Eur J Forest Res (2013) 132:137–150 DOI 10.1007/s10342-012-0663-0 ORIGINAL PAPER Chloroplast DNA polymorphism of an exotic P. mugo Turra population introduced to seaside spit of Kursiu Nerija in Lithuania Darius Danusevičius • Jurata Buchovska • Vidmantas Stanys • Jurat_e Bron_e Šikšnianien_e Virgilijus Baliuckas • Gediminas Brazaitis • Received: 6 January 2012 / Revised: 22 August 2012 / Accepted: 5 October 2012 / Published online: 1 November 2012 Ó Springer-Verlag Berlin Heidelberg 2012 Abstract The aim of this study was to elucidate the introduction history of P. mugo in the unique landscape of the Lithuanian seaside spit of Kursiu Nerija by assessing its genetic structure and the genetic diversity. The individuals were sampled in 12 populations within an area of 3 km 9 50 km along the Lithuanian part of Kursiu Nerija. P. mugo was introduced over 200 years ago to prevent sand erosion by establishing a forest cover. Chloroplast DNA polymorphism of 220 individuals of P. mugo together with 18 P. sylvestris and 11 putative P. sylvestris 9 P. mugo hybrids was assessed by the aid of five microsatellite markers. The standard intra-population diversity indexes were calculated. The intra-specific variation between distinct morphotypes as well as the population differentiation within the most spread P. mugo ssp. rotundata morphotype was assessed based on the haplotype frequencies by hierarchical AMOVA, GST/RST test, UPGMA Communicated by C. Ammer. Electronic supplementary material The online version of this article (doi:10.1007/s10342-012-0663-0) contains supplementary material, which is available to authorized users. D. Danusevičius (&)
G. Brazaitis Faculty of Forestry and Ecology, Aleksandras Stulginskis University, Studentu street 11, 53361 Akademija, Kaunas, Lithuania e-mail: darius.danusevicius@asu.lt J. Buchovska
V. Baliuckas Lithuanian Research Centre for Agriculture and Forestry, Institute of Forestry, Liepu st. 1, Girionys, 53101 Kaunas, Lithuania clustering and PCA methods. The genetic diversity of P. mugo in Kursiu Nerija was high (He = 0.95; 83 different haplotypes). All except one of the P. mugo populations sampled contained a notable share of private haplotypes. AMOVA revealed high intra-specific diversity but low differentiation between the P. mugo populations. Most of the haplotypic variance was within populations. The UPGMA clustering produced groups more corresponding to the sub-species morphotypes than the geography of the populations. There was no geographical pattern of reduction in genetic diversity towards the younger plantations. A strong candidate for a species-specific DNA marker was found. After several events of introduction, the genetic diversity of P. mugo in Kursiu Nerija is very high and is structured based on the sub-species morphotypes rather than geography. The high frequency of shared and notable frequency of private haplotypes in most of the populations indicate that the major part of the P. mugo material originates from a number of geographically and genetically related sources, which more likely are introductions from abroad that the local collections. The high frequency of private haplotypes in the northernmost populations leaves a possibility for minor introductions from other genetically distinct sources. The absence of private haplotypes in one of the sampled populations indicates the use of local seed collections. The large number of shared haplotypes provides a strong evidence for a geneflow among the P. mugo taxa. Keywords Ecotype
Hybridization
Organelle DNA polymorphism
Morphotype
Neringa
Exotic species
rotundata
uncinata
SSR
Species-specific marker Introduction V. Stanys
J. B. Šikšnianien_e Lithuanian Research Centre for Agriculture and Forestry, Institute of Horticulture, Kauno st. 30, Babtai, 54333 Kaunas, Lithuania Kursiu Nerija is a narrow strip of sandy dunes stretching 97 km along the coast of Baltic Sea in western Lithuania and 123 138 Kaliningrad Region in Russia (Curonian spit, abbreviated as Nerija, then a part of the eastern Prussia; Fig. 1). Following the intensive forest cuttings in seventeenth to eighteenth centuries, entire villages were buried by moving sands in Nerija. In 1803, the eastern Prussian administration employed a Danish inspector Bjorn Sorensen to cope with the sand erosion by establishing forest cover. It required tremendous efforts to build the protective dune along whole coast line of Nerija and experimentation with grass cover as well as various tree species until first successful plantations with P. mugo were established in 1825 (Strakauskaite 2004). Regrettably, no written records remained on the introduction sources. Presumably, it could be Denmark because of the contract with the Danish inspector who already had successful experience with afforestation of the sandy dunes in western Denmark with P. mugo (Jørgensen 2006; Strakauskaite 2004). German forest managers Georg David Kuwert and Franz Eph continued the afforestation programme with P. mugo starting at the southern part of Nerija and continuing northwards (Strakauskaite 2004). The afforestation programme intensified after 1860 when villages were buried by the moving sands. The northern part of Nerija from Juodkrante to Smiltyne was afforested the latest during later part of nineteenth to beginning of twentieth century. P. mugo was usually planted on the top of the dunes, where it fits well and was naturalized by forming a local ecotype. Again, there are no records on how the afforestation was done: by collecting the seeds in newly established stands in Nerija and afforesting the new areas northwards (stepping stone model) or by direct introductions from foreign sources or both. In 1904, the large-scale afforestation programme was completed with 2/3 of dunes covered with