This Article Appeared in a Journal Published by Elsevier. the Attached

Home , Abies durangensis, Abies fabri, Abies guatemalensis

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy

Molecular Phylogenetics and Evolution 62 (2012) 263–274

Contents lists available at SciVerse ScienceDirect

Molecular Phylogenetics and Evolution

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

Phylogeny, diversiﬁcation rates and species boundaries of Mesoamerican ﬁrs (Abies, Pinaceae) in a genus-wide context

Érika Aguirre-Planter a,1, Juan P. Jaramillo-Correa a,b,1, Sandra Gómez-Acevedo a, Damase P. Khasa b, ⇑ Jean Bousquet b, Luis E. Eguiarte a, a Departamento de Ecología Evolutiva, Instituto de Ecología, Universidad Nacional Autónoma de México, Apartado Postal 70-275, México, D.F., Mexico b Canada Research Chair in Forest and Enrionmental Genomics, Centre for Forest Research and Institute for Systems and Integrative Biology, Université Laval, Québec, Canada G1V 0A6 article info abstract

Article history: The genus Abies is distributed discontinuously in the temperate and subtropical montane forests of the Received 18 May 2011 northern hemisphere. In Mesoamerica (Mexico and northern Central America), modern firs originated Revised 26 September 2011 from the divergence of isolated mountain populations of migrating North American taxa. However, the Accepted 28 September 2011 number of ancestral species, migratory waves and diversification speed of these taxa is unknown. Here, Available online 10 October 2011 variation in repetitive (Pt30204, Pt63718, and Pt71936) and non-repetitive (rbcL, rps18-rpl20 and trnL- trnF) regions of the chloroplast genome was used to reconstruct the phylogenetic relationships of the Keywords: Mesoamerican Abies in a genus-wide context. These phylogenies and two fossil-calibrated scenarios were Biogeography further employed to estimate divergence dates and diversification rates within the genus, and to test the Chloroplast DNA Conifer hypothesis that, as in many angiosperms, conifers may exhibit accelerated speciation rates in the sub- Mexico tropics. All phylogenies showed five main clusters that mostly agreed with the currently recognized sec- Transverse Volcanic Belt tions of Abies and with the geographic distribution of species. The Mesoamerican taxa formed a single Molecular phylogeny group with species from southwestern North America of sections Oiamel and Grandis. However, popula- Diversification rates tions of the same species were not monophyletic within this group. Divergence of this whole group dated back to the late Paleocene and the early Miocene depending on the calibration used, which translated in

very low diversification rates (r0.0 = 0.026–0.054, r0.9 = 0.009–0.019 sp/Ma). Such low rates were a constant along the entire genus, including both the subtropical and temperate taxa. An extended phylogeographic analysis on the Mesoamerican clade indicated that Abies flinckii and A. concolor were the most divergent taxa, while the remaining species (A. durangensis, A. guatemalensis, A. hickelii, A. religiosa and A. vejari) formed a single group. Altogether, these results show that divergence of Mesoamerican firs coincides with a model of environmental stasis and decreased extinction rate, being probably prompted by a series of range expansions and isolation-by-distance. Ó 2011 Elsevier Inc. All rights reserved.

1. Introduction the fastest speciation rates are usually attained in taxa with high background substitution rates and that have recently expanded Diversification and speciation rates seem to be higher in taxa into new habitats, while the lowest are most often observed in spe- from tropical or subtropical ecosystems than from temperate or cies with both low substitution and extinction rates, and which in- boreal zones (e.g. Wright et al., 2006; Gillman et al., 2010). Such habit stable environments (Lancaster, 2010). Conifers are slowly differences can be prompted by contrasting patterns of population evolving taxa that should conform to this last low speciation mod- isolation, adaptation speed, extinction rates and/or stochastic el. However, based on morphology, conifer taxa appear to be more events that cause reproductive isolation (e.g. Wright et al., 2006; diversified in the subtropical than in the temperate environments Glor, 2010). For instance, in angiosperms, it has been shown that (Farjon, 1990), which might suggest some kind of divergence accel- eration towards the equator. In subtropical conifers, speciation related to geographical isolation is expected to be a predominant ⇑ Corresponding author. Fax: +52 55 5616 1976. driver for diversification, while slow reproductive isolation should E-mail addresses: [email protected] (É. Aguirre-Planter), jaramillo@ be a delaying factor (e.g. Bouille´ et al., 2011). Thus, it could be ecologia.unam.mx (J.P. Jaramillo-Correa), [email protected] (S. Gómez-Acevedo), hypothesized that the higher number of conifer species observed [email protected] (D.P. Khasa), [email protected] (J. in the subtropics is either the result of accelerated speciation due Bousquet), [email protected] (L.E. Eguiarte). 1 These authors contributed equally to this work. to increased historical isolation, or simply an artifact produced

264 É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 by an elevated phenotypic plasticity that has misled experts when 2. Materials and methods describing species based solely on morphological variation (e.g. De Kroon et al., 2005; Prada et al., 2008). 2.1. Sampling and Mesoamerican species The knowledge of diversification rates and speciation processes are of particular importance in megadiverse countries like Between one and five individuals were collected for 31 Abies Mexico, where the lack of such information collides with the species in natural populations, botanical gardens and arboretums large conservation needs (Callmander et al., 2005). Mexico has (see Table A1). These taxa are naturally distributed in North Amer- approximately 30,000 plant taxa (Rzedowski, 1993), from which ica, Europe and Asia, and represent most of the sections described about 10% are of conservation concern (SEMARNAT, 2010). This by Liu (1971), and Farjon and Rushforth (1989) (Table 1). Sampled species richness has been explained by the habitat variation pro- outgroups included Keteleeria davidiana (Bertrand) Beissner (the duced by a complex geological history (Espinosa-Organista et al., sister genus of Abies) and Larix kaempferi (Lambert) Carrière. 2008; Jaramillo-Correa et al., 2009), which provides ideal condi- For the Mesoamerican firs, sampling was far more extensive in tions for the diversification of the many taxa that expanded from an effort to disentangle their evolutionary and phylogeographic both North and South America during the last 30 million years relationships. Needles were collected from between 10 and 30 (Rzedowski, 1978; Graham, 1999). Conifers were among the first adult cone-bearer trees in 36 populations in both Mexico and Gua- of such temperate elements that arrived into Mexico (i.e. some temala, which covered as much as possible the ranges of the eight 23 Ma; Graham, 1999), and since then, they have become one taxa currently described for this region (Fig. 1, Table A1). These the most diverse and important components of the montane for- taxa include: ests of this country. However, in spite of this, the taxonomic status of many Mexican conifers remains dubious and has been 1. Abies guatemalensis Rehder. It is the most southerly dis- questioned in different occasions (e.g. Farjon and Rushforth, tributed species of the genus. It forms small-scattered 1989; Farjon, 1990; Strandby et al., 2009). mountain populations between central Mexico and Hon- Firs (Abies Miller; a predominantly temperate northern genus) duras and El Salvador, at altitudes between 2000 and are among the most abundant and less studied taxa of the mon- 4000 m above the sea level (a.s.l.) (Martínez, 1948; Dona- tane forests of Mesoamerica (i.e. Mexico and northern Central hue et al., 1985; Andersen et al., 2006). America). They are represented by eight threatened species, six 2. Abies religiosa (Humboldt, Bonpland et Kunth) Schlechten- of which are endemic to Mexico (Liu, 1971; Farjon and Rush- dal et Chamisso. This taxon is relatively common along the forth, 1989; SEMARNAT, 2010). Their first appearance in the Transverse Volcanic Belt (TVB) in central Mexico, and is Mesoamerican fossil record points to a late arrival in the Plio- distributed in mostly continuous stands between 2000 cene (i.e. about 5 Ma; Graham, 1976, 1999), which implies a and 3500 m.a.s.l. (Martínez, 1948). Populations of this spe- rather rapid diversification after their establishment. However, cies are the preferred overwintering habitat of the mon- it is unclear whether this date correctly reflects their first com- arch butterfly (Danaus plexippus; Anderson and Browler, ing into the region, or if there were one or more migratory 1996). waves from North America. Morphologically, Mesoamerican firs 3. Abies flinckii Rushforth. Given its discontinuous distribu- have been grouped into three or two different sections according tion in two separate clusters, this species was initially to the most widely recognized classifications (i.e. Liu, 1971; Far- described as two different varieties of A. religiosa and A. jon and Rushforth, 1989. See Table 1), which suggests at least guatemalensis (i.e., A. religiosa var. emarginata and A. guate- two independent expansions from as many ancestors. Unfortu- malensis var. jaliscana, respectively). Nevertheless, mor- nately, recent molecular phylogenetic analyses of this genus phometric (Rushforth, 1989; Strandby et al., 2009) and (i.e. Suyama et al., 2000; Xiang et al., 2004, 2009) had limited genetic surveys (Aguirre-Planter et al., 2000; Jaramillo- sampling at the local scale, and thus low resolution to address Correa et al., 2008) have provided enough evidence to this issue, while previous population studies (Aguirre-Planter grant it the species status. et al., 2000; Jaramillo-Correa et al., 2008) lacked the phyloge- 4. Abies hickelii Flous et Gaussen. This species is limited to a netic context necessary to test this hypothesis. Indeed, when few stands growing around 2500–3000 m.a.s.l. in the east- inferring diversification processes in closely related taxa such ern portion of the TVB and along the Sierra Madre del Sur, as the Mesoamerican firs, a combination of phylogenetic and in southern México (Martínez, 1948; Farjon, 1990). It has phylogeographic analyses on samples covering most of the genus one of the smallest ranges of the genus. range is necessary (e.g. Liston et al., 2007; Xu et al., 2010). 5&6. Abies durangensis Martínez. This taxon has two currently In the present study, we used the variation of repetitive and recognized varieties. Var. durangensis (5) has a scattered non-repetitive regions of the chloroplast genome (cpDNA) to: (i) distribution along the Sierra Madre Occidental in north- determine the putative number of ancestral lineages and western Mexico, at elevations between 2000 and migratory waves that originated the modern Mesoamerican 2900 m.a.s.l, (Martínez, 1948; Farjon, 1990). Var. coahuil- Abies, and (ii) test the hypothesis that the subtropical firs have ensis (6) occurs only in a few stands on damp canyons at experienced an accelerated diversification with respect to their altitudes close to 3000 m.a.s.l., in the northern part of temperate counterparts. To do so, we gathered a large sample the Sierra Madre Oriental (Rushforth, 1989). In this study, from all Mesoamerican species that covered most of their ranges, both varieties were considered as independent taxa (see which was included in a phylogenetic analysis to evaluate their below). relationships with other Abies of the world. Then, we estimated 7. Abies vejari Martínez. The range of this species extends the divergence times and diversification rates of each group of along the Sierra Madre Oriental, from southeastern Coahu- the phylogeny. An exhaustive phylogeographic survey, which ila to Nuevo Leon and in western Tamaulipas (Martínez, complements a previous study on southern Mesoamerican firs 1948; Farjon, 1990). It is a high mountain taxon found (Jaramillo-Correa et al., 2008), was used to improve the resolu- between 2800 and 3300 m.a.s.l. on steep mountain slopes tion of the phylogeny for making inferences on the putative evo- (Farjon, 1990). We are considering as populations of this lutionary history of the genus since its expansion into taxon stands previously recognized as Abies mexicana Mesoamerica. Martínez, then reduced to Abies vejari var. mexicana Author's personal copy

