The Atlantic–Mediterranean watershed, river basins and glacial history shape the genetic structure of Iberian poplars

Ana I de Lucas

Molecular Ecology (2012) 21, 3593–3609 doi: 10.1111/j.1365-294X.2012.05619.x The Atlantic–Mediterranean watershed, river basins and glacial history shape the genetic structure of Iberian poplars D . M A C A Y A - S A N Z , * M . H E U E R T Z , * U . L Ó P E Z - D E - H E R E D I A , † A . I . D E - L U C A S , ‡ E . H I D A L G O , § C . M A E S T R O , – A . P R A D A , * * R . A L Í A * and S . C . G O N Z Á L E Z - M A R T Í N E Z * *Department of Forest Ecology and Genetics, Forest Research Centre, INIA. Carretera de A Coruña km 7.5 – 28040 Madrid, Spain, †Forest Genetics and Physiology Research Group, ETS Forestry Engineering, Technical University of Madrid (UPM), Ciudad Universitaria s ⁄ n – 28040 Madrid, Spain, ‡Laboratorio de Diagnóstico Genético, Departamento de Biotecnologı́a, ITAGRA.CT, Campus Universitario ‘‘La Yutera’’, Avenida de Madrid 44 – 34004 Palencia, Spain, §Departamento de Producción Vegetal y Recursos Forestales, E.T.S. Ingenierı́as Agrarias, Universidad de Valladolid, Avenida de Madrid 44 – 34004 Palencia, Spain, –Unidad de Recursos Forestales, CITA, Gobierno de Aragón, Avenida de Montañana 930 – 50059 Zaragoza, Spain, **Banc de Llavors Forestals, Dirección General de Gestión del Medio Natural, Conselleria de Territorio y Vivienda, Avda. Comarques del Paı́s Valencià 144 – 46913 Quart de Poblet, Valencia, Spain Abstract Recent phylogeographic studies have elucidated the effects of Pleistocene glaciations and of Pre-Pleistocene events on populations from glacial refuge areas. This study investigates those effects in riparian trees (Populus spp.), whose particular features may convey enhanced resistance to climate fluctuations. We analysed the phylogeographic structure of 44 white (Populus alba), 13 black (Populus nigra) and two grey (Populus x canescens) poplar populations in the Iberian Peninsula using plastid DNA microsatellites and sequences. We also assessed fine-scale spatial genetic structure and the extent of clonality in four white and one grey poplar populations using nuclear microsatellites and we determined quantitative genetic differentiation (QST) for growth traits in white poplar. Black poplar displayed higher regional diversity and lower differentiation than white poplar, reflecting its higher cold-tolerance. The dependence of white poplar on phreatic water was evidenced by strong differentiation between the Atlantic and Mediterranean drainage basins and among river basins, and by weaker isolation by distance within than among river basins. Our results suggest confinement to the lower river courses during glacial periods and moderate interglacial gene exchange along coastlines. In northern Iberian river basins, white poplar had lower diversity, fewer private haplotypes and larger clonal assemblies than in southern basins, indicating a stronger effect of glaciations in the north. Despite strong genetic structure and frequent asexual propagation in white poplar, some growth traits displayed adaptive divergence between drainage and river basins (QST > FST), highlighting the remarkable capacity of riparian tree populations to adapt to regional environmental conditions. Keywords: genetic differentiation, glaciations, Iberian Peninsula, Populus, QST > FST, spatial genetic structure. Received 15 December 2011; revision received 16 March 2012; accepted 27 March 2012 Introduction Historical demographic processes caused by the Pleistocene glaciations have contributed to shape the current Correspondence: Santiago C. González-Martı́nez, Fax: +34 913572293; E-mail: santiago@inia.es 2012 Blackwell Publishing Ltd patterns of phylogeographic structure in widespread temperate tree species. As the location of glacial refugia and the ways postglacial colonization took place became elucidated (Nichols & Hewitt 1994; Palmé et al. 2003), researchers increasingly started focusing on phylogeographic patterns within former glacial refugia such 3594 D . M A C A Y A - S A N Z E T A L . as the Mediterranean peninsulas (Rodrı́guez-Sánchez et al. 2010). The existence of numerous refugia in the Mediterranean Peninsulas is believed to have enabled species and populations to persist in these buffered habitats throughout the Quaternary (Medail & Diadema 2009). Even in the southern regions that experienced the severest conditions (e.g. the north-western Iberian Peninsula), Mediterranean species could have persisted in isolated benign locations, similar to the ‘cryptic refugia’ described for boreal species (Anderson et al. 2006; Cheddadi et al. 2006; Petit et al. 2008). Typical signatures of former refuge populations are low withinpopulation genetic diversity, accompanied by large amounts of regional diversity, and high levels of genetic differentiation among populations (Hampe & Petit 2005). Besides the effects of the Quaternary glaciations, some recent studies have also highlighted the importance of older geological events, such as Tertiary plate tectonics, in shaping the current phylogeographic structure of forest trees in former refugial areas (e.g. Quercus suber, Magri et al. 2007; Taxus baccata, González-Martı́nez et al. 2010). Although there is a large body of phylogeographic studies on former glacial refugia regions (see review in Medail & Diadema 2009 for the Mediterranean basin), few have focused on riparian temperate species, whose particular attributes could have enhanced their resilience to past climate changes. First, as riparian temperate species performance depends largely on a single environmental condition, phreatic water availability, that is more related to orography than to climate, climate factors may affect them less than to other plant species. Second, their preferred habitats (valley bottoms, wetlands and deep gorges) are considered ideal to buffer climatic oscillations due to warmer and moister conditions, making them good candidates for ‘refugia within the refugia’ (Medail & Diadema 2009; NietoFeliner 2011). Third, many typical temperate riparian species (e.g. Populus spp., Salix spp., Tamarix spp., Ulmus spp.; Stuefer et al. 2002; Ruiz de la Torre 2006; Slavov & Zhelev 2010) have high levels of clonality, which could reinforce population survival by securing local persistence through unfavourable conditions and allowing rapid colonization of disturbed areas. As a drawback, the dependence of riparian trees on phreatic water leads to a scattered pattern of suitable habitats, separated by large inhospitable areas (e.g. elevated plateaus between river valleys). As a result, riparian populations are exceptionally prone to isolation, and consequently substantial genetic structure has been reported at regional level in different riparian trees (e.g. Cottrell et al. 2005; Fussi et al. 2010). Pervasive population isolation promotes stochastic processes reducing both molecular and quantitative genetic variation (at a rate that depends on the reciprocal of effective population size). Depletion of genetic variation reduces the ability of populations to adapt. Theoretical models have shown an ambiguous role of gene exchange in local adaptation. Gene flow counteracts the effects of selection, as it dilutes local changes in allele frequencies (Lenormand 2002). However, Alleaume-Benharira et al. (2006) showed, using individualbased simulations, that gene flow can also mitigate the effect of drift by replenishing genetic variance in small marginal populations. In species with more specialized ecological requirements, the gene flow from core populations necessary to ensure adaptability of isolated marginal populations is expected to increase (AlleaumeBenharira et al. 2006). As specialized species are prone to geographic isolation but, at the same time, inhabit similar environments across large ranges, gene flow may be critical to keep levels of genetic variance high enough to maintain their evolutionary potential. To our knowledge, no study has hitherto reported on the adaptive consequences derived from the ecological specificity of riparian tree species. In this study, we assessed the genetic diversity and structure of wind-dispersed Iberian poplar species (especially white poplar, Populus alba L.) at local, regional and wide spatial scales using chloroplast and nuclear DNA markers. Additionally, common garden data in white poplar provided insights into the adaptive significance of river-basin isolation in this species. The Iberian Peninsula (IP hereafter) represents an ideal setting for this study, as it harbours numerous refuge areas with distinct environmental features (Gómez & Lunt 2007 and references therein). With regard to riparian habitats, high climatic heterogeneity is accompanied by a complex river system consisting of two drainage basins (watershed of the Atlantic Ocean and the Mediterranean Sea), several main river basins and numerous smaller watercourses originating in coastal mountain ranges. Different parts of the IP currently inhabited by poplar species (notably the Duero basin, in the northwestern range) were severely affected by Quaternary glaciations and present particularly amenable conditions for testing the persistence of riparian species in harsh environments. Within this range, we focused on the white poplar, a widespread species with high colonization capability and marked tolerance to temperature changes, atmospheric dryness and salt stress, if groundwater is available (Ruiz de la Torre 2006). Despite the scarcity of palynological records resulting from poor pollen preservation (Huntley & Birks 1983), leaf fossils of white poplar found in travertine formations have shown its undoubted native presence in the IP (Roiron et al. 2004). Furthermore, in contrast with other European poplar species, white poplar has limited utility to 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3595 humans and negligible commercial value. Human mediated movement of reproductive