forests, and the eroding dunes were stabilized (Strakauskaite 2004). Presently, the plantations of P. mugo amount to 2,500 ha at the Lithuanian part of Nerija. Most of the P. mugo stands reached their natural maturity (raising fire hazard), and variable options were presented regarding its future, including gradual change to P. sylvestris. The local ecotype of P. mugo, however, represents the historical heritage of potential value. The representative stands of P. mugo must be conserved in Nerija, and our study on patterns of genetic variation may provide the guidelines. The Danish history of P. mugo starts in 1768, after professor Johan Daniel Titius from University of Wittenberg suggested afforesting the sand dunes as a measure against soil erosion (Jørgensen 2006 and references therein). In Denmark, the sources the early P. mugo material were primarily the Central and Western Alps (Austria, Switzerland) and to a lesser extent eastern France (Müller 1887; Mar-Möller 1965). P. mugo Turra is endemic to the mountains of Central and southern Europe and is a highly morphologically variable group of closely related taxa with a complex taxonomy 123 Eur J Forest Res (2013) 132:137–150 (Hamernik and Musil 2007). High variation in morphology, adaptation to specific adaptive environments and high within-taxon hybridization rates complicate the taxonomic subdivision (Christensen 1987a, b; Boratynska and Boratynski 2007). Being geneticists, we considered the sensu lato taxonomic approach by Christensen (1987a, b). Based on the studies of range-wide morphology, Christensen (1987a, b) concludes that the ssp. uncinata (forming pure stands in the western range) and ssp. mugo (forming pure stands in the eastern range) form hybrid complexes of the nothotaxon notho ssp. rotundata with an intermediate morphology in the central range. Furthermore, the ssp. rotundata exhibits a clinal variation in morphology from the uncinata-like morphotypes in the west (Pyrenees) towards the ssp. mugo type in the east (Anderson and Hubricht 1938; Heiser 1949). Unlike the populations at the central range, the populations of both ssp. uncinata and ssp. mugo at the corresponding margins of the species range often form pure stands (Christensen 1987a, b). The free inter-mating between the P. mugo taxa (e.g. Heiser 1949; 73) advocates for simple and evolutionarily sound systematics by Christensen (1987a, b). Formation of the putative natural P. mugo 9 P. sylvestris hybrids was first reported based on morphology (Marcet 1967; Christensen 1987a, b; Christensen and Dar 1997; Boratyński et al. 2003; Danusevičius et al. 2012) and confirmed by DNA markers (Kormutak et al. 2008; Wachowiak and Prus-Głowacki 2008). Kormutak et al. (2008) assessed the hybridization rates in the seeds with the cpDNA markers to define the species of the male parent in the seed embryo and the species of the female parent in the gametophyte carrying only the maternal genome. The artificial hybridization studies between P. mugo and P. sylvestris proved their inter-fertility and provided useful markers for hybrid studies (Kormutak et al. 2005; Wachowiak et al. 2004). Most of the polymorphism studies within P. mugo complex were based on morphology (reviews Christensen 1987a, b; Christensen and Dar 1997; Hamernik and Musil 2007), and there are only few based on biochemical and even fewer on DNA markers (Lewandowski et al. 2000; Schmid 2000; Monteleone et al. 2006; Dzialuk et al. 2009; Heuertz et al. 2010). Heuertz et al. (2010) in a range-wide study of P. mugo genetic diversity including the ssp. uncinata, rotundata and mugo based on chloroplast DNA markers found stronger differences between the populations than between the morphologically defined taxa. Dzialuk et al. (2009) described the patterns of chloroplast DNA polymorphism of P. mugo ssp. uncinata in Pyrenees and found high levels of within-population diversity and weak population differentiation, mainly occurring at the margin of the range. The possible explanation was recent fragmentation of historically large population and extensive gene flow among the sub-species. Monteleone et al. Eur J Forest Res (2013) 132:137–150 139 Fig. 1 Location of the populations studied in Kursiu Nerija spit at the Baltic Sea coast in the western part of Lithuania (left). Length of the Lithuanian part of the spit is ca. 60 km. The spit continues to Russia (total length of the spit is ca. 100 km, upper right). P. mugo ssp. uncinata (lower right) and P. mugo 9 P. sylvestris hybrids both were mainly found as single trees at D population. P. mugo ssp. mugo dominates in population R2. The middle right photograph shows the continuous cover of P. mugo on the main dune in D population (the cover is also seen in the background in the middle right photograph) (2006) reported lack of differentiation between the ssp. uncinata, rotundata and mugo based on nuclear RAPD markers in Italian Alps, which suggests an extensive genome sharing between the sub-species. Our study addresses the problem of the genetic structure and diversity of the P. mugo population with largely unknown introduction history to the seaside spit of Nerija. No genetic studies were carried out with the P. mugo material since its introduction to Neringa over 200 years ago. Our hypothesis is that the P. mugo population in Kuršiu˛ Nerija originates from a single or a few of related seed sources and is of low genetic diversity. If true, it may raise concerns when attempting to naturally regenerate the P. mugo stands which are presently approaching the age of natural maturity in Nerija. This issue also has a fundamental interest to investigate what level of genetic diversity is present in a population experiencing several events of introduction (Morgante et al. 1998; see the introduction history