É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 265

Table 1 Comparison of the two most used Abies species classiﬁcations (Liu, 1971; Farjon and Rushforth, 1989). Mexican taxa included in the present study are underlined.

Liu Farjon and Rushforth Subgenus Pseudotorreya I. sect. Bracteata sect. Bracteata A. bracteata A. bracteata II. sect. Balsamea Subgenus Abies A. kawakamii, A. sibirica, A. balsamea, A. lasiocarpa, A. sachalinensis, A. koreana, A. fraseri, A. nephrolepis, A. sect. Momi veitchii A. firma III. sect. Amabilis sect. Homolepides A. amabilis, A. mariesii A. holophylla, A. homolepis, A. mariesii, A. kawakamii IV. sect. Grandis sect. Chensienses A. grandis, A. concolor, A. durangensis, A. guatemalensis, A. flinckii A. chensiensis, A. ernestii V. sect. Oiamel sect. Elateopsis A. religiosa, A. vejari, A. hickeli A. delavayi, A. fargesii, A. recurvata, A. squamata VI. sect. Nobilis sect. Elate A. procera, A. magnifica A. koreana, A. nephrolepis, A. sachalinensis, A. veitchii VII. sect. Momi sect. Pichta A. homolepis, A. recurvata, A. firma, A. beshanzuensis, A. holophylla, A. chensiensis, A. pindrow, A. ziyuanensis A. sibirica VIII. sect. Pseudopicea sect. Pindrau A. delavayi, A. fabri, A. forrestii, A. chengii, A. densa, A. spectabilis, A. fargesii, A. fanjingshanensis, A. A. spectabilis, A. pindrow yuanbaoshanensis, A. squamata sect. Abies IX. sect. Abies A. alba, A. nebrodensis, A. nordmanniana, A. cephalonica A. alba, A. cephalonica, A. nordmanniana, A. nebrodensis, A. ciclica sect. Piceaster X. sect. Piceaster A. pinsapo, A. cilicica, A. numidica A. pinsapo, A. numidica sect. Nobiles A. procera, A. magnifica sect. Oyamel A. religiosa, A. hickeli sect. Vejarianae A. vejari sect. Grandes

A. grandis, A. amabilis, A. durangensis, A. guatemalensis, A. concolor sect. Balsamea A. balsamea, A. lasiocarpa, A. fraseri

Gulf of California U.S.A.

Gulf of Mexico

Sierra Madre Oriental Mexico Sierra Madre Occidental Transverse Volcanic Belt

Pacific Ocean

Circle color Guatemala A. flincki A. durang v. coah A. hickelii A. durangensis Sierra Madre A. religiosa A. vejari del Sur A. guatemalensis A. concolor Honduras

Fig. 1. Geographic location of Mesoamerican Abies populations included in the phylogenetic and phylogeographic surveys. The dashed lines delimitate the main mountain ranges in Mexico that are mentioned in the text.

(Martínez) (Liu, 1971) and most recently assigned to a sub- 8. Abies concolor (Gordon) Lindley ex Hildebrand. This species is species by Farjon (1990). Our ﬁeld observations suggest distributed in SW and western USA (Oregon, California, and that these populations are indeed morphologically similar. Rocky Mountains of Utah, Colorado, Arizona, New Mexico, Author's personal copy

266 É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274

Idaho, and Nevada) and has a few scattered populations in gruent. In all cases, K. davidiana and L. kaempferi were used as northern Mexico (Sonora and North Baja California; Martínez, outgroups. 1948; Farjon, 1990). It occurs between 600 and 3350 m.a.s.l. 2.4. Divergence times and diversification rates Further descriptions of the populations and species included in the present study can be found in Aguirre-Planter et al. (2000, and Estimates of diversification ages were obtained with the references therein) and with the samples stored in the Herbario penalized likelihood method (Sanderson, 2002) implemented in Nacional (MEXU). r8s version 1.71 (Sanderson, 2003). We used the consensus tree obtained with the combined set of three cpDNA sequences (rbcL, rps18-rpl20 and trnL-trnF), and the point estimates of ages (min- 2.2. DNA extraction and manipulation imum, maximum and mean with standard deviation) were obtained using 272 trees topologically identical to this consensus Total DNA was extracted from frozen needles by using a cetyl- tree. The calibration points used were (1) a minimal age of trimethyl ammonium bromide (CTAB) mini-prep protocol (Váz- 44.4 Ma for K. davidiana (LePage, 2003) and (2) two putative quez-Lobo, 1996), or a DNeasy Plant Mini Kit (Qiagen), and its dates for the split between the genera Keteleeria and Abies (node concentration was measured with a GeneSpec spectrophotometer AinFigs 2–4), which were used in two independent analyses. (MiraiBio). Three cpDNA regions, rbcL, rps18-rpl20 and trnL-trnF, The first one (option 1) constrained this node to a minimum were amplified via the polymerase chain reaction (PCR) by using age of 100.4 Ma and a maximum age of 113.8 Ma (Kremp, primer pairs reported elsewhere (Suyama et al., 2000). These three 1967; Xiang et al., 2007), and the second one (option 2) consid- regions have been shown to be the most useful for making phylo- ered minimum and maximum ages of 45.5 Ma and 55.0 Ma, genetic inferences in the genus Abies (Suyama et al., 2000). The PCR respectively (Axelrod, 1976; Erwin and Schorn, 2005). Diversifi- reaction mixture consisted of 3 lLof10Â reaction buffer, 1.5 lLof cation rates were determined following Magallón and Sanderson 30 mM/L magnesium chloride solution, 1.2 lL of a 1.25 mM/L dNTP (2001) by using a formula derived from the method-of-moments solution in equimolar ratio, 0.25 mM for each primer, 25–50 ng of estimator of Rohatgi (1976) on crown groups of the Abies phy- template DNA and 1 U of polymerase in a total volume of 30 lL. logeny (i.e., formula 7 in Magallón and Sanderson, 2001). This Amplification was carried out for an initial denaturation of 95 °C analysis allowed us to obtain two estimates of diversification for 10 min, followed by 30 PCR cycles consisting of 94 °C for 30 s, rates, corresponding to zero extinction (r ) and 90% relative 55 °C for 1 min and 72 °C for 1 min, with a final extension period 0.0 extinction (r ) rates, respectively. of 72 °C for 10 min. After verifying that a single band was amplified 0.9 by examining the PCR products on 2% agarose gels (in TAE), both 2.5. Phylogeographic analysis of Mexican firs DNA strands were sequenced directly in an Applied Biosystems 3130xl DNA Genetic Analyser by using the appropriate primers, a A previous study on southern Mexican species (Jaramillo-Correa Sequenase GC-rich kit (Applied Biosystems) and a dideoxynucleo- et al., 2008) showed that chloroplast microsatellite markers tide chain termination procedure. (cpSSR) were more informative than their mitochondrial counterparts to depict phylogeographical structure across taxa. Thus, in or- 2.3. Phylogenetic analyses der to extend such a survey to the northern Mexican species (i.e. A. durangensis var. durangensis, A. durangensis var. coahuilensis, A. vej- Sequences alignments were made with CLUSTALW (Thompson ari, and A. concolor), and to gain a better resolution of interspecific et al., 1994), as implemented in BIOEDIT 7.0.0 (Hall, 1999), for each relationships than the one obtained with the phylogeny (see Sec- independent data set and manually refined after checking individ- tion 3), the three most informative cpSSRs (Pt30204, Pt63718, ual chromatograms for putative base-calling errors, and to ensure and Pt71936; Vendramin et al., 1996) were amplified, electropho- the correct alignment of conserved bases and avoid redundant in- resed, genotyped, and sequenced as previously described (Jara- dels. Initially, all samples (i.e. between one and five individuals per millo-Correa et al., 2008) for 9–18 individuals from 3 to 5 species, see Table A1) were included, then identical sequences in populations per taxon (Table A1). Genotypes for the southernmost nearby populations of the same species were omitted from the fi- species (i.e., A. guatemalensis, A. hickelii, A. flincki, and A. religiosa) nal alignment and only one of them was kept. Final alignments were gathered from the above mentioned study. The remaining were submitted to the program MODELTEST (Posada and Crandall, non-Mexican taxa used in the phylogenetic analyses were ex- 1998) to find the best-fitting model for each independent DNA re- cluded given the scope of the hypotheses to be tested at the phylo- gion with the Akaike Information Criterion (AIC). Phylogenetic geographical level. analyses were conducted using a Bayesian approach (MRBAYES ver- Multi-locus cpDNA haplotypes (chlorotypes) were defined by sion 3.1.2; Huelsenbeck and Ronquist, 2001) by implementing assembling all observed polymorphisms. The evolutionary rela- gene partitioned data, applying the best fit substitution model, tionships among chlorotypes were depicted with a minimum and allowing unlinked parameter estimation among partitions. spanning tree estimated from DNA sequences with the software Two independent Markov Chain Monte Carlo (MCMC) runs were TCS (Clement et al., 2000), and by using a fix connection limit of performed by using random starting trees and a heating chain 5 steps (Templeton et al., 1992). Ambiguities induced by reticu- scheme (temp = 0.2). Each run included five million generations, lations were corrected based on the methodology proposed by and samples were taken every 200 generations. Then, a 50% Bayes- Crandall and Templeton (1993). The main chlorotype groups ob- ian consensus was created in PAUP⁄ v. 4 (Swofford, 2000) from all served within each population were then plotted onto a map trees sampled after burn-in (i.e. 2500 iterations). A maximum like- and compared with the results of a Bayesian analysis of popula- lihood phylogeny was also constructed with RAxML (Randomized tion structure and the genetic discontinuities estimated with the Axelerated Maximum Likelihood), a program for sequential and Montmonier’s maximum difference algorithm. The Bayesian parallel Maximum Likelihood-based inference of large phyloge- analysis was performed with the software BAPS 4.13 (Corander netic trees that has the advantage of doing rapid bootstrap heuris- et al., 2003; Corander and Marttinen, 2006) by setting the num- tic searches (Stamatakis et al., 2008). These analyses were initially ber of clusters (k) to range between 2 and 40. The optimal par- conducted for each independent dataset and then on the combined tition of populations was then determined by repeating the set of sequences, after verifying that all three partitions were con- analysis 100 times with the k-value that exhibited the minimum Author's personal copy