material is therefore unlikely to have modified the pattern of natural genetic structure in this species. Analysing different aspects of genetic diversity and spatial genetic structure (SGS) in Iberian poplars, we aimed at clarifying the role of climatic fluctuations and orographic barriers on population dynamics in riparian species. The use of different types of molecular markers [plastid DNA (cpDNA) sequences and microsatellites (cpSSRs) and nuclear microsatellites (nSSRs)] allowed us to discriminate among distinct overlapping patterns of SGS and to control for allele homoplasy. The comparison of neutral genetic differentiation patterns to quantitative traits facilitated understanding the role of isolation in promoting local adaptation. More specifically, our goals were to (i) examine the genetic signature of ancient geological divides (the flooding of the Strait of Gibraltar and the rise of the Mediterranean ⁄ Atlantic watershed), setting a temporal framework for main phylogeographical events; (ii) assess regional patterns of diversity and differentiation, informing on the capability of riparian species to survive severe climatic changes in situ and to migrate across extensive inhospitable areas; (iii) evaluate the role of asexual reproduction and fine-scale genetic structure for maintaining population persistence and connectivity within river basins in a water-dependent species; and (iv) test for signs of local adaptation based on genetic differentiation for quantitative traits (as estimated by QST). The comparison of levels of genetic differentiation for molecular markers and quantitative traits addressed the specific question of adaptive divergence vs. genetic drift in a narrow-niche but widespread species. Materials and methods Plant material, sampling and DNA extraction Fifty-nine Iberian poplar populations (n = 628 trees) were sampled (see Fig. 1, details in Table S1, Supporting information), with a focus on Iberian white poplar (Populus alba L.; 44 populations), and representative samples of black (Populus nigra L.; 13 populations) and grey (Populus x canescens (Aiton) Sm.; two populations) poplars. Black poplar was not sampled in the south of the IP as it is relatively scarce in that region. Sampling included three major river basins, two draining to the Atlantic Ocean (Duero and Guadalquivir rivers, in northwestern and southern Spain, respectively) and one to the Mediterranean (Ebro river, northeastern Spain). Several smaller Mediterranean river basins and scattered populations in two additional major Atlantic river basins (Tajo and Guadiana) were also sampled. In 2012 Blackwell Publishing Ltd addition, one white poplar population from the High Atlas in Morocco was sampled to obtain time calibrations based on major biogeographic events separating the Iberian Peninsula from northern Africa (see below). For each population, six leaves from each of n = 10 trees, separated by at least 100 m (to reduce the chance of sampling clonal replicates or related trees), were collected and dried in silica gel prior to cpDNA analysis. To study genetic structure at the local scale using nuclear microsatellites (nSSRs), one population of grey poplar (from the Guadiana river) and four of white poplar (from the Duero, Guadalquivir and Ebro basins; Fig. 1) were more intensively sampled, collecting material from three to five additional trees around each of the ten core individuals (n = 200 trees). All trees were geographically referenced. Total DNA was purified from dried leaves following a slightly modified protocol from Doyle & Doyle (1990). Molecular markers (cpSSRs, nSSRs and cpDNA sequences) Thirteen chloroplast microsatellites (cpSSRs) from Weising & Gardner (1999) and Bryan et al. (1999) were tested in a panel consisting of 17 individuals sampled across the white poplar range in the Iberian Peninsula. Out of 11 cpSSRs that produced a PCR product, only two (ccmp2 and ccmp5) were polymorphic. All samples were amplified at ccmp2 and ccmp5 (missing data of ca. 6%) in 10 lL of final volume, including 5 ng of DNA template, 0.4 units of Taq (Bioline, London, UK), 0.15 lM of each primer (the forward primers labelled with IRD800; Li-Cor Biosciences, Lincoln, NE, USA), 0.1 mM of each dNTP and 2 mM of MgCl2. The PCR profile consisted of 5 min at 94 C, 12–24 cycles (samples with weak amplification at 12 cycles were repeated using 24 cycles) of 1 min at 94 C, 30 s at 49 C (ccmp2) or 50 C (ccmp5), and 1 min at 72 C and a final extension of 10 min at 72 C. PCR fragments were resolved on a Li-Cor 4300 DNA analyser (Li-Cor Biosciences). To reduce the probability of scoring errors, a selection of samples that covered the fragment size range was included as internal standard in each gel. SAGAGT vs. 3.3. was used for gel calibration and scoring (Li-Cor Biosciences). The chloroplast DNA region trnC-petN1 was sequenced in at least one individual per population and cpSSR haplotype (n = 133 individuals), assuming that individuals with the same cpSSR haplotype within populations would also share their trnC-petN1 haplotype, because of lower expected mutation rates in the latter. For each combination of cpSSRs and trnC-petN1, the rpl16-poprpl cpDNA region was sequenced in at least one individual (n = 107 individuals). To sequence the samples in both directions, 30 lL of PCR product 3596 D . M A C A Y A - S A N Z E T A L . (a) Populus alba h03 h04 h05 Ebro basin h06 h07 h08 Duero basin h09 h10 Catalonia h11 h12 h20 h22 h23 h24 h25 Tajo basin h26 h27 h28 h29 Populus nigra h36 h37 Atlantic drainage basin Levante Guadiana basin Mediterranean drainage basin h01 h38 h02 h39 h13 h40 h14 h41 h15 h42 h16 h43 h17 h44 h45 h18 Guadalquivir basin h30 h31 h46 Populus x canescens h51 h33 h52 h34 0 50 h47 h50 h32 h35 h04 h48 h27 h49 h51 100 200 h53 h54 h55 300 km. (b) Ebro basin Duero basin Catalonia Tajo basin Atlantic drainage basin Levante Guadiana basin Guadalquivir basin Populus alba Populus nigra Mediterranean drainage basin Populus x canescens H11 50 H02 H03 H10 0 H01 H04 H12 H01 H05 H13 H02 H06 H14 H04 100 200 H07 300 km. Fig. 1 Geographic distribution and population frequency of haplotypes based on (a) the full data set (cpSSRs and cpDNA sequences) and (b) unique event polymorphisms (UEPs). Red squares indicate populations used to study clonality levels and finescale population genetic structure. Main hydrographic features and the altitudinal pattern (in shadows) are also shown. 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3597 was yielded for each sample and cpDNA region. The PCR mix included 7.5 ng of DNA template, 0.75 units of Taq (Bioline), 0.3 lM of each primer (unlabelled), 0.15 mM of each dNTP and 3 mM of MgCl2. The PCR profile was 3 min at 94 C, 39 cycles of 30 s at 94 C, 30 s at 50 C and 80 s at 72 C, and a final extension of 10 min at 72 C. Fragments were purified with Exo-SAP (Affymetrix, Santa Clara, CA, USA), and sequenced (standard Sanger sequencing) in external facilities (Secugen, Madrid). SEQMAN software (DNASTAR Inc., Madison, Wisconsin, USA) was used to edit and align cpDNA sequences. Five highly-polymorphic nSSRs (ORPM127, ORPM312, PMGC2852, ORPM30 and ORPM344; Lexer et al. 2005) were used to study genetic structure at the local scale. Fourteen additional nSSRs (see Table S2, Supporting information) were used to confirm clonal identity of large clonal assemblies. Protocols for amplification and fragment resolution were the same as for cpSSRs with the following PCR modifications: 2 min initial denaturation, 30 s denaturation during cycles, 5 min final elongation, and annealing temperatures and number of cycles as given in Table S2 (Supporting information). polymorphism types [insertions ⁄ deletions (indels), SSRs, SNPs or short tandem repeats (STRs)] were weighted identically. For comparative purposes, haplotypes from 12 Iberian and French aspens (Populus tremula) were added to the network. Time to the most recent common ancestor in white poplar To obtain insights into the temporal scale of relevant phylogenetic events in white poplar, we obtained estimates of the times to the most recent common ancestor (TMRCAs) for different data sets, based on the standard coalescence (constant size) model and Bayesian analysis using BATWING (Wilson et al. 2003). Salix sp. was used as outgroup (GenBank accessions AJ849602.1, DQ875043, DQ875044.1, DQ875047.1, DQ875048.1 and DQ875049.1). Analyses considered UEP sites to define tree topology and included (set1) all white poplar haplotypes (including Morocco), (set2) all Iberian white poplar haplotypes (excluding Morocco) or (set3) the white poplar haplotypes present in the Mediterranean drainage basin (upper part of the haplotype network, Fig. 2). The analyses were performed based on parameters Haplotypes and haplotype networks CpSSR scores and cpDNA sequences were combined into haplotypes (see Table S3, Supporting information). Owing to SSR mutation mechanisms and high polymorphism in trnC-petN1, homoplasy events (and, therefore, network reticulations) were considered likely in our data. Homoplasy events can introduce severe biases in some analyses, such as those based on coalescence (Provan et al. 2001). Therefore, three haplotype sets were defined: (i) based on the whole set of polymorphisms (‘full data set’, haplotypes coded with a ‘h’ prefix; see Table S3, Supporting information); (ii) based on unique event polymorphisms, that is excluding any polymorphisms that produced reticulations in the haplotype network (‘UEP data set’, haplotypes coded with a ‘H’ prefix; Table S3, Supporting information); and (iii) based on single nucleotide polymorphisms (‘SNP data set’, haplotypes coded with a ‘HR’ prefix; Table S3, Supporting information). Haplotype frequencies per population (see Table S4, Supporting information) were plotted using pie charts on a GIS (ARCMAP version 9.2; ESRI, Redlands, CA, USA) that also included main physiographical features. Statistical parsimony networks were constructed with TCS vs. 1.21 (Clement et al. 2000) based on genetic distance matrices among haplotypes for the three haplotype sets. Because locus-specific mutation rates were not available, distances among different variants at each polymorphic site were considered equal. Likewise, all 2012 Blackwell Publishing Ltd Fig. 2 Statistical parsimony network representing the minimum number of polymorphic site differences among haplotypes. The network was constructed considering only unique event polymorphisms (UEPs). The inset represents a network using only single nucleotide polymorphisms (SNPs). Notice the unexpected location of H08, H09 and H14. 