above). Similar cases may exist with other species and may have a broader interest. There also is a practical application of our results for gene conservation of the naturalized local ecotype of P. mugo. The objective of our study was to assess the genetic structure and the genetic diversity of the P. mugo plantations in Kursiu Nerija by the aid of the neutral chloroplast DNA markers. Materials and methods The plant material We used the following morphology to identify the P. mugo sub-species (Fig. 2, adult trees): P. mugo ssp. uncinata as monocormic tree with stem diameter exceeding 15 cm and tree height over 8 m (usually straight stem at the base, Fig. 2); P. mugo ssp. rotundata as tall polycormic or monocormic tree with stem diameter less than 15 cm, tree height less than 8 m (usually bending stem at the base; Fig. 2) and P. mugo ssp. mugo as polycormic prostrate shrub reaching the height of 2 m (Fig. 2). The age of a typical ssp. rotundata tree in the population N1 (Nida) was 86 years, that is, established in 1926 (estimated dendrochronologically from stem disc sampled at root collar with diameter of 7 cm). Morphologically, P. mugo ssp. rotundata dominates in Kursiu Nerija and was chosen to study the genetic diversity of the introduced material. For comparison, one morphologically distinct P. mugo spp. mugo stand and single trees of P. mugo spp. uncinata, Pinus sylvestris as well as putative P. mugo x sylvestris hybrids were are also sampled. The individuals possessing the key morphological traits common for both P. sylvestris and P. mugo were identified as putative hybrids in our earlier study (Danusevičius et al. 2012). In total, 249 individuals 123 140 Eur J Forest Res (2013) 132:137–150 Fig. 2 Examples of trees in the sampled populations. For most populations, the most frequent private haplotypes in that population are illustrated. The vertical line indicates the height of 2 m. Arrow points at the sampled tree. Population code above the photograph, tree code below the photograph. The haplotype is given below the photograph. For populations with no private haplotype, the most common haplotype is shown (AV). Populations AVU and DU represent sub-species uncinata, S2—ssp. mugo. HMxS—hybrid P. mugo 9 P. sylvestris. For P. sylvestris (Sylv), a special polycormic morphotype found in Kursiu Nerija is shown from 12 populations were studied (Table 2; Figs. 1, 2). The needles for the DNA analysis were sampled in autumn 2010 and spring 2011. All the sampled populations are artificial plantations, except R2 ssp. mugo, which is a naturally regenerating population. For ssp. rotundata, the sampling strategy was to include the trees from a number of places (populations) over the large planted area with minor interruptions starching on the main dune along the south–north gradient of the spit. At each population of ssp. rotundata, the trees were chosen so that they would represent the common morphotypes observed at the corresponding population and are located at least 10 m apart. In this way, we chose to sample more morphologically diverse than uniform trees within the ssp. rotundata limits. The sample tracking path of 100–200 m within a plantation was made. Exceptions were (a) the AV population representing a circular plantation of ca. 9 ha, surrounded by P. sylvestris, in which the trees were sampled over a 123 Eur J Forest Res (2013) 132:137–150 141 Table 1 Description of the loci studied Locus Speciesa Allele number Sizeb Range Size (range) Heuertz et al. (2010)c Population differentiation p value (s.e.)d pt15169 Mugo 5 123 122–126 125 (119–129) 0.000 (0.000) Sylvest. 5 126 124–128 128 (126–130) pcp17987-1 Mugo 5 108 106–110 Sylvest. 3 110 109–111 Mugo 5 143 141–145 145 (142–148) Sylvest. 3 145 144–146 147 (146–150) pcp87314 Mugo Sylvest. 5 3 112 112 111–115 112–114 0.0511 (0.022) pcp30277 Mugo 6 117 115–120 0.000 (0.000) Sylvest. 5 139 135–139 Mugo 4 169 168–171 Sylvest. 3 172 170–172 pt71936 pcp17987-2 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) The 11 hybrid individuals excluded a Sylvest. means P. sylvestris; number of trees were 18 for P. sylvestris and 220 for P. mugo (sensu lato) b The size indicates the most common allele, range—minimum and maximum alleles c The study included all three P. mugo sub-species—uncinanta, rotundata, mugo d The p value of the chi-square statistics for the exact population differentiation test based on haplotype frequencies Table 2 The geographical location and intra-population genetic diversity estimates of the P. mugo populations and P. sylvestris controls in Kursiu Nerija Population (id)1 S2 Latitude (km)3 Long. 1 Smiltine (S2) R 558380 (47) 218060 2 Smiltine (S1) R 55836 (44) 3 Juodkrante (D) R 55833 (37) 4 Avino ragas (AV) R 5 Naglis (R2) 6 Pervalka (R1) No Hr He D2sh 9.8 8.2 0.97 6.8 0.20 8.3 7.2 0.94 4.9 0.09 20.4 9.1 0.98 5.5 0 0.16 11.0 7.5 0.95 6.1 16 4 0.13 12.9 8.2 0.96 3.2 17 13 3 0.18 10.7 8.2 0.96 6.5 0 N Ah Ph Ch 14 11 5 0.17 21807 15 10 5 218070 31 25 5 558300 (30) 218060 25 14 M 558260 (18) 218050 23 R 558250 (16) 218050 0 Ne 7 Preila (P1) R 55822 (10) 21803 16 12 2 0.19 9.8 8.0 0.96 3.1 8 Nida (N1) R 558170 (0) 208590 23 15 1 0.13 12.3 8.0 0.96 3.9 9 10 Juodkrante (DU) Avino ragas (AVU) U U 55833 (37) 558300 (30) 218070 218060 38 18 22 8 4 0 0.18 0.33 13.6 5.2 7.9 5.2 0.95 0.86 9.5 4.3 11 Nerija (HMxS) H K. Nerija – 11 9 9 0.18 8.1 8.0 0.96 5.1 12 Nerija (Sylv) Sy K. Nerija – 18 15 15 0.17 12.5 8.8 0.97 5.3 Total – – – 249 – 53 Mean – – – 20.7 14.2 4.4 – – – – – 0.2 11.2 7.9 0.95 5.3 Altitude of the localities varied between 5 and 20 m a.s.l. The P. mugo ssp. rotundata populations are arranged from south to north as in Fig. 1 a Population name and its id used in further tables and figures b Sub-species abbreviation: R—rotundata, M—mugo, U—uncinata, H—hybrid