É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 267

Abies hickeli 7 Abies guatemalensis 11 Abies hickeli 1 Abies durangensis 35 Abies guatemalensis LX Abies guatemalensis PBL Abies guatemalensis SMI Abies guatemalensis 2 Abies hickeli 46 Abies durangensis 24 Abies religiosa 13 Abies flinckii 19 Abies flinckii 14 Abies religiosa 12 Abies religiosa 15 Abies guatemalensis TAJ Abies guatemalensis 42 Abies grandis Abies concolor 1 Group I Abies guatemalensis 41 Abies religiosa 22 (Mesoamerica + Abies guatemalensis LC Western North America) Abies flinckii 16 Abies guatemalensis 10 Abies hickeli 4 Abies religiosa 48 Abies concolor 33 D Abies vejari 28 Abies concolor 37 Abies durangensis 31 Abies durangensis var. coahuilensis 25 Abies durangensis var. coahuilensis 26 C Abies vejari 27 Abies concolor 2 Abies concolor 36 Abies vejari 29 E Abies magnifica Abies amabilis Group II (Western North America + Japan) Abies mariesii Abies balsamea Abies sibirica Abies lasiocarpa Abies fraseri B Abies holophylla Abies koreana Abies nephrolepis Group III Abies sachalinensis (Eurasia + Northern Abies veitchii Abies firma North America) Abies homolepis H Abies fabri A Abies fargesii Abies recurvata G Abies kawakami Abies alba Abies pinsapo I Abies nordmanniana Group IV (Europe + Mediterranean) F Abies numidica Abies bracteata Group V (California) Keteleeria davidiana Larix kaempferi

0.0050

Fig. 2. Maximum-likelihood phylogenetic tree (50% bootstrap consensus) of 33 Abies taxa inferred from three chloroplast markers (rbcL, rps18-rpl20, trnL-trnF). Thicker horizontal lines represent branches supported by bootstrap values over 95%. Dashed lines represent branches supported by bootstrap values between 50 and 95%. Roman numbers correspond to meaningful groups from an evolutionary or systematic point of view (see Results and Discussion).

log-likelihood. The genetic discontinuities were determined with probabilities, especially when the concatenated matrix and the the software BARRIER ver. 2.2 (Manni et al., 2004) by using a pair- partitioned models were used. Nevertheless, the Bayesian tree wise Slatkin’s linearized FST distance matrix between populations had higher overall support than the ML phylogeny. estimated with ARLEQUIN 3.5 (Excoffier and Lischer, 2010). The The first group, Group I (node D), was composed by all Meso- number of barriers ranged from 1 to 10 and statistical support american species and A. grandis. All these taxa correspond to sec- for these genetic discontinuities was determined from 100 boot- tions Oiamel and Grandis from Farjon and Rushforth (1989). Some strap replicates of the genetic distance matrix, which were ob- smaller assemblages could be observed within this large group, tained by resampling individuals within populations. Only which were mainly composed by populations from the northern those barriers with high statistical robustness (i.e. above 80%) and southern parts of Mesoamerica, but without showing any clear were kept. or well supported assortment of species (Figs. 2–4). The second group (II, node E) was related to the first group, and included 3. Results two other species from western North America (A. amabilis and A. magnifica) and one from Japan (A. mariesii). These three taxa 3.1. Phylogenetic analyses have been assigned to sections Amabilis and Nobilis by Farjon and Rushforth (1989). All the remaining Asian species were found in A total of 1284, 541, and 553 nucleotides were sequenced for the third group (III, node H) together with the taxa from northern rbcL, rps18-rpl20 and trnL-trnF, respectively. The final alignment and eastern North America (i.e., A. balsamea, A. lasiocarpa and A. for the Mexican Abies taxa included 35 different sequences, fraseri). The species from this group belong to sections Momi and which are all available in Genbank (accession nos. JN935603– Balsamea, respectively (Farjon and Rushforth, 1989). The European JN935765). The best fitting substitution models found with the and Mediterranean taxa (A. alba, A. pinsapo, A. nordmanniana and A. AIC were TIM + I for rbcL and K81uf + G for spacers rps18-rpl20 numidica), namely those classified in sections A. and Piceaster by and trnL-trnF, respectively. The monophyly of the genus Abies Farjon and Rushforth (1989), were found in the fourth group (IV, was confirmed in both the ML and Bayesian phylogenies, which node I), and the rare and California-endemic A. bracteata formed exhibited a same main basic topology composed by five groups a distinct basal group (V, node F) that diverged at an early stage (Figs. 2–4). These groups were all supported by high posterior from all the others (Figs. 2–4). Author's personal copy

268 É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274

Abies vejari 27 Abies durangensis var. coahuilensis 26 Abies durangensis 31 Abies durangensis var. coahuilensis 25 Abies vejari 29 Abies concolor 36 Abies hickeli 7 Abies flinckii 19 Abies guatemalensis 11 Abies hickeli 1 Abies guatemalensis SMI Abies religiosa 12 Abies religiosa 15 Abies religiosa 13 Abies guatemalensis 2 Abies guatemalensis LX Abies hickeli 46 Abies flinckii 14 Abies guatemalensis PBL Group I Abies duangensis 35 Abies durangensis 24 (Mesoamerica + Abies grandis Abies guatemalensis 41 Western North America) Abies concolor 1 Abies guatemalensis 42 Abies guatemalensis TAJ Abies religiosa 22 Abies flinckii 16 Abies guatemalensis LC Abies vejari 28 Abies concolor 2 Abies concolor 37 Abies guatemalensis 10 D Abies concolor 33 Abies religiosa 48 C Abies hickeli 4 Abies amabilis E Abies magnifica Abies mariesii Group II (Western North America + Japan) Abies nephrolepis Abies sachalinensis Abies koreana Abies firma Abies veitchii Abies homolepis B Abies holophylla Group III Abies balsamea Abies sibirica (Eurasia + Northern Abies lasiocarpa Abies fraseri North America) Abies fabri H Abies fargesii A Abies kawakami Abies recurvata G Abies alba F Abies pinsapo I Abies nordmanniana Group IV (Europe + Mediterranean) Abies numidica Abies bracteata Group V (California) Larix kaempferi Keteleeria davidiana

Fig. 3. 50% Bayesian consensus tree of 33 Abies taxa inferred from three chloroplast markers (rbcL, rps18-rpl20, trnL-trnF). The letters and roman numerals and roman numerals indicate the groups discussed in the text. Thicker horizontal lines represent branches supported by posterior probabilities over 0.95. Dashed lines represent branches supported by posterior probabilities between 0.5 and 0.95. Roman numbers correspond to meaningful groups from an evolutionary or systematic point of view (see Results and Discussion).