3598 D . M A C A Y A - S A N Z E T A L . scaled by population size N (which have higher precision; Wilson et al. 2003). A uniform prior distribution was assumed for h (h = 2 Nl, l being the mutation rate), and an estimation of TMRCA, T, was obtained in coalescence units, that is, scaled by N. For each data set, three independent MCMC were run with a burn-in period of 10 000 iterations and a main run of 20 000 to obtain a total of 60 000 iterations for the estimation of T. To obtain unscaled coalescence times in years, we assumed that Iberian and African white poplar lineages diverged after the flooding of the Strait of Gibraltar (some 5.33 Ma; Duggen et al. 2003). This assumption seems reasonable as post flooding seed dispersal across the Strait of Gibraltar has been suggested to be very rare (if any) for other forest trees with large dispersal capabilities (see Jaramillo-Correa et al. 2010). The barrier to seed dispersal results from distance across the strait (14.3 km at its narrowest point) and from accelerated wind speed through the strait due to funnelling by the steep-sided land masses on both sides (wind normally blows from the East, that is, ‘Levanter’, during poplar dispersal season; Dorman et al. 1995). Considering that the product T · h is proportional to the unscaled TMRCA (t) in years (i.e. T · h = 2l · t · g, where g is the generation time), rough estimates of t were obtained for set2 (the spread of Iberian haplotypes) and set3 (the divergence across Mediterranean lineages) by computing T · h proportions with respect to set1 [i.e. (T · h)set2 ⁄ (T · h)set1 = tset2 ⁄ tset1 where t for set1, tset1, is fixed at 5.33 Ma] under the assumption of constant mutation rates across runs (a reasonable hypothesis within species). Finally, effective population sizes for each set run were computed as N = t ⁄ (T · g) considering a generation time, g, of 20–60 years. These values are slightly higher than those found in the literature (e.g. 15 years for P. tremula in Ingvarsson 2008) because in white poplar generation times are probably extended due to extensive clonal propagation. Genetic diversity and differentiation across drainage and river basins Nei’s (1978) unbiased haplotypic within-population genetic diversity (expected heterozygosity, HE) was calculated based on the full data set (cpSSRs and cpDNA sequences) for all poplar species and populations using Arlequin vs. 3.1 (Excoffier et al. 2005). HE was also computed by pooling individuals according to drainage basin (Atlantic vs. Mediterranean), river basin or latitude (North vs. South). Non-parametric KolmogorovSmirnov tests were used to compare levels of genetic diversity at each spatial scale. Computation of haplotypic richness, before and after rarefaction, and the number of private haplotypes after rarefaction was carried out using RAREFAC (Petit et al. 1998). Estimates of genetic differentiation among populations and regions for white and black poplars were obtained based on the full and the UEP data sets. Hierarchical Analyses of Molecular Variance (AMOVAS) were performed with Arlequin vs. 3.1 grouping populations by drainage basins, river basins or latitude. AMOVAs for the UEP data set considered haplotype frequencies (FST) or distances among haplotypes (NST). Global F-statistics were computed using SPAGEDI vs. 1.3d software (Hardy & Vekemans 2002). Because F-statistics have recently been shown to perform badly when levels of diversity are dissimilar among populations, we also computed the D estimator by Jost (2008) using DEMEtics vs. 0.8.0 in R environment (Gerlach et al. 2010). Significance of D estimates was evaluated using bootstrap resampling. In the extensively sampled white poplar, isolation by distance (IBD) analysis was used to test the effect of water-dependency on genetic structure. In the absence of barriers to gene flow, the ratio FST ⁄ (1 ) FST) is expected to increase linearly as a function of the distance between pairs of populations (or its logarithm in two-dimensional analyses). Hence, a single linear regression slope is expected for all population pairs. If water-dependency influences the genetic structure, smaller differentiation is expected for pairs of populations belonging to the same river basin than for pairs belonging to different basins. Consequently, a flatter regression slope is expected within and a steeper slope between basins (see box 2 of Guillot et al. 2009). IBD was assessed using only the full data set in one (within main river basins) or two (within and between basins) dimensions. In one dimension, both ‘resistance’ distances (the distance measured following the river course, see Guillot et al. 2009) or Euclidean distances were used; otherwise Euclidean distances were applied. The significance of IBD (regression slope greater than zero) was tested with permutations. Statistical analyses were carried out with SPAGeDi vs. 1.3d (FST estimates) and Statistica vs. 10 (significance of regression slopes). Genetic structure at the local scale Five highly-polymorphic nSSRs were used to identify clonal assemblies and to evaluate fine-scale SGS in one grey and four white poplar populations (see Fig. 1). Clone identity in the Aranda de Duero population (where unusually large clonal assemblies were found) was confirmed using 14 additional nSSRs. Individuals with the same genotype (ramets) were identified using Gimlet vs. 1.3.2 (Valière 2002). All individuals (i.e. genets and ramets) were used together to produce an 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3599 overall estimate of SGS. The relative kinship coefficient Fij of Loiselle et al. (1995) was computed for all pairs of individuals within populations using SPAGeDi, and considering the whole sample (n = 201) allele frequencies as reference. Fij was regressed on the Euclidean distance between individuals (linear environment), and deviation from zero (presence of SGS) of the regression slope b was tested with permutations. Values of b were used to compare SGS strength across populations. The SGS patterns were plotted averaging Fij in five distance classes (0–25, 25–50, 50–100, 100–200 and >200 m) including a similar number of sample pairs. The average kinship in distance classes over 100 m (the distance among trees chosen for wide-range population sampling) is relevant to evaluate the probability of including clonal replicates or related individuals in data analyses performed at larger spatial scales. Common gardens and genetic differentiation for quantitative traits Two of the genotyped white poplar populations from the Ebro and three from the Guadalquivir basin were included in a quantitative genetic study to determine trait differentiation among populations and river basins. The Ebro and Guadalquivir were chosen because they represent typical locations of the species and belong to different drainage basins (the Ebro drains to the Mediterranean Sea and the Guadalquivir to the Atlantic Ocean). For each population, two to four open-pollinated families (15 in total: eight from the Ebro and seven from the Guadalquivir) averaging c. 40 plants ⁄ family were established in a common garden following a complete block design with eight blocks. Total height at age 1 and 3 years (HT1 and HT3), and stem diameter at the base (DSB3) and stem form (FOR3) at age 3 years were measured in all individuals. Stem form was evaluated as a discrete variable, with values from 1 for straight stems, to 2 for arched stems without inflection, and 3 for sinuous stems with at least one inflection. Variance components for basin, population and family effects were obtained by Restricted Maximum Likelihood (REML) using the following model: yijklm ¼ l þ rm þ pðrÞlðmÞ þ fðpÞkðlÞ þ bj þ ðf bÞkj þ eijklm where yijklm is the phenotypic value of the variable for the ith tree from the kth family in the lth population in the mth river basin located in the jth block, l is the overall mean, rm is the effect of the mth river-basin, p(r)l(m) is the effect of the lth population within the mth river-basin, f(p)k(l) is the effect of the kth family within the lth population, bj is the effect of the jth block, 2012 Blackwell Publishing Ltd (f · b)kj is the family per block interaction and eijklm the residual. Overall genetic differentiation in quantitative traits (QST) and quantitative genetic differentiation at different hierarchical levels (among populations within river basins and among river basins) were estimated from variance components for the four traits. To disentangle the effects of genetic drift from those of adaptive divergence within and among river basins, QST estimates for each trait were compared with FST computed based on the same populations using the allozyme data set (7 loci, average of 60 diploid individuals per populations) in Alba (2000). Allozymes were preferred to nSSRs as their lower polymorphism makes them more suitable for unbiased FST estimation (Jost 2008). The bootstrap procedure outlined in Whitlock (2008) was used to test for QST > FST, generating 1000 bootstraps for each statistic (over individuals for traits and over loci for allozymes). In addition, 95% confidence intervals of the bootstrap distributions of QST and FST were compared. Results Haplotypes and haplotype networks The chloroplast loci showed a total of 36 polymorphic sites, including three mononucleotide microsatellites (one within