mugo (female) x sylvestris (male), Sy—Pinus sylvestris control sampled over whole Lithuanian part of Kursiu Nerija (ca 40 km range) c Distance in km from the southernmost population of Nida (N1) is given in the parenthesis; N number of individuals, Ah number of haplotypes detected in each population, Ph number of private haplotypes, Ch frequency of the most common haplotype in particular population, Ne effective number of haplotypes (the inverse of the probability that two randomly chosen haplotypes are identical), HR haplotypic richness (Mousadik and Petit 1996), He the Nei’s index of genetic diversity estimated without bias (Nei 1973), D2sh the mean genetic distance between individuals within populations (Goldstein et al. 1995, applied to cpSSRs by Morgante et al. 1998) 1.5 km path around and within the plantation and (b) the D population representing a large continuous ssp. rotundata cover from Juodkrante to Klaipeda, where the trees were sampled at three points located within a 5 km interval. The sampled trees were labelled and photographed. The ssp. uncinata does not form stands in Nerija and was found 123 142 dispersed as single trees among the ssp. rotundata plantations (usually at the edge) and were sampled over a greater area of ca. 10 km interval (DU population) and ca. 2 km interval (AVU population). The rather distinct ssp. mugo stand was found at one location only (R2 population). Mainly mature trees were sampled. The molecular analyses We used five highly polymorphic paternally inherited chloroplast SSR markers (Table 1): Pt15169, Pt71936, pcp71987, pcp87314, pcp30277 (originally isolated from Pinus thunbergii by Vendramin et al. (1996) and tested with range-wide P. mugo material by Heuertz et al. (2010) and Dzialuk et al. (2009)). The forward primers were fluorescence labelled for the PCR multiplexing. The DNA was extracted from 100 mg of fresh needles by the CTAB protocol (Doyle and Doyle 1990). The DNA concentration measured with the IMPLEN nanophotometer. The PCR was carried in 10 ll volumes, containing 10x reaction buffer, 2.5 mM MgCl2, 0.2 mM each of dNTP, 30 ng of DNA and 0.5 units of Taq DNR polymerase) with ‘Mastercycler’ thermocycler (Eppendorf). The PCR started with 12 min of denaturation at 95 °C, followed by 25 cycles of denaturation for 30 s at 94 °C, annealing for 45 s at 60 °C and extension at 72 °C for 10 min. The fluorescence-labelled PCR products were separated by the capillary electrophoresis by the GeneScan-500 LIZ standard on the ABI 3130 genetic analyser (Applied Biosystems). The fragments were analysed with the GeneMapper 4.0 software (Applied Biosystems). Eur J Forest Res (2013) 132:137–150 The intra-specific and inter-population differentiation was assessed by the hierarchal AMOVA on the haplotype frequencies by using both the SSR distance types—the sum of squared allele size differences (the RST type distance, considering not only that the alleles are different but also how large the difference is) and based on the number of different alleles (the FST type distance, considering the allele identity only) (2000 permutations; Arlequin ver. 3.1., Excoffier and Lischer 2010). For P. mugo alone, the population differentiation was assessed by comparing the GST and NST fixation indexes calculated with the PERMUT cpSSR version 2.0 software (Pons and Petit 1996). GST is an FST estimate that is based on the haplotype frequencies alone (not considering the population structure of haplotypes), while the NST considers the genetic relatedness among the haplotypes (Pons and Petit 1996). If the NST value is greater than GST value for the molecular data, then the closely related haplotypes tend to be located within a similar area, indicating a geographical structure of the molecular data. The population grouping was assessed with the Principal Coordinates Analysis (PCA) on the population RST distance matrix (from AMOVA, GenAlEx software; Peakall and Smouse 2006) and the UPGMA clustering of the haplotypes by pairwise coancestry genetic distance (DR, Reynolds et al. 1983), the node consistency was tested with 10 000 bootstrap replicates (TFPGA software; Miller 1997). Population similarity was also compared by estimating the number of shared haplotypes between the populations (Arlequin ver. 3.1). Finally, the map plots of population proportions of several most frequent, shared and private haplotypes were made. Data analysis The multilocus haplotypes were constructed, and the allele frequencies for each locus and population were calculated by the aid of the Haplotype Analysis ver. 1.05 software (Eliades and Eliades 2009) and Arlequin ver. 3.1 software (Excoffier and Lischer 2010). The intra-population diversity was studied by calculating the following statistics available in the Haplotype Analysis software: frequencies of private (Ph) and most common haplotypes (Ch), number of different haplotypes (A), effective number of haplotypes (Ne, the inverse of the probability that two randomly chosen haplotypes are identical), haplotypic richness (HR, algorithm used by Petit et al. (1998) in the RAREFAC software; the number of individuals in the smallest population was 11), the Nei’s index of genetic diversity estimated without bias (He, Nei 1973), the mean genetic distance between individual haplotypes within populations (D2sh , Goldstein et al. 1995, applied to cpSSRs by Morgante et al. 1998, considering the allele size difference between the haplotypes). 