3.2. Divergence times and diversification rates variants per species. Sequencing of the amplified fragments confirmed that most of the polymorphisms were due to variation in When assuming that the split between Abies and Keteleeria the mononucleotide stretches. However, as in a previous survey occurred between 113.8 and 100.4 Ma (Kremp, 1967; Xiang in southern Mexican firs (Jaramillo-Correa et al., 2008), some et al., 2007; option 1), the crown age of origin for the genus duplications and SNPs were observed in the regions flanking Abies (node B in Fig. 3) was estimated in 82 ± 13 Ma. When con- the SSRs. Only one of these SNPs was new with respect to this sidering that the genus is composed by 48 species, this estimate previous study, and helped define 31 new haplotypes that were resulted in diversification rates of 0.0386 and 0.0205, for the exclusive to A. concolor and A. durangensis var. durangensis zero (r0.0) and 90% extinction (r0.9) rates, respectively. The esti- (Fig. 5). The total number of chlorotypes, including those from mated age for Group I, which included the eight Mesoamerican the previous study, was 210, and 91.4% of them had frequencies species and A. grandis (node D in Fig. 3), was 58 ± 12 Ma (Ta- below 0.01 when averaged over the total sample size for Mexi- ble 2). This group had a diversification rate twice as high can firs (data available from the authors upon request). Forty-

(r0.0 = 0.0256, r0.9 = 0.0091) as its closely related Group II nine of these cpDNA types were common and shared across spe- (r0.0 = 0.0142, r0.9 = 0.0041; node E in Fig. 3), which apparently cies, thus suggesting that they represent putative ancestral poly- originated later, at 48 ± 15 Ma. The Asian–northern North Amer- morphisms or that they have been more recently spread through ican Group III (node H) would have originated 48 ± 13 Ma ago, interspecific gene flow (Fig. 5). and exhibited the highest diversification rate of the genus The chlorotype network (Fig. 5) revealed three significantly dif-

(r0.0 = 0.0542, r0.9 = 0.0256), while the European–Mediterranean ferent groups of haplotypes, exclusively defined by the indels and Group IV (node I) would be slightly younger (43 ± 13 Ma) and SNPs flanking the SSRs, which should represent unique and non- would have a lower diversification rate than Group III homoplastic mutational events (Vachon and Freeland, 2010). As

(r0.0 = 0.0290, r0.9 = 0.0096). The divergence time between Groups mentioned above, the first group was entirely formed by individu- III, IV and V (node F, Fig. 3) was estimated in 73 ± 13 Ma, while als of A. concolor and A. durangensis var. durangensis and distributed the separation between Groups I and II (node C) should have oc- in north-western Mexico. The second group was mostly composed curred about 72 ± 12 Ma ago. On the other hand, when assuming by individuals of A. flincki, and was restricted to the westernmost the less conservative split dates between Abies and Keteleeria (i.e. part of the Transmexican Volcanic Belt in central Mexico. This between 55 and 45.5 Ma; Axelrod, 1976; Erwin and Schorn, group was already observed in the previous survey on southern 2005; option 2), the age estimates obtained were roughly half Mexican firs (Jaramillo-Correa et al., 2008), and in the present of those produced when assuming option 1 (Table 2). study, only a few additional trees of the northern A. vejari were assigned to it. The third group enclosed individuals from all species, and was widely distributed across Mesoamerica. Interestingly, all 3.3. Phylogeographic structure in Mesoamerican Abies individuals of A. durangensis var. coahuilensis were included in this group, while half of the trees of A. durangensis var. durangensis All three cpSSRs were polymorphic in the northern Mexican were assigned to the first group. Both, the partition obtained with taxa (A. durangensis var. durangensis, A. durangensis var. coahuil- BAPS and the genetic breaks disclosed with barrier coincided with ensis, A. vejari and A. concolor), exhibiting between 2 and 9 size Author's personal copy

É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 269

Abies vejari 27 Abies durangensis var. coahuilensis 26 Abies durangensis 31 Abies durangensis var. coahuilensis 25 Abies vejari 29 Abies concolor 36 Abies hickeli 7 Abies flinckii 19 Abies guatemalensis 11 Abies hickeli 1 Abies guatemalensis SMI Abies religiosa 12 Abies religiosa 15 Abies religiosa 13 Abies guatemalensis 2 Abies guatemalensis LX Abies hickeli 46 Abies flinckii 14 Abies guatemalensis PBL Abies duangensis 35 Abies durangensis 24 Abies grandis Abies guatemalensis 41 Abies concolor 1 Abies guatemalensis 42 Abies guatemalensis TA J Abies religiosa 22 Abies flinckii 16 Abies guatemalensis LC Abies vejari 28 Abies concolor 2 Abies concolor 37 Abies guatemalensis 10 D Abies concolor 33 Abies religiosa 48 C Abies hickeli 4 Abies amabilis E Abies magnifica Abies mariesii Abies nephrolepis Abies sachalinensis Abies koreana Abies firma Abies veitchii Abies homolepis B Abies holophylla Abies balsamea Abies sibirica Abies lasiocarpa Abies fraseri Abies fabri H Abies fargesii Abies kawakami G Abies recurvata Abies alba F Abies pinsapo I Abies nordmanniana Abies numidica Abies bracteata Keteleeria davidiana

65.0 54.8 33.7 23.8 5.3 1.8 Ma Option 1 Cretaceous Paleocene Eocene Oligocene Miocene Pliocene

33.7 23.8 5.3 1.8 Ma Option 2 Eocene Oligocene Miocene Pliocene

Fig. 4. Chronogram derived from the consensus of the 272 compatible Bayesian trees. The letters indicate the groups discussed in the text (see Results and Discussion). Thicker horizontal lines represent branches supported by posterior probabilities between 0.95–1.00. Dashed lines represent branches supported by posterior probabilities between 0.5 and 0.95.

Table 2 Mean, minimum and maximum age estimates (standard deviations in parentheses) derived from 272 sampled Bayesian trees that were topologically identical to the consensus phylogenetic tree obtained for the genus Abies, and diversiﬁcation rates determined for each species group under zero (r0.0) and 90% relative extinction (r0.9) rates by using two different calibration points.

Nodec Option 1a Option 2b Mean aged Minimumd Maximumd Diversiﬁcation rate Mean aged Minimumd Maximumd Diversiﬁcation rate

r0.0 r0.9 r0.0 r0.9 A 111.99 (3.40) 100.86 113.80 – – 53.80 (2.28) 46.25 55.00 – – B 82.21 (12.85) 54.21 107.27 0.0386 0.0205 38.77 (5.59) 24.70 51.10 0.0819 0.0435 C 71.88 (12.12) 48.07 105.17 0.0260 0.0102 33.90 (5.72) 22.42 50.02 0.0552 0.0216 D 58.58 (12.32) 32.11 100.56 0.0256 0.0091 27.67 (5.84) 15.31 48.52 0.0543 0.0192 E 48.55 (14.24) 21.01 96.47 0.0142 0.0041 22.95 (6.68) 9.61 45.84 0.0302 0.0088 F 73.50 (12.95) 47.64 105.08 0.0385 0.0191 34.64 (6.08) 22.23 50.03 0.0817 0.0406 G 59.86 (13.67) 32.08 99.73 0.0473 0.0234 28.25 (6.39) 15.57 47.39 0.1002 0.0497 H 47.98 (12.85) 21.93 89.06 0.0542 0.0256 22.69 (5.97) 11.15 41.03 0.1147 0.0541 I 43.06 (13.13) 16.90 88.56 0.0290 0.0096 20.38 (6.12) 8.52 40.80 0.0614 0.0203

a Option 1 assumes a minimal age of 44.4 Ma for Keteleeria (LePage, 2003), and minimum and maximum ages of 100.4 Ma and 113.8 Ma, respectively (Kremp, 1967; Xiang et al., 2007), for the split between Keteleeria and Abies (node A). b Option 2 assumes the same minimal age for Keteleeria, and ages between 45.5 Ma and 55.0 Ma for the split of this genus from Abies (Axelrod, 1976; Erwin and Schorn, 2005). c Letters correspond to nodes in Figs. 2 and 3. d All ages are given in million years. Author's personal copy

270 É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274

A Clade II Clade I Circle filling A. flincki A. durang v. coah B A. hickelii A. durangensis A. religiosa A. vejari A. guatemalensis A. concolor

Clade III

Circle filling Clade I

Clade II

Clade III

Circle border A. flincki A. durang v. coah A. hickelii A. durangensis A. religiosa A. vejari A. guatemalensis A. concolor

Fig. 5. Minimum-spanning network (A) and geographic location of the main cpSSR groups and genetic barriers (dashed arrows) (B) in populations of eight Mesoamerican Abies taxa. The network represents the most parsimonious relationships among chlorotypes with a ﬁx connection limit of ﬁve steps, where each cpDNA-type is represented by a circle whose size is proportional to its abundance. Small white dotted circles represent missing chlorotypes. The map depicts the relative frequency of each of the three main groups delineated in the network for each population. The barriers were determined with Montmonier’s maximum-difference algorithm on FST-distances derived from chlorotype frequencies, and coincided with the groups detected with a Bayesian analysis of population structure. Their thickness is equivalent to their statistical robustness.