the trnC-petN1 region), 13 SNPs, 17 indels and three short tandem repeats (STRs). Combining all polymorphisms, 54 haplotypes were resolved (Table S3, Supporting information): 36 for Populus alba (one shared with Populus x canescens), 2 exclusive for P. x canescens and 16 for Populus nigra. The highly polymorphic trnCpetN1 region alone resolved 26 haplotypes, while the less variable rpl16-poprpl resolved only ten. Considering only UEPs, the number of haplotypes was reduced to 14 (Table S3, Supporting information). In terms of the full and UEP data sets, there were no shared haplotypes across species (Figs 1 and 2). However, cpSSRs alone were unable to distinguish among species, with P. nigra and P. alba sharing seven haplotypes. This fact highlights the limited value of cpSSRs for phylogenetic inference in poplars. Only three haplotypes had a wide distribution, h23, h37, h53, at frequencies of 0.054, 0.046 and 0.032, respectively. Interestingly, the most abundant haplotypes had a restricted distribution (h24, h25 and h41, at frequencies of 0.069, 0.107 and 0.058). Several haplotypes were confined to a single population (over 40% of private haplotypes for white and black poplars) or river basin (over 70% of private haplotypes for white and grey poplars). Haplotype networks based on the full data set showed a great number of reticulations and could not be resolved. However, the reduced data sets based on 3600 D . M A C A Y A - S A N Z E T A L . UEPs or SNPs yielded interpretable networks (Fig. 2). Many frequent white poplar haplotypes (H02-H05) were closely related to the widespread H01 that is found in both Atlantic and Mediterranean drainage basins. Some less abundant P. alba haplotypes (H06 and H07) lay close to P. nigra haplotypes and were exclusive of the southern part of the Atlantic drainage basin, probably indicating an older presence of white poplar in this range. Finally, we have to note one very divergent P. nigra haplotype (H14) and the wide separation (with intermingling P. nigra haplotypes) of (i) Iberian and North African (H08) P. alba; and (ii) P. alba and the closely related P. tremula (both in section Populus). Time to the most recent common ancestor 0.8 The scaled TMRCA estimates from Bayesian inference (T, in coalescent units) in white poplar followed unimodal, asymmetric (gamma-like) distributions with averages of 3.31 for set1 (see Material and methods), of 2.19 for set2 and 1.91 for set3. T · h distributions (Fig. 3) were significantly different for the three data sets (Kolmogorov-Smirnov tests, P < 0.01) indicating divergence of Moroccan and Iberian white poplar lineages 1.93 times (CI: 0.74–4.13 at 95%) before the spread of Iberian lineages, and 4.32 times (CI: 1.71–8.81 at 95%) before the lineage radiation of the Mediterranean drainage basin in the IP. Considering the time of the last known direct communication between the IP and northern Africa (the Messinian Salinity Crisis that finished about CpDNA diversity at different geographic scales Overall, based on the full data set, haplotypic diversity per population was lower in P. alba (average of 0.317) than in P. nigra (average of 0.409) (Table S1, Supporting information). Sympatric populations often showed contrasting diversity levels for each species (e.g. Henares population, P. alba: 0.000, P. nigra: 0.600; see also Fig. S1, Supporting information) reflecting their different demographic history. Haplotypic diversity was extremely variable at the population level for both white and black poplars, and there were no significant differences between populations belonging to different drainage basins or latitudinal ranges (KolmogorovSmirnov tests, P > 0.1). By contrast, genetic diversity at the regional level (sensu Hampe & Petit 2005; see Material and methods) was much greater for southern Iberian river basins than for northern basins in white poplar (Table 1). Similarly, after adjustment for uneven sample size via rarefaction (Petit et al. 1998), 26 haplotypes (21 private) were found in southern Iberian river basins compared to ten (five private) in the north. Finally, regional genetic diversity was similar for Mediterranean and Atlantic drainage basins (see Table 1), thus indicating a more important role of latitude than drainage basin for explaining the current standing genetic variation of Iberian white poplar. As black poplar is scarce in southern Iberia, comparative data are not available for this species. 0.4 CpDNA differentiation across drainage and river basins 0.0 0.2 Density 0.6 Mediterranean haplotypes Iberian haplotypes All haplotypes 5.33 Ma; Duggen et al. 2003), the spread of Iberian haplotypes was dated to c. 2.76 Ma, and the divergence across Mediterranean lineages to 1.23 Ma. If we assume a generation time, g, for P. alba of 20–60 years, effective population sizes (N) are estimated to 21 000–63 000 for Iberian white poplar (10 667–32 000 for the Mediterranean range). 0 5 10 15 T×θ Fig. 3 Density plots of unscaled Time to the Most Recent Common Ancestor, TMRCA, obtained by coalescence simulations using BATWING (see text for details). Three different sets of runs are shown including (from right to left): (i) all white poplar haplotypes; (ii) only Iberian haplotypes; and (iii) haplotypes that are present in Mediterranean populations. In white poplar, genetic differentiation as estimated by FST and Jost’s D was significant for almost all spatial scales (populations, river basins, latitudinal groups and drainage basins; see Table 2 and below for exceptions), with overall values of FST = 0.670 (0.735 for the UEP data set) and D = 0.929 (0.559 for the UEP data set, Table 2). The main factors causing genetic structure (based on the more reliable UEP data set) in this species were river and drainage basins with FCT ⁄ NCT ⁄ D values of 0.320 ⁄ 0.223 ⁄ 0.511 and 0.374 ⁄ 0.260 ⁄ 0.569, respectively. In addition, judging by Jost’s D, the river basins with the lowest (and non-significant) levels of internal differentiation were the northern Duero and Catalonia. 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3601 Table 1 Haplotypic genetic diversity and allelic richness in white and black poplars from the Iberian Peninsula; number of sampled individuals (n), number of haplotypes (A), number of haplotypes after rarefaction (A¢), number of private haplotypes after rarefaction (Ap¢) and Nei’s expected heterozygosity HE (standard deviation). The minimum sample size in each category was used as reference for rarefaction Species ⁄ group White poplar River basin Duero Catalonia Ebro Levante Tajo Guadiana Guadalquivir Latitude Northern Southern Drainage basin Atlantic Mediterranean Overall Black poplar* River basin Duero Catalonia Ebro Levante Tajo Drainage basin Atlantic Mediterranean Overall A¢ Ap ¢ n A 64 28 86 91 30 10 127 4 2 7 15 5 2 11 3.00 1.91 4.40 6.63 4.35 2.00 5.66 1.28 0.00 1.12 3.89 3.35 0.00 4.00 178 258 10 28 10.00 26.49 5.00 21.80 0.797 (0.018) 0.940 (0.005) 231 205 436 19 19 33 18.95 19.00 26.64 13.97 14.00 11.65 0.904 (0.008) 0.863 (0.016) 0.939 (0.004) 10 24 54 6 30 3 5 9 2 5 2.60 3.21 4.03 2.00 3.14 0.60 1.04 2.46 0.00 0.50 0.644 0.721 0.831 0.333 0.674 40 84 124 7 13 16 7.00 11.09 16.00 3.00 7.79 7.00 0.767 (0.046) 0.892 (0.012) 0.908 (0.009) HE (SD) 0.543 0.304 0.733 0.899 0.779 0.356 0.847 (0.059) (0.094) (0.037) (0.012) (0.040) (0.159) (0.016) (0.101) (0.058) (0.024) (0.215) (0.076) *Black poplar was not sampled in the southern Iberian Peninsula as it is relatively scarce in that region. Latitudinal differentiation was not significant for FST or NST and very low for D. Finally, looking at the full data set, only five haplotypes (out of a total of 36) were shared across drainage basins and numbers of private haplotypes were very similar in each region (16 in the Atlantic vs. 14 in the Mediterranean). In black poplar, patterns of genetic differentiation were less clear, probably due to reduced sampling (only northern Iberian populations) and higher human-mediated transfer of seeds and cuttings among populations (Galán-Cela et al. 2003). Despite overall genetic differentiation similar to white poplar (FST = 0.627 and D = 0.600 with the UEP data set), black poplar showed lower (and non-significant for FST or NST) genetic differentiation across drainage basins (0.130 ⁄ 0.032 ⁄ 0.432 for FCT ⁄ NCT ⁄ D) and inconsistent values for differentiation across river basins (low and non-significant for F- and N-statistics but relatively high for Jost’s D; Table 2). 2012 Blackwell Publishing Ltd Isolation by distance (i.e. positive slopes) was found in white poplar, albeit with different strengths at different spatial scales. Regression slopes showed stronger (and significant) IBD among river basins than within them, highlighting the isolation effect produced by the dependence of white poplar on water courses (Table 3; see also Fig. S2, Supporting information). Within-basin IBD was found in the Duero when regressing on the logarithm of Euclidean distance and (marginally, 0.05 < P < 0.10) in the Guadalquivir regressing on resistance distance. No IBD was found in the Ebro basin. Levels of clonality and genetic structure at the local scale In white poplar, the five highly polymorphic nSSRs identified 6–13 genets per population, with an average 4.2 ramets per genet (n ⁄ NG, Table 4). Clone size in this species was highly variable (from a few metres to several kilometres). Three out of four white poplar populations had average clone sizes below 100 m. Larger clonal assemblies, with one of them extending over 15 km (Fig. 4), were identified and confirmed with 14 additional SSRs in Aranda de Duero from the northern Duero basin (a tundra-like area during glacial times). This population also contained a higher number of genets (13) than other populations (6–9), resulting in similar levels of (significant) overall fine-scale genetic structure (Table 4). All populations (including Aranda de Duero) had lower and non-significant kinship at >100 m distance classes (Table 