123 Results Efficiency of the cpSSR markers All the loci studied were polymorphic, producing from 8 to 12 alleles (Table 1, S1). The primer pair pcp17987 amplified two loci for all material including P. sylvestris. It is an efficient marker for studying the cpDNA polymorphism for both P. sylvestris and P. mugo totalling to 15 alleles over the two loci (Table 1, S1). For all the loci except pcp87314, there was a tendency for a higher number of repeats in P. sylvestris and the putative hybrids than in P. mugo (Table 1; S1); especially the locus pcp30277 produced alleles sized above 135 bp solely for the P. sylvestris (18 trees) and the hybrid material (11 trees), whereas for all the remaining 220 P. mugo individuals, the allele size ranged below 120 bp (Table 1; S1). Thus, in comparison with P. mugo, P. sylvestris possessed more private haplotypes by combining the alleles with a higher Eur J Forest Res (2013) 132:137–150 number of repeats at most of the loci (Table 1; S1). The locus pcp30277 was the most efficient for both revealing the intra-specific polymorphism and discriminating between P. mugo and P. sylvestris by producing six P. mugo-specific and another six P. sylvestris-specific alleles (Table 1; Fig. S1). Description of haplotypes Overall the loci, the allele size variants combined into 83 different haplotypes (Tables S2, S3), of these 53 were population private (29 were private for P. mugo sensu lato (s.l.), 15 for P. sylvestris and 9 for the putative hybrid material) (Table S4). Forty-five of the haplotypes were detected only once. For P. sylvestris, 15 of 18 individuals produced a private haplotype. Interestingly, 9 of 11 putative P. mugo 9 P. sylvestris hybrids also possessed a unique haplotype and all these nine haplotypes were different from the P. sylvestris as well as the P. mugo haplotypes (Table 2; S2). The geographical distribution of the five most frequent haplotypes together with the shares of private and the remaining less frequently shared haplotypes is shown in Fig. 3. Most of the frequent haplotypes were shared among the populations, except for the northernmost ssp. rotundata S2 population, ssp. uncinata AVU population and ssp. mugo R2 population (Fig. 3). For instance, the most frequent haplotypes h26 and h23 were found in 23 and 19 copies, respectively, in 7 of 10 P. mugo populations (Table S2). The adjacent populations tended to possess different haplotypic structure (Tables S1, S2; Fig. 3). The ssp. uncinata population AVU had no private haplotypes and shared 6 haplotypes (h26, h11, h13, h50, h28 and h15; Table S3) with the other ssp. uncinata population (DU). These could be the ssp. uncinata-specific haplotypes. Interestingly, both the ssp. uncinata populations shared most of the haplotypes with these populations of ssp. rotundata in which they were found (AVU with AV and DU with D, Table 5). The ssp. rotundata population N1 distinguishing by the relatively taller and monocormic trees also shared most of its haplotypes with the ssp. uncinata populations (Table 5). The northernmost ssp. rotundata populations (S1, S2) contained the lowest number of shared and the highest number of private haplotypes and so were most genetically different from the rest (Table 5; Fig. 2). Intra-population diversity The within-population diversity parameters based on the haplotype frequencies showed that the two adjacent ssp. uncinata populations behaved contrastingly DU possessed higher genetic diversity than AVU (He for DU = 0,95; He for AVU = 0,86; Table 2; Fig. 3). The AVU was 143 especially distinguishing from the rest by high probability to sample identical haplotypes (Ne is the inverse of the probability that two randomly chosen haplotypes are identical, thus, the lower the Ne, the higher the probability to sample identical haplotypes; Table 2). The ssp. mugo population R2 possessed medium diversity estimates but comparably high distance among haplotypes (D2sh ) meaning fewer but more different haplotypes (Table 2). For the ssp. rotundata, the northern D population had highest haplotypic richness (HR) and diversity (He) but medium withinpopulation distance between haplotypes (D2sh ) (meaning that there were many haplotypes but differing in few alleles) (Table 2). The northernmost S2 population possessed slightly lower haplotypic diversity than the rest (Table 2). There was no marked difference in haplotypic diversity and richness between northern and southern ssp. rotundata populations (Table 2). However, there was a tendency for a higher number of private haplotypes and the h27 h23 S2 h11 h6 10 km h26 S1 h23 h11 h6 D h26 h23 h26 h23 h11 AV h27 h11 h6 Juodkrante h6 h23 R1 DU (uncinata) h26 h27 h6 AVU (uncinata) h11 Pervalka h26 h26 h27 h23 P1 Preila h6 R2 (mugo) Nida h26 h23 h11 N1 h27 h26 h27 h23 h6 h11 Shared Private Fig. 3 Distribution of the multilocus haplotypes in the populations. The five most frequent haplotypes in the whole material (frequency [0.04; found in 10 and more individuals) are shown separately. The remaining haplotypes were pooled into two groups—shared haplotypes (found in more than one population) and private (found only in a particular population). P. sylvestris and the P. mugo 9 P. sylvestris hybrids possessed private haplotypes only (Table 2) and not shown in this figure 123 144 Eur J Forest Res (2013) 132:137–150 Table 3 The AMOVA results based on RST and FST Source variation df Variance components % variation p value (s.e.) RST Among species 4 67.60 91.98 0.0401 (s.e. = 0.0065) Among populations within species 7 0.05 0.07 0.2024 (s.e. = 0.0112) 237 5.90 7.95 0.0000 (s.e. = 0.0000) Among species 4 0.39 18.58 0.0176 (s.e. = 0.0042) Among populations within species 7 0.04 1.93 0.0225 (s.e. = 0.0044) 237 1.66 79.49 0.0000 (s.e. = 0.0000) Within populations FST Within populations higher distance among haplotypes (D2sh ) in the northern pupations (S2, S1, D, AV) than in the southern populations (P1, N1) (Table 2). Inter-population differentiation The differentiation between the Pinus mugo populations in Nerija was low with the GST and RST values of 0.004 and 0.002, respectively. This indicates