these three groups. However, the populations from Guatemala conservative calibration approach (option 1 in Table 2). However, were additionally separated from the Mexican stands in these caution has been advised when relying on pollen without associated two analyses (Fig. 5). macroremnants for performing calibrations, due to unknown homo- plasious characters with now-extinct genera (Erwin and Schorn, 4. Discussion 2005; Willyard et al., 2007). The two calibrations approaches used yielded a late Cretaceous 4.1. Monophyly and divergence of the genus Abies and an early Eocene crown-group diversification time, respectively, with the final divergence leading to the extant species occurring in In the present study, we confirmed the monophyly of the genus the Miocene (option 1) and the Pliocene (option 2). Such differ- Abies, and showed that it contains five strongly supported groups ences between scenarios are rather common in phylogenetic stud- of species. These groups mostly agree with the recognized sections ies in the Pinaceae (e.g. Willyard et al., 2007; Gernandt et al., 2008), from the most widely used classifications (Liu, 1971; Farjon and given that the choices made for calibrating the phylogenies repre- Rushforth, 1989), and with the continental location of species sent a significant source of uncertainty, especially for the times (Figs. 2–4). The majority of these concurrences were already ob- elapsed between the origin of taxa and both the earliest recogniz- served and discussed in previous works relying on chloroplast able fossil, and the divergence that originated the extant species and nuclear DNA data (Suyama et al., 2000; Kormut’ák et al., (Willyard et al., 2007). However, irrespective of which scenario is 2008; Xiang et al., 2009), which themselves backed earlier results the most appropriate for calibrating the Abies phylogeny, it seems of controlled crosses within and across sections (e.g. Critchfield, that two periods were key for the diversification of this genus: the 1988; Kormut’ák et al., 2008). Eocene and the Miocene. Although the topologies of the estimated phylogenies do not al- The Eocene was a period of falling global temperatures (e.g., Za- low inferring precisely where Abies first appeared, the basal position chos et al., 2001; Miller et al., 2008), favorable to range expansions of the morphologically divergent A. bracteata and its extant distribu- of temperate taxa such as most conifers (Farjon, 2005). Similar cal- tion in southwestern North America (Griffin and Critchfield, 1972) ibrated phylogenies of Juniperus and Cupressus revealed that their would support an origin in this region of the world. This observation first waves of diversification also occurred during this epoch would support to some degree the less conservative approach (op- (Mao et al., 2010), while studies in Pinus suggested that divergence tion 2 in Table 2) used as calibration point. Indeed, evidence from within the subgenus Strobus also began at the end of this period fragmentary organs indicates that Abies was distributed in the conif- (Gernandt et al., 2008). Thus, as proposed for many Tertiary taxa erous forests of western North America during the Eocene (Axelrod, (Milne and Abbot, 2002; Mao et al., 2010), the earlier Abies diver- 1976; Erwin and Schorn, 2005), relatively close to the region where sification (i.e. of either its basal lineages according to option 1; or the modern range of A. bracteata is located. Nevertheless, a much ear- the crown group according to option 2) might have been prompted lier presence of Abies in Eurasia (i.e. between the Cretaceous and the by isolation-by-distance and adaptation to marginal environments Paleocene) was inferred from pollen deposits in China (Xiang et al., after an expansion during the cooler periods of the Eocene, to then 2007) and Siberia (Kremp, 1967), which would support the most diminish during the more climatically stable Oligocene. Author's personal copy

É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 271

On the other hand, the Miocene divergence (i.e. of either mod- recent as the last glacial maximum. During this period, conifer for- ern populations: option 1; or major species groups: option 2) coin- ests reached their highest development in this country (e.g., Loz- cides with the uplift of major mountain ranges in Asia, in North ano-García et al., 1993; Vázquez-Selem and Heine, 2004), which America and in Mesoamerica (e.g., Spicer et al., 2003; Wilson and could have facilitated southward migrations and/or hybridization Pitts, 2010; see below). This period was accompanied by another with local related taxa, such as inferred from the chlorotype distri- progressive decrease of the global temperature (Zachos et al., bution in A. concolor and A. durangensis var. durangensis (Fig. 5). 2001). Altogether, these factors should have resulted in new habi- Nevertheless, these phylogeographical hypotheses are yet to be tats suitable for further expansions of temperate taxa (Farjon, tested through more detailed analyses combining coalescent simu- 2005) which, coupled with the isolating effects of the uprising lations on nuclear variation and ecological niche modeling (e.g. mountains, should have prompted diversification in Abies and Richards et al., 2007; Carstens et al., 2009; Martínez-Méndez other conifers (e.g. Bouillé and Bousquet, 2005; Willyard et al., et al., unpublished results). 2007; Mao et al., 2010; Xu et al., 2010). 4.3. Diversification rates in subtropical and temperate Abies 4.2. Monophyly and divergence of Mesoamerican firs The correlation between increased rates of diversification and The Mesoamerican species were all located into a single group past geological or climatic events and/or the occurrence of adap- (I) together with two taxa from western North America (A. concolor tive radiations have been usually invoked to explain the higher and A. grandis), but without forming monophyletic groups within species diversity in the tropics and subtropics than in the temper- this lineage (node D, Figs. 2–4). This large group conforms to a ate regions (e.g. Wright et al., 2006; Glor, 2010). However, in spite wide phylogeographic unit distributed from the southern parts of of exhibiting more species towards southern latitudes, the overall the Vancouver Island and the shores of British Columbia, along diversification rates of Abies were low (0.0386–0.0819 sp/Ma, this the coast and down into the northern parts of Baja California, study) and virtually identical in the subtropical (i.e. Mesoamerican and from the Pacific Northwest along the Rocky Mountains into Group I, node D) and temperate groups (II, node E to IV node I). the main Mesoamerican cordilleras. Its related Group II (node E) These rates are indeed comparable to those of other conifers (e.g. is also distributed in this region, although its southern limit is re- Juniperus 0.078 sp/Ma, Mao et al., 2010) and of slowly evolving stricted to central California and its north bounds spans further, angiosperms, such as the Nympheales (0.031 sp/Ma) and the Win- into southern Alaska. This group additionally appears to have expe- teraceae (0.0358 sp/Ma; Magallón and Sanderson, 2001). In angio- rienced an ancient long distance migration into eastern Asia, prob- sperms, low diversification rates are usually compatible with a ably via the Bering Land Bridge, where it diverged into modern A. model of environmental stasis and decreased extinction (Lancaster, mariesii (Suyama et al., 2000; Xiang et al., 2009). According to 2010), which could be further extended to conifers regardless of the estimated divergence times, Groups I and II separated between the environment they inhabit. the late Cretaceous and the early Eocene (node C, Fig. 4, Table 2), Low diversification rates are indeed expected in conifers. These and remained isolated for enough time to develop reproductive taxa, as most wind-pollinated trees, have extensive gene flow, out- barriers. Indeed, previous controlled crosses between A. concolor crossing mating systems, long generation times, and slow-develop- (Group I) and A. magnifica (Group II) only produced a few non-via- ing isolation mechanisms that allow hybridization in secondary ble offspring (Critchfield, 1988), while a survey on cpDNA variation contact zones, which decrease the likelihood of extinction and found no haplotype shared between these two widely sympatric diversification (e.g. Bousquet et al., 1992; Perron and Bousquet, species (Oline, 2008). 1997; Verdú, 2002; reviewed by Petit and Hampe (2006)). Conifers Both the current distribution of species from Group I (node D) further exhibit high amounts of repetitive elements (Grotkopp and their divergence dates suggest that they originated between et al., 2004; Morgante, 2006) and reduced rates of recombination the Paleocene and the Miocene, after a unique southward expan- at both the genome-wide and within-gene scale (Jaramillo-Correa sion that probably followed the mountain chains of western North et al., 2010), which also hamper rapid speciation processes. Thus, America. These ranges, collectively known as the American Cordil- as in slowly evolving angiosperms, diversification in conifers is ex- lera (Dickinson, 2004), formed in two phases that coincide with the pected to be driven by stochastic forces and/or variation at numer- key periods for the Abies diversification: first during the Laramide ous loci with small phenotypic effects, rather than on rapid Orogeny that took place from the late Cretaceous to the late Paleo- adaptation to new environments (e.g. Bousquet et al., 1992; Eckert cene (English and Johnston, 2004), and then during the Miocene et al., 2009). and the Pliocene, which originated most of the mountain ranges Although a higher proportion of differentially fixed variation at in central and southern Mexico (Hay and Soeding, 2002). The first these loci across isolated montane taxa can account for the higher appearance of temperate conifers in the fossil record of Mexico number of Abies taxa in the subtropics, it could also be hypothe- dates back to the early Miocene, about 23 Ma (Graham, 1999), sized that these figures are an artifact of phenotypic plasticity that which supports the hypothesis of an early southward expansion has mislead taxonomists. Such a possibility is supported by previ- and divergence (option 1). However, the first appearance of Abies ous analyses on terpene, allozyme, cytoplasmic DNA and morpho- pollen in this record only occurs until the Pliocene (5 Ma; Graham, metric variation, which found little evidence for distinguishing so 1976), which coincides with the hypothesis of a much later diver- many individual species within the Mesoamerican Abies complex gence for the Mesoamerican firs (option 2). (Aguirre-Planter et al., 2000; Nava-Cruz et al., 2006; Jaramillo-Cor- However, a detailed phylogeographic analysis of Mesoamerican rea et al., 2008; Strandby et al., 2009). The morphometric study of firs with cpSSRs suggest multiple southward migrations from Strandby et al. (2009) further reported a significant correlation be- North America, as two particular species, A. concolor and A. flinckii, tween anatomical dissimilarities and geographic distance in the appeared to diverge significantly from the others (Fig. 5). The dif- southern parts of this complex, which was interpreted as the action ferentiation of A. flinckii was previously interpreted as the result of contrasting environmental pressures on a widely distributed of a first migratory wave that entered Mexico via the Sierra Madre gene pool that has significant levels of phenotypic plasticity. Such Occidental, which then diverged in the western part of the Trans- a theory has already been invoked to explain anatomical and bio- Mexican Volcanic Belt (see Jaramillo-Correa et al., 2008 and refer- chemical variation in conifers, including Abies (e.g. Parker et al., ences therein). On the other hand, A. concolor might represent a rel- 1981; Huber et al., 2004; Baquedano et al., 2008), but then again, atively late arrival into northwestern Mexico, which could be as the integrative analysis of genomic, phenotypic and ecological Author's personal copy