4), suggesting that samples of individuals separated by >100 m for cpDNA analysis consisted largely of unrelated individuals. The grey poplar population was characterized by just four genets with wide-ranging distances (up to 24 km) among ramets. This surprising clonal structure could be a consequence of propagation by farmers, as grey poplar is the only source of softwood in the region, and occurs mostly in managed environments (e.g. abandoned watermills, farms, etc.). Genetic differentiation for quantitative traits in white poplar Two of the four quantitative traits showed significant overall genetic differentiation: HT3 with QST = 0.569 ± 0.149 (SD) and FOR3 with QST = 0.696 ± 0.114 (SD). Both traits had over three to sixfold higher differentiation among basins (0.435 ± 0.195 and 0.592 ± 0.120, respectively) than within them (0.135 ± 0.046 and 0.104 ± 0.102, respectively). Populations from the Ebro basin had generally taller and straighter individuals. HT1 and DSB3 were not significantly different among river basins. Given that secondary growth in trees is 3602 D . M A C A Y A - S A N Z E T A L . Table 2 Genetic differentiation among populations ⁄ groups at various hierarchical levels in white and black poplars from the Iberian Peninsula. Differentiation was measured considering haplotypic frequencies (F-statistics and Jost’s D-statistics) or taking into account genetic distances among haplotypes (N-statistics). Estimates are provided for haplotypes resolved using the complete data set or, alternatively, using only UEPs (see text for details). All genetic differentiation estimates are significant at a = 0.05 unless stated otherwise (ns). NA: not available or not possible to compute White poplar Black poplar Full set UEPs Full set UEPs Group Level* F-statistics F-statistics N-statistics F-statistics F-statistics N-statistics River basin FCT FSC FST FCT FSC FST FCT FSC FST 0.165 0.616 0.679 0.090 0.653 0.685 0.082 0.655 0.683 0.670 0.320 0.632 0.750 )0.001ns 0.735 0.734 0.374 0.654 0.783 0.735 0.223 0.586 0.678 0.028ns 0.660 0.669 0.260 0.605 0.707 0.665 0.048ns 0.524 0.547 NA NA NA 0.043ns 0.532 0.552 0.542 0.170ns 0.568 0.642 NA NA NA 0.130ns 0.600 0.652 0.627 0.043ns 0.312 0.341 NA NA NA 0.032ns 0.324 0.346 0.331 0.773 NA 1.000 0.794 NA 0.661 NA NA NA NA 0.744 0.737 0.892 0.889 0.586 NA 0.000ns 0.465 NA 0.723 NA NA NA NA 0.432 0.695 0.501 0.600 Latitude Drainage basin Overall Jost’s D-statistics River basin Duero Catalonia Ebro Levante Tajo Guadalquivir Latitude North South Drainage basin Atlantic Mediterranean Overall Among Within Within Within Within Within Within Among Within Within Among Within Within 0.892 0.353 0.076ns 0.764 0.895 0.985 0.807 0.896 0.759 0.935 0.926 0.888 0.846 0.929 0.511 0.008ns 0.000ns 0.544 0.678 0.031ns 0.268 0.052 0.534 0.576 0.569 0.227 0.549 0.559 *FCT refers to genetic differentiation among groups (i.e. river basins, latitudes or drainage basins), FSC to genetic differentiation among populations within groups and FST to genetic differentiation among populations without considering groups. less important for early establishment than height differences, significant genetic differentiation for stem diameter may become apparent in later common garden assessments as trees mature. Interestingly, when compared to neutral markers for the same populations using Whitlock’s (2008) approach (Table S5, Supporting information), QST for FOR3 was significantly higher than FST among river basins (P-value: 0.033, with FST and QST 95% confidence intervals of 0.011–0.238 and 0.337–0.774, respectively), but not among populations within them (P-value: 0.781, with FST and QST 95% confidence intervals of 0.086–0.247 and 0.000–0.303, respectively). For HT3, a similar trend was observed (P-value for QST > FST among river basins: 0.151, P-value for QST > FST within river basins: 0.715), but a high QST variance among river basins for this trait (QST 95% confidence intervals of 0.065–0.733) rendered the comparison not significant. Discussion Haplotype networks and shared polymorphism across species The paradoxical position of most black poplar haplotypes within the white poplar network and closeness to aspen is consistent with previous hypotheses of ancient hybridization followed by capture of Populus alba’s chloroplast by Populus nigra (Smith & Sytsma 1990). Hamzeh & Dayanandan (2004) observed a cyto-nuclear incongruence for the phylogenetic position of black poplar. They placed this species in the section Populus 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3603 hybridization events in shared glacial refugia and postglacial recolonization (e.g. Petit et al. 2002) and is maintained by recurrent interspecific gene flow (Lexer et al. 2006). In contrast, in species that do not currently hybridize but that may have hybridized in the past, relatively recent reproductive isolation would result in a progressive loss of shared haplotypes while retaining close phylogenetic relationships. Our results in P. alba and P. nigra are in agreement with the second explanation. The highly divergent P. nigra haplotype H14 (differing by 12 mutations from the closest haplotype in the network) could then be more closely related to the genuine, pre-introgressed P. nigra plastid genome. Alternatively, haplotype introgression from commercial Euroamerican clones (Vanden Broeck et al. 2006) or from the ornamental Lombardy cultivar (Chenault et al. 2011) has been shown for P. nigra. Sequencing of a diverse array of commercial clones (n = 14) found H14 among them (not shown), pointing to introgression, despite being generally rare (<5%; Heinze & Lickl 2002 and references therein; but see Ziegenhagen et al. 2008 and Smulders et al. 2008a that reported c. 20–50% introgressed offspring in Elbe and Rhine rivers, respectively), as the most-likely explanation. As DNA sequences revealed a complete segregation among species, shared variants at ccmp2, ccmp5 and the trnC-petN1 microsatellite are probably due to homoplasy rather than ancient hybridization. Microsatellites usually have higher mutation rates than other regions of the genome (reflected by a higher number of variants Table 3 Isolation by distance (IBD) within and among river basins of white poplar from the Iberian Peninsula (see also Fig. S2, Supporting information). Pairwise genetic distances expressed as FST ⁄ (1 ) FST) were regressed on the logarithm of the Euclidean distance. For main river basins, regression slopes with ‘resistance’ distances (i.e. geographic distances following the river course) are also shown; *0.05 < P < 0.10, **0.01 < P < 0.05, ***P < 0.01. NA: not available or not possible to compute Regression slopes Different basins Same basin Duero Ebro Guadalquivir Overall log (Euclidean distance) Resistance distance 0.86** 0.19 0.93** )1.24 0.26 0.79*** NA NA 0.3E-05 1.8E-05 0.3E-05* NA close to P. alba and P. tremula on the basis of cpDNA, but in the section Aigeiros on nuclear DNA evidence. Its inclusion in Aigeiros is in agreement with classical studies based on morphology (Eckenwalder 1996) and was also supported by nuclear AFLP markers (Cervera et al. 2005). Haplotype sharing has been widely described in sympatric, related tree species [for instance in European ashes (Heuertz et al. 2006) or in white oaks (Petit et al. 2002)]. In species that hybridize readily, such as ashes and oaks, haplotype sharing commonly occurred during Table 4 Number and size of clonal assemblies, and fine-scale spatial genetic structure for four white poplar and one grey poplar populations. All clone sizes are given as maximum among-ramet distance in metres; max (L): size of the largest clone, L: mean clone size, min (L): size of the smallest clone, n: number of samples and NG: number of genets; standard errors of the regression slope (SE) are computed by a jackknife resampling procedure. NA: not available or not possible to compute Average kinship (Fij) by distance class Population White poplar Aranda de Duero La Alfranca Jimena Villamanrique de la Condesa‡ Overall Grey poplar Montes de Toledo Slope River basin Max (L) L Min (L) n NG 0–25 Duero 558.37† 163.58† 11.40† 36 13 0.324 0.272 0.209 0.074 )0.007 )1.2E-05 3.1E-06 25.69 30.30 NA 5.77 9.90 NA 49 33 32 9 6 8 0.386 0.164 0.421 0.340 NA NA 0.038 0.209 NA )0.028 )0.078 NA 0.082 0.114 NA )3.6E-04 1.2E-04 )1.1E-03 2.8E-04 NA NA 86.74 5.77 150 36 0.386 0.245 0.115 0.038 0.092 )8.2E-06 2.9E-06 50 0.267 0.286 0.514 0.516 0.198 )1.6E-05 2.1E-06 Ebro 32.32 Guadalquivir 62.61 Guadalquivir NA 558.37 Tajo 23749.85 13641.84 211.24 4 25–50 50–100 100–200 >200 *All slopes are significantly different from zero with P < 0.001, as tested by permutation. † Excluding one very large and widespread clone (four ramets stretched over 15 km; see Results). ‡ Spatial coordinates were not available for this population. 