weak genetic differences between the P. mugo populations representing different parts of Nerija and leads to an assumption that the genetic structure of P. mugo in Nerija depends on other factors than the geographical locations within Nerija and its introduction sources may well fall within few geographically close areas. The analysis of molecular variance (AMOVA) based on haplotype frequencies including P. sylvestris, and the putative hybrids calculated by the FST and RST methods yielded slightly different results (Table 3). For the FST, the AMOVA showed that all the sources of variation considered by us had a significant effect on the haplotype frequencies, the major part of the haplotypic variation being within populations (Table 3). For the RST, which considers the allele size difference, the major part of the variation was among the species and the effect of population within species was not significant (in agreement with the RST over GST test described above, Table 3). The differences between the sub-species were significant for both RST and FST methods, but for the later statistics, the significance was greater (Table 3). The test of the pairwise population differentiation (Markov chain length: 100,000 steps, Arlequine) is summarized in Table 4, where P. sylvestris and the hybrid haplotypes were significantly different from the rest. The ssp. mugo population (R2) was significantly different from both ssp. uncinata populations (DU and AVU) and ssp. rotundata populations of AV (central) and R1, the latter being the adjacent population (Table 4; Figs. 1, 2). Among ssp. rotundata, most of the populations were not significantly different in their haplotype frequencies, except the significant differences between the central AV and northernmost S1, S2 populations (Table 4). The northern D population possessed not significantly different haplotypic constitution from all the other populations, including spp. uncinata. The two adjacent northern populations S1 and S2 had significantly different Table 4 Results of the test of differentiation between populations Population AV AVU D AVU 0.198 D 0.884 DU 0.836 0.133 0.745 N1 0.075 0.022 0.523 DU N1 P1 R1 R2 S1 SYLV 0.005 0.017 0.034 0.149 0.511 P1 0.224 0.025 0.618 0.394 0.520 R1 0.579 0.000 0.853 0.330 0.112 0.212 R2 0.004 0.004 0.060 0.061 0.236 0.564 0.040 S1 0.052 0.000 0.796 0.047 0.004 0.057 0.093 S2 0.007 0.002 0.104 0.014 0.341 0.301 0.229 0.181 0.001 SYLV HMxS 0.000 0.000 0.000 0.000 0.002 0.008 0.000 0.001 0.000 0.001 0.006 0.014 0.007 0.018 0.001 0.003 0.003 0.008 0.000 The non-differentiation exact p values for the chi-square statistics are given (Markov chain length: 100,000 steps) 123 S2 Eur J Forest Res (2013) 132:137–150 haplotypes. The southernmost population N1 was significantly different than the northern S1 population and showed a similarity with ssp. uncinata in DU (compare haplotypes in Fig. 3). Population groups The UPGMA clustering showed that the haplotypic variation followed species- and sub-species-like structure and revealed two major groups: P. sylvestris and the putative hybrid being markedly different from the rest (Fig. 4). More interestingly, the P. mugo complex clustered into ssp. mugo-like (G1), ssp. uncinata-like (G2) and ssp. rotundata-like groups (G3, G4; Fig. 4). However, for the P. mugo UPGMA groups, less than 50 % of the bootstraps returned the same cluster structure, indicating minor differences between the groups. (G1) contains the ssp. mugo R2 population and the two rotundata populations S2 and P1; note that, the S2 represents the shrubby morphological margin of ssp. rotundata closer to ssp. mugo (see Fig. 2). (G2) contains the two ssp. uncinata populations and the southernmost ssp. rotundata population N1, in which represent the morphological margin of ssp. rotundata closer to ssp. uncinata (the uncinata-like morphotypes dominated: thick, straight but small, less than 10 metres tall, Fig. 3). (G3) rotundata subgroup with central and southern populations of rotundata D, R1, AV, for all of which a relatively higher variation of shrub-like to tree-like morphotypes, was observed (tree-like morphotypes not shown, but some of the shrubby forms shown in Fig. 4). (G4) rotundata subgroup of a single population S1, which can be morphologically characterized as less variable and typical representative of rotundata (ca. 2.5–3 m tall, curvy, well expressed bending at the base, polycormic), and with this rotundata-like purity, S1 is distinct from the more morphologically variable D, and R1 and AV populations. 145 Regarding the PCA results, the first two principal components explained 98.0 % of the total variation. The first PC differentiated the P. mugo sub-species complex from P. sylvestris and accounted for 90 % of the total variation. The second PC represented the variation between P. mugo populations and accounted for the remaining 8 % (Fig. 5). The differentiation between ssp. uncinata and the rest seems to be weaker than in the UPGMA clustering (Fig. 5). Only the ssp. uncinata (AVU) stands out, but the other ssp. uncinata population (DU) is within a group ssp. rotundata. However, the ssp. mugo population R2 stands out as a rather separate group (Fig. 5). For P. mugo ssp. rotundata, the contrasting differences were between the Nida (N1, southmost) and Juodkrante (D, north) populations, thought, the northernmost (S and S2) remained in the intermediate PCA cluster (Fig. 5). Morphologically, trees in the N1 population were taller and contained more of straight monocormic uncinata-like morphotypes than in the D population (Fig. 2). Discussion General All the six loci studied were polymorphic and provide a reliable estimate of the polymorphism of P. mugo population in Neringa at the paternally inherited chloroplast genome level. The