272 É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 datasets, and common garden experiments seems necessary to test 2010). Thus, for threatened species complexes, such as the Meso- this hypothesis. american firs, defining evolutionary lineages with restricted gene flow and significant differences in adaptive traits at the population 4.4. Hybridization, ancestral polymorphism or lack of taxonomic level seems more appropriate (e.g. de Querioz, 1998; Fraser and resolution? Bernatchez, 2001). Accordingly, each one of the three cpSSR groups detected herein should represent independent lines of ancestry, Altogether, the results of this and previous studies (see above) while populations within these lineages correspond to isolated suggest that three main Abies lineages can be distinguished in Mes- units evolving under the action of stochastic forces for the last gen- oamerica: two of them corresponding to A. flinckii and A. concolor, erations. Each Mesoamerican fir population should be thus moni- and a third one formed by the remaining taxa. Such a sub-division tored for demographic treats such as inbreeding and genetic might further imply that a redefinition of the taxonomy of the drift, while conservation programs at a larger scale should be Mesoamerican species is necessary, as proposed by as Strandby based, for the time being, on the three main phylogeographic et al. (2009). Nevertheless, the lack of genetic differentiation seems groups detected herein. However, the integration of additional eco- to be a constant feature in Abies complexes across the world. For logical and landscape genetic data into gene-niche association gra- instance, in the eastern Mediterranean, no significant genetic dif- dients are urgently needed to improve the resolution of these ferences could be observed among A. bornmuelleriana, A. cephalo- groups, especially for the more threatened northern taxa. nica, A. equi-trojani, and A. nordmanniana with either mitochondrial or cpDNA markers (Liepelt et al., 2010), and similar Acknowledgments results were obtained for other complexes in North America (Pot- ter et al., 2010) and Asia (Wang et al., 2011). These comparisons The authors would like to thank J. Beaulieu, I. Gamache, (Cana- and the low diversification rates detected herein, point that these dian Forest Service), C. Sayre (VanDusen Botanical Garden), F.T. Le- groups have not had sufficient time of physical and genetic isola- dig (Univ. of California Davis), P. Delgado, D. Gernandt, A. Keiman, tion to fully differentiate, and that most of them have retained Y. Nava, C. Saenz, and G.R. Furnier (Institute of Biology and Ecology ancestral polymorphisms. This phenomenon has been further ob- of UNAM) for valuable help in sampling, and A. Gagné, S. Senne- served in many distantly related and reproductively isolated coni- ville, and S. Gérardi (Centre for Forest Research of Univ. Laval), fers, and it is likely favored by the large ancestral population sizes and J. Hicks and A. Kelly (Oregon State Univ.) for laboratory assis- and extensive gene flow characteristics of these trees (Bouillé and tance. The authors also thank U. Strandby and M. Sørensen for pro- Bousquet, 2005; Chen et al., 2010), and/or by recurrent introgres- viding samples from Guatemala. The authors’ gratitude is also sion in secondary contact zones between species with incomplete extended to S. Ramírez-Barahona for his help with some figures reproductive isolation, such as reported for many Mesoamerican and comments, and to N. Martínez-Méndez, D. Piñero, V. Souza species complexes (e.g. Matos and Schaal, 2000; Moreno-Letelier (Institute of Ecology of UNAM), and Susana Magallón (Institute of and Piñero, 2009; Peñaloza-Ramírez et al., 2010). Biology of UNAM) for comments and logistic support along the On the other hand, the fact that some species, including the project. Further valuable comments from two anonymous review- Mesoamerican A. concolor and A. flinckii, and the Asian A. gracilis, ers helped improve a previous version of the manuscript. Acknowl- A. holophylla and A. nephrolepis, do appear well delineated from a edgements are additionally extended to the Ministère du genetic point of view (Semerikova et al., 2011; Jiang et al., 2011; développement économique de l’innovation et de l’exportation present study), might imply that the number of DNA regions sur- du Québec, the Natural Sciences and Engineering Research Council veyed is actually insufficient to adequately differentiate all sur- of Canada (Discovery program), the Consejo Nacional de Ciencia y veyed taxa. Indeed, a recent Pinus phylogeny exhibited a much Tecnología (CONACYT, Grant SEP-2004-CO1-46475-Q), the Comi- higher resolution by using a massive sequencing approach on the sión Nacional para el Conocimiento y el uso de la Biodiversidad chloroplast genome (Parks et al., 2009). However, such studies (CONABIO, Grant B138), the Dirección General de Asuntos de Per- have to be placed into a more detailed population genetics context sonal Académico, UNAM, Programa de Apoyo a Proyectos de Inves- in order to capture all the putative interactions that closely-related tigación e Innovación Tecnológica, (PAPIIT Grant IN224309-3), and and putatively hybridizing taxa have experienced during their evo- the Programa de Apoyo a las Divisiones de Estudios de Postgrado, lution (e.g., Liston et al., 2007; Bouille´ et al., 2011). Such integration UNAM (PADEP-UNAM) in Mexico, for financial support. This paper would imply the massive sequencing of hundreds or even thou- was written during a sabbatical leave of L.E. Eguiarte at the Univer- sands of genomes, which is the standard sample size in phylogeo- sity of California Irvine with support of UC-MEXUS, CONACYT and graphic surveys, and which would render such studies currently DGAPA-UNAM. non-viable from an economical and bioinformatic point of view. More practical approaches include the one used herein, which inte- grates non-repetitive (i.e. DNA sequences) and repetitive (cpSSR) Appendix A. Supplementary material datasets into both a genus-wide and a geographically restricted context (Liston et al., 2007; Vachon and Freeland, 2010), and the Supplementary data associated with this article can be found, in use of multiple nuclear SSRs, which can be rapidly developed and the online version, at doi:10.1016/j.ympev.2011.09.021. analyzed, and provide direct evidence of isolation or hybridization events. References 4.5. Implications for conservation Aguirre-Planter, E., Furnier, G.R., Eguiarte, L.E., 2000. Low levels of genetic variation within and high levels of genetic differentiation among populations of species The lack of taxonomic consensus often collides with the conser- of Abies from southern México and Guatemala. American Journal of Botany 87, vation needs of megadiverse countries such as Mexico or Guate- 362–371. Andersen, U.S., Prado-Córdova, J.P., Sørensen, M., Kollmann, J., 2006. Conservation mala (Callmander et al., 2005). For instance, the indiscriminate and utilization of Abies guatemalensis Rehder (Pinaceae) – an endangered grouping of species into one single taxon has left endangered spe- endemic conifer in Central America. Biodiversity and Conservation 15, 3131– cies unprotected (e.g., Rieseberg, 1991), while hybrid or introgres- 3151. Anderson, J.B., Browler, L.P., 1996. Freeze-protection of overwintering monarch sant populations that do not have conservation status under many butterflies in México: critical role of the forest as a blanket and an umbrella. legislations have also been left unguarded (e.g. Fitzpatrick et al., Ecological Entomology 21, 107–116. Author's personal copy