2012 Blackwell Publishing Ltd b* SE 3604 D . M A C A Y A - S A N Z E T A L . ED50 30N X686635 Y4606230 0 50 100 200 m La Alfranca GENOTYPES 1 2 3 4 5 6 7 8 9 ED50 30N X414930 Y4617020 0 GENOTYPES 1 2 3 0 4 50 5 100 6 7 8 9 10 11 12 13 5 km Aranda de Duero 200 m Fig. 4 Spatial distribution of clonal replicates from two contrasting white poplar populations. The La Alfranca population (top) had smaller and less spread clonal assemblies than the Aranda de Duero population (bottom) that includes a clone stretching over c. 15 km. in this study), making them more prone to homoplasy (Provan et al. 2001). This condition makes them useful for local and contemporary studies, especially those where high levels of variability are desirable (e.g. parentage analysis), but discourages their use for phylogeographical inference in poplars. The noteworthy lack of shared, or even closelyrelated, haplotypes between P. alba and P. tremula in the IP contrasts with expectation, as these species often hybridize and they are largely sympatric in this range. However, a larger sampling of P. tremula should be carried out to confirm this observation. Finally, the pronounced divergence of the Moroccan endemic H08 from Iberian haplotypes indicates an ancient divergence of North African and Iberian lineages (see below), as previously noticed by Fussi et al. (2010) based on a limited sample of PCR-RFLP haplotypes. Different species, different histories Black poplar, while showing a similar degree of overall genetic differentiation as white poplar, displayed higher regional haplotypic genetic diversity and lower levels of population genetic structure among river and drainage basins. These results indicate that the two species have different demographic histories. In particular, they point to more frequent gene exchange across river and drainage basins, and ⁄ or generally higher effective populations sizes in black poplar, which is to be expected in view of its higher tolerance to cold temperatures (Galán-Cela et al. 2003; Ruiz de la Torre 2006) and previous literature (e.g. Smulders et al. 2008b). Alternatively, the lower genetic structure in black poplar could reflect a higher seed and cutting transfer by humans across regions. However, the high number of private 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3605 haplotypes found in the IP in this (c. 40%) and in other studies (e.g. Cottrell et al. 2005) does not support this alternative hypothesis. Similar patterns of high diversity and low differentiation have been observed in the more cold-tolerant species of other European tree genera, such as the six native Iberian pine species (Soto et al. 2010), or in the thermophilous Fraxinus angustifolia and the more coldtolerant F. excelsior (Heuertz et al. 2006). The bases of these patterns are likely to be better survival of the cold-tolerant species during the cold stages of past glaciations and early colonization of new territory, compared to thermophilous tree species. Our findings are relevant because they extend these observations to riparian trees that are normally not considered to be dependent on regional climatic patterns and, thus, are excluded from models of future species distributions based on climate predictions (e.g. Benito-Garzón et al. 2008 for Iberian trees). Moreover, our findings suggest that near-future predicted climatic change may affect Iberian poplar species differentially, giving a competitive advantage to the more drought- and salt-tolerant white poplar compared to black poplar. Competitive exclusion from the already scarce riparian habitat would possibly drive this already threatened species (Lefèvre et al. 1998 and references therein) to lower effective population numbers and, eventually, to local extinctions. Genetic signatures of ancient events in white poplar The genus Populus appeared during the transition from the Secondary to the Tertiary era and diversified into different sections and species during the warm Paleogene period (Eckenwalder 1996). Modern species are believed to have evolved during the global cooling in the beginning of the Neogene (c. 23 Ma). During this period, still warmer and wetter than today, and before the beginning of the Quaternary, modern poplars would have spread across the IP and northern Africa. The North African and Iberian lineages of white poplar would have diverged after their last possible contact at the end of the Messinian Salinity Crisis c. 5.33 Ma, when the Mediterranean Sea was desiccated (Krijgsman et al. 1999; Duggen et al. 2003). The subsequent flooding of the Mediterranean basin has been associated to a genetic discontinuity at the level of the Strait of Gibraltar in various organisms (Rodrı́guez-Sánchez et al. 2008; see Jaramillo-Correa et al. 2010 for some forest trees). Within the IP, a marked differentiation between the Atlantic and Mediterranean drainage basins was found in white poplar. This pattern has also been found in other Iberian tree species (Rodrı́guez-Sánchez et al. 2010). This disjunction probably reflects a genetic 2012 Blackwell Publishing Ltd signature of ancient geological events, considering that the main Iberian mountain systems attained their current configuration during the late Miocene. The fact that F-statistics using the UEP data indicated stronger differentiation (0.374 vs. 0.082) than using the complete polymorphism set (assumed to be affected by recent mutation) also pointed at ancient phylogeographic processes. Consistent with its lower sensitivity to mutation, Jost’s D statistic did not reflect these differences. The reasons for the significant differentiation between drainage basins (Atlantic vs. Mediterranean) are not obvious, considering that major Iberian mountain systems run from west to east, thus mostly preventing latitudinal migration (i.e. among river basins but not between drainage areas). One explanation that can apply to plants, and more specifically to riparian trees, is related to the vegetation altitudinal shifts produced by glacial climatic oscillations (Hewitt 1996; Rodrı́guezSánchez et al. 2010). The relatively benign climate before the Pleistocene should have favoured extensive gene exchange across drainage basins, even for lowland species. Then, during the Pleistocene cold periods, altitudinal limits for plant species lowered and riparian trees probably became confined to the lower river courses. This process resulted in distributions close to the western and eastern coastal fringes of the IP where the main Iberian rivers flow into the sea. In this way, the Atlantic and Mediterranean drainage basins would have become effectively separated while migration along the coastlines (where mountain ranges are lower) would have connected river basins. Our results in white poplar suggest that increasing isolation between Atlantic and Mediterranean drainage basins occurred c. 1.12 Ma (lower bound of 3.11 Ma), in agreement with the proposed scenario related to Pleistocene cooling. Regional and population effects of glacial times in white poplar The patterns of genetic diversity and structure in white poplar reflect the effects of Pleistocene climatic oscillations in several ways. First, regional genetic diversity was higher and private haplotypes were four times more abundant in the southern Iberian river basins, which were warmer than the northern basins. Secondly, genetic structure among populations was much more pronounced in the southern Guadalquivir and Levante basins than in the northern Duero and Catalonia basins. Thirdly, in the formerly tundra-like Duero basin, significant IBD was found only when considering Euclidean geographical distances, but not ‘resistance’ distances. This IBD pattern is more consistent with a rapid isotropic postglacial spread than with a long-term build-up of SGS along linear favourable environments. Fourthly, 3606 D . M A C A Y A - S A N Z E T A L . clonal assemblies were apparently larger (with some clones extending up to c. 15 km) in the colder Duero basin than in the southern basins. Asexual propagation could have helped Iberian poplars to survive in harsh glacial environments and to colonize new territory rapidly once ecological conditions improved (Silvertown 2008). Pleistocene glacial oscillations lowered temperature and humidity globally. Palaeoecological inferences indicate that during the glacial maxima, areas in the westernand northernmost IP (like the Duero basin) were barely habitable by arboreal vegetation (González-Sampériz et al. 2010). Northern populations of the thermophilous white poplar show a genetic depauperation that seems to reflect these past events. The hostile climatic conditions suffered during Pleistocene glacial periods in these areas could have pushed white poplar populations towards one of two fates: (i) an important population size reduction, but persistence in sheltered ‘cryptic refugia’ (sensu Stewart & Lister 2001); or (ii) local extinction followed by postglacial recolonization. The first scenario would have resulted in reduced genetic diversity but would have maintained common local haplotypes in surviving populations (Provan & Bennett 2008). In the second situation, diversity would have been reduced due to founder effects, and the region would have been replenished with (non-local) haplotypes from the colonizing populations. Our data support the first scenario, showing significant genetic differentiation among northern and southern river basins and presence of private haplotypes in both latitudes. The existence of cryptic local refugia and recent spread of surviving genotypes is also a plausible explanation for the high haplotypic diversity observed in white poplar populations from Austria (Fussi et al. 2010) and the discovery of huge clonal assemblies of the species in Sardinia and Malta (Brundu et al. 2008; Fussi et al. 2012). Signals of glacial survivorship in scattered populations situated beyond the estimated persistence limit have been widely observed in boreal and alpine latitudes (Hewitt 2004; Opgenoorth et al. 2010), including for some Salicaceae (e.g. Palmé et al. 2003 for Salix sp.). Evidence for local adaptation in white poplar Several decades of common garden experiments have revealed the widespread occurrence of locally adapted populations in forest trees (see reviews in McKay & Latta 2002; Latta 2004; Savolainen et al. 2007), including some Populus species (see Fig. 4 in Savolainen et al. 2007 for Populus balsamifera and Populus tremuloides). The higher quantitative (QST) than molecular (FST) genetic differentiation found across river basins, albeit not within river basins, for some growth traits in white poplar suggests that this species is also locally adapted, but at wider spa- tial scales (i.e. river basins that can span hundreds of kilometres) than in other temperate trees. White poplar populations have typically low population sizes, given their dependence on phreatic water and the high anthropization of Iberian riparian habitats, which breaks the continuity of riparian forests. Human impact is most noteworthy in the low- and medium water courses where white poplar is more abundant (Ruiz de la Torre 2006). Our