locus pcp30277 is a good candidate for a species-specific marker, for which there was a 15 bp difference between the fragments of P. sylvestris (including the putative hybrids) and the P. mugo s.l. (Table 1; S1). This size difference is separable by agarose gels and after the validation with a geographically broader material could be a cheap marker discriminating the morphologically questionable individuals. In contrast to Heuertz et al. (2010), there was no haplotype sharing between P. mugo s.l. and P. sylvestris in our Fig. 4 The UPGMA clustering of the populations using the coancestry distance (Reynolds et al. 1983) (left). The numbers show the proportion of similar replicates for a specific node from the bootstrapping procedure (10,000 permutations) and so indicate the similarity of the populations within a node (higher number means more similar). The four emerging groups of P. mugo populations are outlined and showed on the location map to the right 123 146 Eur J Forest Res (2013) 132:137–150 Fig. 5 Population clustering based the two first principal components from the PCA analysis. The clusters of populations are outlined subjectively based proximity and depicted on the map (right) study. We used more loci together with the species-specific locus 30277 (Table 1), which as such prevented the haplotype sharing. For the two loci used by Heuertz et al. (2010), we also found the shared alleles (Table 1; S1). Another reason may be that, our P. sylvestris may represent a different refugial line than studied by Heuertz et al. (2010) (Naydenov et al. 2007). Haplotypic variation reflects sub-species differentiation UPGMA clustering indicates the sub-species morphotype may have stronger effect on the seemingly random haplotypic variation of the P. mugo material than the geographical location of the populations (Fig. 4). However, for the P. mugo material, less than 50 % of the bootstraps returned the same cluster structure, indicating minor differences between the groups. The ssp. mugo group (G1) mainly contained polycormic small shrubs (as population R2 in Fig. 2), except P1 which morphologically resembled the typical spp. rotundata population S1 and maybe therefore were close to each other in the PCA plot (Fig. 5). The haplotypes in P1 and R2 populations were relatively more uniform (D2sh in Table 2). The ssp. uncinata group (G2) of tall monocormic trees was genetically closer to ssp. rotundata than to ssp. mugo (Fig. 4). Even being in the same UPGMA node, the two ssp. uncinata populations were variable, the DU population being much more polymorphic (Fig. 2). This may be connected to a higher number of samples in DU or 123 sampling over a greater area than in AVU. Both the ssp. uncinata populations contained a low number of private haplotypes, indicating mating with other P. mugo subspecies. Morphologically, the northernmost ssp. rotundata population N1 was closer to ssp. uncinata than the other ssp. rotundata populations and could partially originate from mating between ssp. uncinata and rotundata. The main difference between the remaining two ssp. rotundata population groups G3 and G4 is that the S1 population (forming the G4 group alone) has a high number of private but less diverse haplotypes (Ph vs. D2sh in Table 2) and so may represent more pure ssp. rotundata population (based on our morphological observation as well). However, this finding was not straightforward. Firstly, poor sandy soil and high salinity may markedly affect the tree morphology and disturb the sub-species identification (we found individuals of intermediate morphotypes between the P. mugo sub-species, especially between ssp. rotundata and uncinata). Secondly, the cpSSRs capture the paternity variation and are affected by mating among parents of possibly variable sub-species in original seed collection areas. Also note that the ssp. uncinata morphotypes shared most of the haplotypes with the ssp. rotundata populations in which they were selected (Table 5). Our finding contrasts the earlier studies, where no significant P. mugo sub-species differentiation was observed (Monteleone et al. 2006; Heuertz et al. 2010). The latter author in the range-wide study with 3 cpSSR markers (2 of which we used also) found that the P. mugo haplotypic Eur J Forest Res (2013) 132:137–150 147 Table 5 Number of shared haplotypes between the P. mugo populations Population AV AV AVU AVU D DU N1 P1 R1 R2 6 13 10 7 6 8 3 3 7 6 4 2 2 1 1 1 18 10 8 9 5 8 3 6 D 13 7 DU 10 6 18 N1 7 4 10 11 11 S1 S2 7 8 5 5 5 5 4 5 2 5 P1 6 2 8 7 5 R1 8 2 9 8 4 4 4 R2 3 1 5 5 5 5 3 S1 3 1 8 5 2 3 2 5 3 5 3 2 4 3 3 4 4 S2 1 3 5 5 5 4 4 4 2 2 Mean 6.3 3.6 9.2 8.3 5.9 4.9 4.9 3.8 3.2 3.7 P. sylvestris and the hybrid material possessed private haplotypes only variation follows geographical rather than sub-species structure. The reason for the different results could be a higher resolution and number of markers as well as small geographical range in our study. Monteleone et al. (2006) used RAPD markers and also failed to find difference between the P. mugo sub-species in Italy. RAPD being efficient indicators of diversity may not be the right type of marker for genetic differentiation studies. In conifers, the cpSSR markers reflect the haplotype of the male parent (Neale and Sedoroff 1998) and are useful in revealing the mating patterns. As regards the intermating within the P. mugo taxa, both the uncinata populations shared most of their haplotypes with these populations of ssp. rotundata in which they were found (AVU with AV and DU with D, Table 5). This may indicate genome sharing between the P. mugo sub-species and favours the sensu lato taxonomic approach (Christensen 1987a, b). The ssp. uncinata population AVU had no private haplotypes and shared 6 haplotypes (h26, h11, h13, h50, h28 and h15; Table S3) with the other ssp. uncinata population (DU). These could be the ssp. uncinata-specific haplotypes. Almost all P. sylvestris individuals had private haplotypes because it is another species than P. mugo. Whereas, the high number of private haplotypes in the putative hybrid material may indicate that male parents of the putative hybrids are of non-local and genetically different origin than the local P. sylvestris. This leads to an assumption that these hybrids were set outside Lithuania in the foreign sources of introduction (e.g. Denmark). How diverse is the material of P. mugo in Nerija? Theoretically, the exotic P. mugo population in Nerija experiencing several stages of introduction (Alps?