É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274 273

Axelrod, D.I., 1976. History of coniferous forests, California and Nevada. University Graham, A., 1999. The tertiary history of the northern temperate element in the of California Publications in Botany 70, 1–62. northern Latin American biota. American Journal of Botany 86, 32–38. Baquedano, F.J., Valladares, F., Castillo, F.J., 2008. Phenotypic plasticity blurs Griffin, J.R., Critchfield, W.B., 1972. The Distribution of Forest Trees in California. ecotypic divergence in the response of Quercus coccifera and Pinus halepensis USDA For. Serv. Re. Pap. PSW-82. Pacific Southwest For. Range Exp. Sta., to water stress. European Journal of Forest Research 127, 495–506. Berkeley, USA. Bouillé, M., Bousquet, J., 2005. Trans-species shared polymorphisms at orthologous Grotkopp, E., Rejmánek, M., Sanderson, M.J., Rost, T.L., 2004. Evolution of genome nuclear gene loci among distant species in the conifer Picea (Pinaceae): size in pines (Pinus) and its life-history correlates: supertree analyses. Evolution implications for the long-term maintenance of genetic diversity in trees. 58, 1705–1729. American Journal of Botany 92, 63–73. Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and Bouille´, M., Senneville, S., Bousquet, J., 2011. Discordant mtDNA and cpDNA analysis program fro Windows95/98/NT. Nucleic Acids Symposium Series 41, phylogenies indicate geographic speciation and reticulation as driving factors 95–98. for the diversification of the genus Picea. Tree Genetics and Genomes 7, 469– Hay, W.W., Soeding, E., 2002. The Cenozoic uplift-climate change paradox. 484. International Journal of Earth Sciences 91, 746–774. Bousquet, J., Strauss, S.H., Doerksen, A.H., Price, R.A., 1992. Extensive variation in Huber, D.W., Ralph, S., Bohlmann, J., 2004. Genomic hardwiring and phenotypic evolutionary rate of rbcL gene sequences among seed plants. Proceedings of the plasticity of terpenoid-based defences in conifers. Journal of Chemical Ecology National Academy of Sciences United States of America 89, 7844–7848. 30, 2399–2418. Callmander, M.W., Schatz, G.E., Lowry II, P.P., 2005. IUCN red list assessment and the Huelsenbeck, J.P., Ronquist, F.R., 2001. MRBAYES. Bayesian inference of global strategy for plant conservation: taxonomists must act now. Taxon 54, phylogenetic trees. Bioinformatics 17, 754–755. 1047–1050. Jaramillo-Correa, J.P., Aguirre-Planter, E., Khasa, D.P., Eguiarte, L.E., Piñero, D., Carstens, B.C., Soute, H.N., Reid, N.M., 2009. An information-theoretical approach to Furnier, G.R., Bousquet, J., 2008. Ancestry and divergence of subtropical phylogeography. Molecular Ecology 18, 4270–4282. montane forest isolates: molecular biogeography of the genus Abies Chen, J., Källman, T., Gyllestrand, N., Lascoux, M., 2010. New insights on the (Pinaceae) in southern Mexico and Guatemala. Molecular Ecology 17, 2476– speciation history and nucleotide diversity of three boreal spruce species and a 2490. tertiary relict. Heredity 104, 3–14. Jaramillo-Correa, J.P., Beaulieu, J., Khasa, D.P., Bousquet, J., 2009. Inferring the past Clement, M.D., Posada, D., Crandall, K., 2000. TCS: a computer program to estimate from the present phylogeographic structure of North American forest trees: gene genealogies. Molecular Ecology 9, 1657–1660. seeing the forest for the genes. Canadian Journal of Forest Research 39, 286– Corander, J., Marttinen, P., 2006. Bayesian identification of admixture events using 307. multilocus molecular markers. Molecular Ecology 15, 2833–2843. Jaramillo-Correa, J.P., Verdú, M., González-Martínez, S.C., 2010. The contribution of Corander, J., Wladmann, P., Sillanpää, M.J., 2003. Bayesian analysis of population recombination to heterozygosity differs among plant evolutionary lineages and differentiation between populations. Genetics 163, 367–374. life-forms. BMC Evolutionary Biology 10, 22. Crandall, K.A., Templeton, A.R., 1993. Empirical tests of some predictions from Jiang, Z.Y., Peng, Y.L., Hu, X.X., Zhou, Y.F., Liu, J.Q., 2011. Cytoplasmic DNA coalescent theory with applications to infraspecific phylogeny reconstruction. variation in and genetic delimitation of Abies nephrolepis and Abies Genetics 134, 2008–2012. holophylla in norheastern China. Canadian Journal of Forest Research 41, Critchfield, W.B., 1988. Hybridization of the California firs. Forest Sciences 34, 139– 1555–1561. 151. Kremp, G.O.W., 1967. Catalog of Fossil Spores and Pollen, vol. 26. The Pensylvania De Kroon, H., Huber, H., Stuefer, J.F., Van Groenendael, J.M., 2005. A modular State University, Pennsylvania, USA. concept of phenotypic plasticity in plants. New Phytologist 166, 73–82. Kormut’ák, A., Lee, S.-W., Hong, K.-N., Yang, B.-H., Hong, Y.-P., 2008. Crossability de Querioz, K., 1998. The general lineage concept of species, species criteria, and the relationships between Korean firs Abies koreana, A. nephrolepis, and A. holophylla process of speciation. In: Howard, D.J., Berlocher, S.H. (Eds.), Endless Forms: and some other representatives of the genus Abies. Biologia 63, 94–99. Species and Speciation. Oxford University Press, Oxford, UK, pp. 57–75. Lancaster, L.T., 2010. Molecular evolutionary rates predict both extinction and Dickinson, W.R., 2004. Evolution of the North American Cordillera. Annual Review speciation in temperate angiosperm lineages. BMC Evolutionary Biology 10, of Earth and Planetary Sciences 32, 13–45. 162. Donahue, J.K., Dvorak, W.S., Gutiérrez, E.A., Kane, M.B., 1985. Abies guatemalensis:A LePage, B.A., 2003. The evolution, biogeography and ecology of the Pinaceae based Two Year Status Report. Bulletin of Tropical Forestry Number 3. CAMCORE. on fossil and extant representatives. Acta Horticulturae 615, 29–52. School of Forest Resources, North Carolina State University, Raleigh, North Liepelt, S., Mayland-Quellhorst, E., Lahme, M., Ziegenhagen, B., 2010. Contrasting Carolina, USA. geographical patterns of ancient and modern genetic lineages in Mediterranean Eckert, A.J., Bower, A.D., Wergzyn, J.L., Pande, B., Jermstad, K.D., Krutovsky, K.V., Abies species. Plant Systematics and Evolution 284, 141–151. Clair, St.J.B., Neale, D.B., 2009. Association genetics of coastal Douglas fir Liston, A., Parker-Defeniks, M., Syring, J.V., Willyard, A., Cronn, R., 2007. Interspecific (Pseudotsuga menziessi var. menziesii, Pinaceae). I. Cold-hardiness related traits. phylogenetic analysis enhances intraspecific phylogeographical inference: a Genetics 182, 1289–1302. case study in Pinus lambertiana. Molecular Ecology 16, 3926–3937. English, J.M., Johnston, S.T., 2004. The Laramide orogeny: what were the driving Liu, T.-S., 1971. A Monograph of the Genus Abies. Department of Forestry, College of forces? International Geological Review 46, 833–838. Agriculture, National Taiwan University, Taipei, Taiwan, China. Erwin, D.M., Schorn, H.E., 2005. Revision of the conifers from the Eocene Thunder Lozano-García, M.S., Ortega-Guerrero, B., Caballero-Miranda, M., Urrutia- Mountain flora, Idaho. USA Review of Paleobotany and Palynology 137, 125– Fucugauchi, J., 1993. Late Pleistocene and Holocene paleoenvironments of 145. Chalco Lake, central México. Quaternary Research 40, 332–342. Espinosa-Organista, D., Ocegueda Cruz, S., Aguilar-Zu´ ñiga, C., Flores Villela, O., Magallón, S., Sanderson, M.J., 2001. Absolute diversification rates in angiosperm Llorente-Bousquets, J., 2008. El conocimiento biogeográfico de las especies y su clades. Evolution 55, 1762–1780. regionalización natural. In Soberón, J., Halffter, G. y Llorente-Bousquets, J. (Eds.), Manni, F., Guérard, D., Heyer, E., 2004. Geographic patterns of (genetic, Capital natural de México, vol. I. Conocimiento actual de la biodiversidad. morphologic, linguistic) variation: how barriers can be detected by using CONABIO, México, pp. 33–65. Monmonier’s algorithm. Human Biology 76, 173–190. Excoffier, L., Lischer, H.E.L., 2010. Arlequin suite ver 3.5: a new series of programs to Mao, K., Hao, M., Liu, J., Adams, R.P., Milne, R.I., 2010. Diversification and perform population genetics analyses under Lynux and Windows. Molecular biogeography of Juniperus (Cupressaceae): variable diversification rates and Ecology Resources 10, 564–567. multiple intercontinental dispersals. New Phytologist 188, 254–272. Farjon, A., 1990. Pinaceae: Drawings and Descriptions of the Genera Abies, Cedrus, Martínez, M., 1948. Los Abies mexicanos. Anales del Instituto de Biología de México Pseudolarix, Keteleeria, Nothotsuga, Tsuga, Cathaya, Pseudotsuga, Larix and Picea. 19, 11–104. Koelt Scientific Books, Königstein, Germany. Matos, J.A., Schaal, B.A., 2000. Chloroplast evolution in the Pinus montezumae Farjon, A., 2005. A Monograph of Cupressaceae and Sciadopitys. Kew: Royal complex: a coalescent approach to hybridization. Evolution 54, 1218–1233. Botanical Gardens, Kew, UK, 643 p. Miller, K.G., Browning, J.V., Aubry, M.-P., Wade, B.S., Katz, M.E., Kulpecz, A.A., Farjon, A., Rushforth, K.D., 1989. A classification of Abies Miller (Pinaceae). Notes Wright, J.D., 2008. Eocene–Oligocene global climate and sea-level changes: St. from the Royal Botanical Garden Edinburgh 46, 59–79. Stephens Quarry, Alabama. Geological Society of America Bulletin 120, 34–53. Fitzpatrick, B.M., Johnson, J.R., Kump, K.D., Smith, J.J., Voss, S.R., Shaffer, H.B., 2010. Milne, R.I., Abbot, R.J., 2002. The origin and evolution of tertiary relict floras. Rapid spread of invasive genes into a threatened native species. Proceedings of Advances in Botanical Research 38, 281–314. the National Academy of Sciences United States of America 107, 3606–3610. Moreno-Letelier, A., Piñero, D., 2009. Phylogeographic structure of Pinus strobiformis Fraser, D.J., Bernatchez, L., 2001. Adaptive evolutionary conservation: towards a Engelm across the Chihuahuan Dessert filter-barrier. Journal of Biogeography unified concept for defining conservation units. Molecular Ecology 10, 2741– 36, 121–131. 2752. Morgante, M., 2006. Plant genome organization and diversity: the year of the junk! Gernandt, D.S., Magallón, S., López, G.G., Flores, O.Z., Willyard, A., Liston, A., 2008. Current Opinion in Biotechnology 17, 168–173. Use of simultaneous analyses to guide fossil-based calibrations of Pinaceae Nava-Cruz, Y., Espinosa-García, F., Furnier, G.R., 2006. Niveles y patrones de phylogeny. International Journal of Plant Sciences 169, 1086–1099. variación química en resinas de las hojas del género Abies del norte de México. Gillman, L.N., Keeling, D.J., Gardner, R.C., Wrigh, S.D., 2010. Faster evolution of Agrociencia 40, 229–238. highly conserved DNA in tropical plants. Journal Evolutionary Biology 23, 1327– Oline, D.K., 2008. Geographic variation in chloroplast haplotypes in the California 1330. red fir-noble fir species complex and the status of Shasta red fir. Canadian Glor, R.E., 2010. Phylogenetic insights on adaptive radiation. Annual Review of Journal of Forest Research 38, 2705–2710. Ecology Evolution and Systematics 41, 251–270. Parker, W.H., Maze, J., Brafield, G.E., 1981. Implications of morphological and Graham, A., 1976. Studies in neotropical paleobotany. II. The Miocene communities anatomical variation in Abies balsamea and A. lasiocarpa (Pinaceae) from of Veracuz, México. Annals of the Missouri Botanical Garden 64, 787–842. western Canada. American Journal of Botany 68, 843–854. Author's personal copy