findings suggest the role of gene flow over mesoscale distances, replenishing genetic variation and counteracting local genetic drift. Lack of IBD patterns also suggests frequent gene exchange along river courses within basins, at least for those Iberian rivers where glacial impact was low (see above). In this scenario, the relative homogeneity of riparian habitats would have counteracted the arrival of maladapted genotypes, preventing the development of ‘migration meltdown’ processes (i.e. self-reinforced processes in which immigration of maladapted genotypes decreases local density, which in turn increases immigration rates bringing in more maladapted genotypes; Lenormand 2002), which can eventually result in population extinction. Theoretical models have shown the potential beneficial effects of gene flow for small peripheral populations (Alleaume-Benharira et al. 2006), and experimental evidence is accumulating (e.g. Sexton et al. 2011). The existence of local adaptation and specialized phenotypes has direct consequences for the adaptive response of white poplar to future environments, such as those predicted by the Intergovernmental Panel on Climatic Change (IPCC, http://www.ipcc.ch, accessed on May, 2012). Indeed, past adaptation processes have most likely generated a wide array of standing genetic variation that may prove of utility beyond the current range of the species. This expectation highlights the need for a wider exploration of genetic resources in this species as well as for the establishment of large multisite common gardens. Conclusions Past climate conditions have left genetic signatures in riparian tree species, such as the Iberian poplars. Some of these signatures reflect early Pleistocene events that led to differentiation of gene pools in the Mediterranean and Atlantic drainage basins. Genetic diversity is higher and genetic differentiation lower in cold-tolerant black poplar than in thermophilous white poplar, and we speculate that cold-tolerance resulted in better survival and higher gene exchange across geographical barriers of this species during past glaciations, as shown elsewhere for other cold-tolerant trees. Patterns of IBD in white poplar reflect its dependence on phreatic water, resulting in higher IBD among river basins than within. 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3607 At the local scale, SGS is greatly influenced by the widespread existence of clonal assemblies extending, in a few cases, up to several kilometres. Nevertheless, the presence of numerous genets of small clone size points out to asexual propagation as a means for maintaining genetic diversity under harsh environments (rather than reducing the effective population size) and for colonizing new territory rapidly. Asexual propagation did not seem to prevent local adaptation in white poplar. Gene flow at mesoscale distances seems sufficient to counteract genetic drift and to promote local adaptation at the river-basin scale. Riparian trees occupy very specialized niches surrounded by large inhospitable areas. The existence of locally adapted phenotypes, even in highly structured species such as the Iberian poplars, is remarkable and suggests some resilience of poplar populations to environmental change and a capacity to adapt when confronted with new environments. Acknowledgements We are indebted to José Antonio Mancha, Santiago de Blas and Fernando del Caño for help in sample collection, and Carmen Garcı́a Barriga and Marı́a Teresa Cervera for lab assistance and advice. We are also grateful to P. C. Grant for reviewing the English language. This work was supported by project REPROFOR (AGL2005-07440-C02-01, Ministry of Education and Science). The contribution of the EU EVOLTREE Network of Excellence and the Collaborative Project on ‘Conservation of Forest Genetic Resources’ between the MARM and the National Institute for Agriculture and Food Research and Technology, INIA (AEG06-054) is also acknowledged. We want to specially acknowledge the contribution of Nuria Alba to this study. She contributed to the collection of the samples of Populus alba, the isozyme analysis and the quantitative evaluation. She passed away on January 2nd, 2012. References Alba N (2000) Variabilidad genética de Populus alba L. mediante caracteres isoenzimáticos y fenotı́picos. PhD Dissertation, Polytechnic University of Madrid, Madrid. Alleaume-Benharira M, Pen IR, Ronce O (2006) Geographical patterns of adaptation within a species’ range: interactions between drift and gene flow. Journal of Evolutionary Biology, 19, 203–215. Anderson LL, Hu FS, Nelson DM, Petit RJ, Paige KN (2006) Ice-age endurance: DNA evidence of a white spruce refugium in Alaska. Proceedings of the National Academy of Sciences of the United States of America, 103, 12447–12450. Benito-Garzón M, Sánchez de Dios R, Sáinz Ollero H (2008) Effects of climate change on the distribution of Iberian tree species. Applied Vegetation Science, 11, 169–178. Brundu G, Lupi R, Zapelli I et al. (2008) The origin of clonal diversity and structure of Populus alba in Sardinia: evidence from nuclear and plastid microsatellite markers. Annals of Botany, 102, 997–1006. 2012 Blackwell Publishing Ltd Bryan GJ, McNicoll J, Ramsay G, Meyer RC, De Jong WS (1999) Polymorphic simple sequence repeat markers in chloroplast genomes of Solanaceous plants. Theoretical and Applied Genetics, 99, 859–867. Cervera MT, Storme V, Soto A et al. (2005) Intraspecific and interspecific genetic and phylogenetic relationships in the genus Populus based on AFLP markers. Theoretical and Applied Genetics, 111, 1440–1456. Cheddadi R, Vendramin GG, Litt T et al. (2006) Imprints of glacial refugia in the modern genetic diversity of Pinus sylvestris. Global Ecology and Biogeography, 15, 271–282. Chenault N, Arnaud-Haond S, Juteau M et al. (2011) SSRbased analysis of clonality, spatial genetic structure and introgression from the Lombardy poplar into a natural population of Populus nigra L. along the Loire River. Tree Genetics & Genomes, 7, 1249–1262. Clement M, Posada D, Crandall KA (2000) TCS: a computer program to estimate gene genealogies. Molecular Ecology, 9, 1657–1659. Cottrell JE, Krystufek V, Tabbener HE et al. (2005) Postglacial migration of Populus nigra L.: lessons learnt from chloroplast DNA. Forest Ecology and Management, 206, 71–90. Dorman CE, Beardsley RC, Limeburner R (1995) Winds in the Strait of Gibraltar. Quarterly Journal of the Royal Meteorological Society, 121, 1903–1921. Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus, 12, 13–15. Duggen S, Hoernle K, van den Bogaard P, Rüpke L, Morgan JP (2003) Deep roots of the Messinian salinity crisis. Nature, 422, 602–606. Eckenwalder JE (1996) Systematics and evolution of Populus. In: Biology of Populus and Its Implications for Management and Conservation (eds Stettler RF, Bradshaw HD, Heilman PE, Hinckley TM), pp. 7–32. NRC Research Press, Ottawa, Ontario. Excoffier L, Laval G, Schneider S (2005) Arlequin (version 3.0): an integrated software package for population genetics data analysis. Evolutionary Bioinformatics, 1, 47–50. Fussi B, Lexer C, Heinze B (2010) Phylogeography of Populus alba L. and Populus tremula L. in Central Europe: secondary contact and hybridisation during recolonisation from disconnected refugia. Tree Genetics & Genomes, 6, 439–450. Fussi B, Bonello J, Calleja E, Heinze B (2012) Combining the use of molecular techniques and archival documentary evidence to trace the origin of Populus alba in a Central Mediterranean archipelago. European Journal of Forest Research, 131, 347–354. Galán-Cela P, Gamarra-Gamarra R, Garcı́a-Viñas JI (2003) Árboles y arbustos de la Penı́nsula Ibérica e Islas Baleares, 4th edn. Eds Jaguar, Madrid. Gerlach G, Jueterbock A, Kraemer P, Deppermann J, Harmand P (2010) Calculations of population differentiation based on GST and D: forget GST but not all of statistics!. Molecular Ecology, 19, 3845–3852. Gómez A, Lunt DH (2007) Refugia within refugia: patterns of phylogeographic concordance in the Iberian Peninsula. In: Phylogeography of Southern European Refugia: Evolutionary Perspectives on the Origins and Conservation of European Biodiversity (eds Weiss S, Ferrand N), pp. 155–188. Springer, Dordrecht, The Netherlands. 3608 D . M A C A Y A - S A N Z E T A L . González-Martı́nez SC, Dubreuil M, Riba M et al. (2010) Spatial genetic structure of Taxus baccata L. in the western Mediterranean Basin: past and present limits to gene movement over a broad geographic scale. Molecular Phylogenetics and Evolution, 55, 805–815. González-Sampériz P, Leroy SAG, Carrión JS et al. (2010) Steppes, savannahs, forests and phytodiversity reservoirs during the Pleistocene in the Iberian Peninsula. Review of Palaeobotany and Palynology, 162, 427–457. Guillot G, Leblois R, Coulon A, Frantz AC (2009) Statistical methods in spatial genetics. Molecular Ecology, 18, 4734–4756. Hampe A, Petit RJ (2005) Conserving biodiversity under climate change: the rear edge matters. Ecology Letters, 8, 461– 467. Hamzeh M, Dayanandan S (2004) Phylogeny of Populus (Salicaceae) based on nucleotide sequences of chloroplast trnT-trnF region and nuclear rDNA. American Journal of Botany, 91, 1398–1408. Hardy OJ, Vekemans X (2002) SPAGEDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes, 2, 618–620. Heinze B, Lickl E (2002) Rare, but steady, introgression in Austrian black poplar as a long-term risk? In: Genetic Diversity in River Populations of European Black Poplar (eds Van Dam B, Bordacs S), pp. 169–175. Csiszár Nyomda Ltd, Budapest. Heuertz M, Carnevale S, Fineschi S et al. (2006) Chloroplast DNA phylogeography of European ashes, Fraxinus sp. (Oleaceae): roles of hybridization and life history traits. Molecular Ecology, 15, 2131–2140. Hewitt GM (1996) Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society, 58, 247–276. Hewitt GM (2004) Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 359, 183–195. Huntley B, Birks H (1983) An Atlas of Past and Present