturwissenschaften 89, 221–225.

Wei, X.-X., Wang, X.-Q., 2003. Phylogenetic split of Larix: evidence from paternallyinherited cpDNA trnT-trnF region. Plant Syst. Evol. 239, 67–77.

Wei, X.-X., Yang, Z.-Y., Li, Y., Wang, X.-Q., 2010. Molecular phylogeny andbiogeography of Pseudotsuga (Pinaceae): Insights into the floristicrelationship between Taiwan and its adjacent areas. Mol. Phylogenet. Evol.55, 776–785.

Weng, C., Jackson, S.T., 2000. Species differentiation of North American spruce(Picea) based on morphological and anatomical characteristics of needles. Can. J.Bot. 78, 1367–1383.

Willyard, A., Syring, J., Gernandt, D.S., Liston, A., Cronn, R., 2007. Fossil calibration ofmolecular divergence infers a moderate mutation rate and recent radiations forPinus. Mol. Biol. Evol. 24, 90–101.

Wojciechowski, M.F., 2005. Astragalus (Fabaceae): a molecular phylogeneticperspective. Brittonia 57, 382–396.

Won, H., Renner, S.S., 2006. Dating dispersal and radiation in the gymnospermGnetum (Gnetales) – clock calibration when outgroup relationships areuncertain. Syst. Biol. 55, 610–622.

Wright, J.W., 1955. Species crossability in spruce in relation to distribution andtaxonomy. Forest Sci. 1, 319–349.

Xiang, Q.-P., Xiang, Q.-Y., Guo, Y.-Y., Zhang, X.-C., 2009. Phylogeny of Abies(Pinaceae) inferred from nrITS sequence data. Taxon 58, 141–152.

Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, andaberrations in global climate 65 Ma to present. Science 292, 686–693.

Zou, J.-B., Peng, X.-L., Li, L., Liu, J.-Q., Miehe, G., Opgenoorth, L., 2012. Molecularphylogeography and evolutionary history of Picea likiangensis in the Qinghai–Tibetan Plateau inferred from mitochondrial and chloroplast DNA sequencevariation. J. Syst. Evol. 50, 341–350.