274 É. Aguirre-Planter et al. / Molecular Phylogenetics and Evolution 62 (2012) 263–274

Parks, M., Cronn, R., Liston, A., 2009. Increasing phylogenetic resolution at low Suyama, Y., Yoshimaru, H., Tsumura, Y., 2000. Molecular phylogenetic position of taxonomic levels using massively parallel sequencing of chloroplast genomes. Japanese Abies (Pinaceae) based on chloroplast DNA sequences. Molecular BMC Biology 7, 84. doi:10.1186/1741-7007-7-84. Phylogenetics and Evolution 16, 271–277. Peñaloza-Ramírez, J.M., González-Rodríguez, A., Mendoza-Cuenca, L., Caron, H., Swofford, D.L., 2000. PAUP⁄. Phylogenetic Analysis Using Parsimony (Ã and other Kremer, A., Oyama, K., 2010. Interspecific gene flow in a multispecies oak hybrid methods). Version 4. Sinauer Associates, Sunderland Massachusetts, USA. zone in the Sierra Tarahumara of Mexico. Annals of Botany 105, 389–399. Templeton, A.R., Crandall, K.A., Sing, C.F., 1992. A cladistic analysis of phenotypic Perron, M., Bousquet, J., 1997. Natural hybridization between black spruce and red associations with haplotypes inferred from restriction endonuclease mapping spruce. Molecular Ecology 6, 725–734. and DNA sequence data. III. Cladogram estimation. Genetics 132, 619–633. Petit, R.J., Hampe, A., 2006. Some evolutionary consequences of being a trees. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the Annual Review in Ecology Evolution and Systematics 37, 187–214. sensitivity of progressive multiple sequence alignment through sequence Potter, K.M., Frampton, J., Josserand, S.A., Nelson, C.D., 2010. Evolutionary history of weighting, position specific gap penalties and weight matrix choice. Nucleic two endemic Appalachian conifers revealed using microsatellite makers. Acids Research 22, 4673–4680. Conservation Genetics 11, 1499–1513. Vachon, N., Freeland, J.R., 2010. Phylogeographical inferences from chloroplast Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. DNA: quantifying the effects of mutations in repetitive and non-repetitive Bioinformatics 14, 817–818. sequences. Molecular Ecology Resources 11, 279–285. Prada, C., Schizas, N.V., Yoshioka, P.M., 2008. Phenotypic plasticity or speciation? A Vázquez-Lobo, A., 1996. Evolución de hongos endófitos del género Pinus. Tesis de case from a clonal marine organism. BMC Evolutionary Biology 8, 47. licenciatura, Facultad de Ciencias, Universidad Nacional Autónoma de México, doi:10.1186/1471-2148-8-47. México, DF. Richards, C.L., Carstens, B.C., Knowles, L.L., 2007. Distribution modelling and Vázquez-Selem, L., Heine, K., 2004. Late quaternary glaciations of México. In: Ehlers, statistical phylogeography: an integrative framework for generating and J., Gibbard, P.L. (Eds.), Quaternary Glaciations – Extent and Chronology, Part III: testing alternative biogeographical hypotheses. Journal of Biogeography 34, South America, Asia, Africa, Australia, Antarctica. Elsevier, Amsterdam, The 1833–1845. Netherlands, pp. 223–242. Rieseberg, L.H., 1991. Hybridization in rare plants: insight from case studies in Vendramin, G.G., Lelli, L., Rossi, P., Morgante, M., 1996. A set of primers for the Cercocarpus and Helianthus. In: Falk, D.A., Holsinger, K.E. (Eds.), Genetics and amplification of 20 chloroplast microsatellites in Pinaceae. Molecular Ecology 5, Conservation of Rare Plants. Oxford University Press, New York, pp. 171–181. 111–114. Rohatgi, V.K., 1976. An Introduction to Probability Theory and Mathematical Verdú, M., 2002. Age at maturity and diversification in woody angiosperms. Statistics. Wiley, New York. Evolution 56, 1352–1361. Rushforth, K.D., 1989. Two new species of Abies (Pinaceae) from western Mexico. Wang, J., Abbot R.J., Peng, Y.L., Du, F.K., Liu, J.-Q., 2011. Species delimitation and Notes from the Royal Botanical Garden Edinburgh 46, 101–109. biogeography of two fir species (Abies) in central China: cytoplasmic DNA Rzedowski, J., 1978. Vegetación de México. Editorial Limusa, Mexico, 432 p. variation. Heredity 107, 362–370. Rzedowski, J., 1993. Diversity and origins of the phanerogamic flora of Mexico. In: Willyard, A., Syring, J., Gernandt, D.S., Liston, A., Cronn, R., 2007. Fossil calibration Ramamoorthy, T.P., Bye, R., Lot, A., Fa, J. (Eds.), Biological Diversity of Mexico: and molecular divergence infers a moderate mutation rate and recent Origins and Distribution. Oxford University Press, New York, USA, pp. 129–144. radiations for Pinus. Molecular Biology and Evolution 24, 90–101. Sanderson, M.J., 2002. Estimating absolute rates of molecular evolution and Wilson, J.S., Pitts, S.P., 2010. Illuminating the lack of consensus among descriptions divergence times: a penalized likelihood approach. Molecular Biology and of earth history data in the North American deserts: a resource for biologists. Evolution 19, 101–109. Progress in Physical Geography 34, 419–441. Sanderson, M.J., 2003. R8s: inferring absolute rates of molecular evolution and Wright, S., Keeling, J., Gillman, L.N., 2006. The road from Santa Rosalia: a faster divergence times in the absence of a molecular clock. Bioinformatics 19, 301– tempo of evolution in tropical climates. Proceedings of the National Academy of 302. Sciences United States of America 103, 7718–7722. SEMARNAT, 2010. Norma Oficial Mexicana NOM-059-SEMARNAT-2010. Protección Xiang, Q.-P., Xiang, Q.-Y., Liston, A., Zhang, X.-C., 2004. Phylogenetic relationships in ambiental de especies nativas de México de flora y fauna silvestre – Anexo Abies (Pinaceae): evidence from PCR–RFLP of the nuclear ribosomal DNA Normativo III: Lista de especies en riego. SEMARNAT, UNDP. internal transcribed spacer region. Botanical Journal of the Linnean Society 145, Semerikova, S.A., Semerikov, V.L., Lascoux, M., 2011. Post-glacial history and 425–435. introgression in Abies (Pinaceae) species of the Russian Far East inferred from Xiang, X., Cao, M., Zhou, Z., 2007. Fossil history and modern distribution of the both nuclear and cytoplasmic markers. Journal of Biogeography 38, 326–340. genus Abies (Pinaceae). Frontiers of Forestry in China 2, 355–365. Spicer, R., Harris, N., Widdowson, M., Herman, A., Guo, S., Valdes, P., 2003. Constant Xiang, Q.-P., Xiang, Q.-Y., Guo, Y.-Y., Zhang, X.-C., 2009. Phylogeny of Abies elevation of southern Tibet during over the past 15 million years. Nature 421, (Pinaceae) inferred from nrITS sequence data. Taxon 58, 141–152. 622–624. Xu, T., Abbot, R.J., Milne, R.I., Mao, K., Du, F.K., Wu, G., Ciren, Z., Miehe, G., Liu, J., Stamatakis, A., Hoover, P., Rougemont, J., 2008. A rapid bootstrap algorithm for the 2010. Phylogeography and allopatric divergence of cypress species (Cupresssus RAxML web servers. Systematics Biology 57, 758–771. L.) in the Qinghai–Tibetan Plateau and adjacent regions. BMC Evolutionary Strandby, U., Christensen, K.I., Sørensen, M., 2009. A morphometric study of the Biology 10, 194. Abies religiosa–hickelii–guatemalensis complex (Pinaceae) in Guatemala and Zachos, J.C., Pagani, M., Sloan, E.T., Billups, K., 2001. Trends, rhythms and aberrations Mexico. Plant Systematics and Evolution 280, 59–76. in global climate 65 Ma to present. Science 292, 686–694.