Pollen Maps for Europe: 0–13000 Years Ago. Cambridge University Press, Cambridge. Ingvarsson PK (2008) Multilocus patterns of nucleotide polymorphism and the demographic history of Populus tremula. Genetics, 180, 329–340. Jaramillo-Correa JP, Grivet D, Terrab A et al. (2010) The Strait of Gibraltar as a major biogeographic barrier in Mediterranean conifers: a comparative phylogeographic survey. Molecular Ecology, 19, 5452–5468. Jost L (2008) GST and its relatives do not measure differentiation. Molecular Ecology, 17, 4015–4026. Krijgsman W, Hilgen FJ, Raffi I, Sierro FJ, Wilson DS (1999) Chronology, causes and progression of the Messinian salinity crisis. Nature, 400, 652–655. Latta RG (2004) Gene flow, adaptive population divergence and comparative population structure across loci. New Phytologist, 161, 51–58. Lefèvre F, Legionnet A, Vries S, Turok J (1998) Strategies for the conservation of a pioneer tree species, Populus nigra L., in Europe. Genetics Selection Evolution, 30, S181–S196. Lenormand T (2002) Gene flow and the limits to natural selection. Trends in Ecology & Evolution, 17, 183–189. Lexer C, Fay MF, Joseph JA, Nica MS, Heinze B (2005) Barrier to gene flow between two ecologically divergent Populus species, P. alba (white poplar) and P. tremula (European aspen): the role of ecology and life history in gene introgression. Molecular Ecology, 14, 1045–1057. Lexer C, Kremer A, Petit RJ (2006) Shared alleles in sympatric oaks: recurrent gene flow is a more parsimonious explanation than ancestral polymorphism. Molecular Ecology, 15, 2007–2012. Loiselle BA, Sork VL, Nason J, Graham C (1995) Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae). American Journal of Botany, 82, 1420–1425. Magri D, Fineschi S, Bellarosa R et al. (2007) The distribution of Quercus suber chloroplast haplotypes matches the palaeogeographical history of the western Mediterranean. Molecular Ecology, 16, 5259–5266. McKay JK, Latta RG (2002) Adaptive population divergence: markers, QTL and traits. Trends in Ecology & Evolution, 17, 285–291. Medail F, Diadema K (2009) Glacial refugia influence plant diversity patterns in the Mediterranean Basin. Journal of Biogeography, 36, 1333–1345. Nei M (1978) Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, 89, 583–590. Nichols RA, Hewitt GM (1994) The genetic consequences of long-distance dispersal colonization. Heredity, 72, 312–317. Nieto-Feliner G (2011) Southern European glacial refugia: a tale of tales. Taxon, 60, 365–372. Opgenoorth L, Vendramin GG, Mao K et al. (2010) Tree endurance on the Tibetan Plateau marks the world’s highest known tree line of the Last Glacial Maximum. New Phytologist, 185, 332–342. Palmé AE, Semerikov V, Lascoux M (2003) Absence of geographical structure of chloroplast DNA variation in sallow, Salix caprea L. Heredity, 91, 465–474. Petit RJ, El Mousadik A, Pons O (1998) Identifying populations for conservation on the basis of genetic markers. Conservation Biology, 12, 844–855. Petit RJ, Csaikl UM, Bordacs S et al. (2002) Chloroplast DNA variation in European white oaks – phylogeography and patterns of diversity based on data from over 2600 populations. Forest Ecology and Management, 156, 5–26. Petit RJ, Hu FS, Dick CW (2008) Forests of the past: a window to future changes. Science, 320, 1450–1452. Provan J, Bennett KD (2008) Phylogeographic insights into cryptic glacial refugia. Trends in Ecology & Evolution, 23, 564– 571. Provan J, Powell W, Hollingsworth PM (2001) Chloroplast microsatellites: new tools for studies in plant ecology and evolution. Trends in Ecology & Evolution, 16, 142–147. Rodrı́guez-Sánchez F, Perez-Barrales R, Ojeda F, Vargas P, Arroyo J (2008) The Strait of Gibraltar as a melting pot for plant biodiversity. Quaternary Science Reviews, 27, 2100– 2117. Rodrı́guez-Sánchez F, Hampe A, Jordano P, Arroyo J (2010) Past tree range dynamics in the Iberian Peninsula inferred through phylogeography and palaeodistribution modelling: a review. Review of Palaeobotany and Palynology, 162, 507–521. Roiron P, Ali AA, Guendon JL, Carcaillet C, Terral JF (2004) Evidence of the indigenous status of Populus alba L. in the western Mediterranean Basin. Comptes Rendus Biologies, 327, 125–132. 2012 Blackwell Publishing Ltd G E N E T I C S T R U C T U R E I N I B E R I A N P O P L A R S 3609 Ruiz de la Torre J (2006) Flora Mayor. Organismo Autónomo Parques Nacionales, Dirección General para la Biodiversidad, Madrid. Savolainen O, Pyhajarvi T, Knurr T (2007) Gene flow and local adaptation in trees. Annual Review of Ecology Evolution and Systematics, 38, 595–619. Sexton JP, Strauss SY, Rice KJ (2011) Gene flow increases fitness at the warm edge of a species’ range. Proceedings of the National Academy of Sciences of the United States of America, 108, 11704–11709. Silvertown J (2008) The evolutionary maintenance of sexual reproduction: evidence from the ecological distribution of asexual reproduction in clonal plants. International Journal of Plant Sciences, 169, 157–168. Slavov GT, Zhelev P (2010) Salient biological features, systematics and genetic variation of Populus. In: Genetic and Genomics of Populus (eds Jansson S, Bhalerao RP, Groover AT), pp. 15–38. Springer, Dordrecht, The Netherlands. Smith RL, Sytsma KJ (1990) Evolution of Populus nigra (sect Aigeiros) – introgressive hybridization and the chloroplast contribution of Populus alba (sect Populus). American Journal of Botany, 77, 1176–1187. Smulders MJM, Beringen R, Volosyanchuk R et al. (2008a) Natural hybridisation between Populus nigra L. and P. x canadensis Moench: hybrid offspring competes for niches along the Rhine river in the Netherlands. Tree Genetics & Genomes, 4, 663–675. Smulders MJM, Cottrell J, Lefèvre F et al. (2008b) Structure of the genetic diversity in black poplar (Populus nigra L.) populations across European river systems: consequences for conservation and nature development. Forest Ecology and Management, 255, 1388–1399. Soto A, Robledo-Arnuncio JJ, González-Martı́nez SC, Smouse PE, Alia R (2010) Climatic niche and neutral genetic diversity of the six Iberian pine species: a retrospective and prospective view. Molecular Ecology, 19, 1396–1409. Stewart JR, Lister AM (2001) Cryptic northern refugia and the origins of the modern biota. Trends in Ecology & Evolution, 16, 608–613. Stuefer JF, Erschbamer B, Huber H, Suzuki J-I (2002) Ecology and Evolutionary Biology of Clonal Plants. Kluwer Academic Pub. Group, Dordrecht, The Netherlands. Valière N (2002) GIMLET: a computer program for analysing genetic individual identification data. Molecular Ecology Notes, 2, 377–379. Vanden Broeck A, Cottrell J, Quataert P et al. (2006) Paternity analysis of Populus nigra L. offspring in a Belgian plantation of native and exotic poplars. Annals of Forest Science, 63, 783– 790. Weising K, Gardner RC (1999) A set of conserved PCR primers for the analysis of simple sequence repeat polymorphisms in chloroplast genomes of dicotyledonous angiosperms. Genome, 42, 9–19. Whitlock MC (2008) Evolutionary inference from QST. Molecular Ecology, 17, 1885–1896. Wilson IJ, Weale ME, Balding DJ (2003) Inferences from DNA data: population histories, evolutionary processes and forensic match probabilities. Journal of the Royal Statistical Society: Series A (Statistics in Society), 166, 155–188. Ziegenhagen B, Gneuss S, Rathmacher G et al. (2008) A fast and simple genetic survey reveals the spread of poplar 2012 Blackwell Publishing Ltd hybrids at a natural Elbe river site. Conservation Genetics, 9, 373–379. D.M.S. develops a PhD on genetic structure, including clonal structure, and hybridization of Iberian poplars. M.H. is interested in population and evolutionary genetics of plant species from biodiversity hotspots at temperate and tropical latitudes. U.L.H. has broad interests in phylogeography and ecological genetics of forest trees. A.I.L. and E.H’s research focuses on the study of genetic diversity and population structure, and the molecular characterization of commercial woody plants. C.M. and A.P. are interested in conservation and use of plant genetic resources, in particular those of poplars and other trees. R.A. and S.C.G.M. have broad interests in population genetics and genomics of forest trees, ecological genetics and the evolution of Mediterranean plants, using quantitative and molecular genetics approaches. Data accessibility GenBank accessions for cpDNA haplotypes: JQ782847– JQ782883 (see Table S3, Supporting information for correspondence between accession numbers and haplotypes and Table S4, Supporting information for haplotype counts per population). Chloroplast microsatellite data deposited in the Dryad repository doi:10.5061/dryad.9hd71135. Supporting information Additional supporting information may be found in the online version of this article. Table S1 Population names, location and basic description, including expected heterozygosity estimates (HE). Table S2 Annealing temperatures, number of PCR cycles and source of nuclear microsatellites. Table S3 Haplotype definition. Table S4 Haplotype counts per population. Full haplotypes are coded with ‘h’ and UEP haplotypes with ‘H’. Table S5 Test for QST > FST following Whitlock (2008) for total height (HT3) and stem form (FOR3) at age 3. Fig. S1 Distribution of haplotypic diversity in Iberian poplars calculated as Nei’s unbiased expected heterozygosity (HE) for the full data set of cpSSRs and cpDNA sequences. Fig. S2 Scatter plots of pairwise genetic distances expressed as FST ⁄ (1 ) FST) against the logarithm of the Euclidean distance within and among river basins of white poplar from the Iberian Peninsula. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

RELATED PAPERS

RELATED TOPICS

Log In

The Atlantic–Mediterranean watershed, river basins and glacial history shape the genetic structure of Iberian poplars

The Atlantic–Mediterranean watershed, river basins and glacial history shape the genetic structure of Iberian poplars

Related Papers

RELATED PAPERS

RELATED TOPICS