–Denmark–Lithuanian) should possess a reduced genetic diversity. However, this was not the case, and the total genetic diversity of P. mugo material in Nerija was high (He = 0.953), which was similar to the corresponding He estimates from the range-wide cpSSR studies of P. mugo s.l. (He = 0,956; 3 cpSSR loci, Heuertz et al. (2010)), ssp. uncinata in Pyrenees of (He = 0.987; 10 cpSSR loci; Dzialuk et al. 2009) and even P. sylvestris in Spain (He = 0.978; Robledo-Arnuncio et al. 2005). The haplotypic diversity was also high—83 haplotypes at 6 loci (compare 100 haplotypes at 3 loci by Heuertz et al. (2010) and 174 at 9 loci by Dzialuk et al. (2009)). The D2sh values showing similarity between the haplotypes within populations varied between 3 and 10 in our study and were less variable than in Heuertz et al. (2010) (1–19) and Dzialuk et al. (2009) (4–14). The effective number of haplotypes (Ne) was 11.2 and 16.9 in our study and Dzialuk et al. (2009), respectively. The D2sh seemed to be little depending on the number of individuals within population (Table 2, e.g. N1 population). Such levels and structure of genetic diversity indicate low coancestry levels of the P. mugo plantations in Neringa and, in case of natural regeneration, leave low probability for an inbreeding depression in their progeny. Interestingly, in the Dzialuk et al. (2009) investigation on the cpSSR polymorphism of the range-wide ssp. uncinata material, the locus pt71936 was monomorphic. In our study, at the locus pt71936, the ssp. uncinata alleles ranged from 141 to 145 and overlapped with the allele range for ssp. rotundata and ssp. mugo (Table S1). This may indicate high mating rates between the P. mugo sub-species within the Danish or Lithuanian P. mugo populations (Christensen 1987a). This intra-specific introgression may increase the genetic diversity of the P. mugo in Kursiu Nerija close to the estimates found in the autochthonous populations (Table 2, Dzialuk et al. 2009; Heuertz et al. 2010). Christensen (1987a, b) morphological observations support the introgressive hybridization hypothesis, where the western 123 148 European ssp. uncinata-type and the eastern European ssp. mugo-type are separated by an intermediate morphotypes in the central introgression zone (Heiser, 1973). A morphological support for this intra-specific hybridization in our material is the large number of the intermediate morphotypes between ssp. uncinata and rotundata (the thick, monocormic, straight but not as tall as expected for uncinata morphotypes). Inferences on introduction of P. mugo If the first plantations in Nerija were established in 1825 not far from the southern N1 population (Strakauskaite 2004), then it would have been practically more convenient to collect the local seed than importing them when establishing the N1 population in 1926 and the rest of the plantations (which seemed to be of similar age as the N1). Unfortunately, to our knowledge, there are no records on the origin of the reproductive material used in Nerija before the WWII. What our genetic data may reveal is how diverse is the genetic material and is there any geographical structure of P. mugo plantations in Nerija. If the northern populations in Nerija (S1, S2, D) would originate from the seed collections in the southern populations (N1, P1), one would expect low population differentiation, high haplotype sharing, absence of private haplotypes and low level of genetic diversity, especially of the later established northern populations. Some of these assumptions were true in our material: no geographical structure (neighbours were assigned to different clusters, Fig. 3), the population differentiation was low (NST \ GST; the AMOVA in Table 3), the northern and southern populations share most of the 5 most frequent haplotypes. However, there was a notable proportion of private haplotypes and high genetic diversity in most of the populations (Fig. 3; Tables 2, 5). There even was a tendency of the southern populations (N1 and P1) to possess lower genetic diversity than the northern populations S1 and D (supposed to be established later) (Table 2). This contradicts the expectation that the northern material originates from the seed collections in the southern material in Nerija (see also high number of private haplotypes in southern populations S1 and S2, Fig. 3). The above given genetic data indicate that the greater part of the P. mugo plantations in Nerija originates within one geographical region (such as Denmark), but there also could be introductions from other less genetically related sources (populations S1, S2 with high proportions of private haplotypes) as well as collections within Nerija itself (population AV with no private haplotypes at all). Another observation is that in spite of the rare application of genetic principles in forestry back in the nineteenth century, the seed for the P. mugo afforestation was collected from sufficient number of female parents to form genetically 123 Eur J Forest Res (2013) 132:137–150 diverse plantations. Exceptions are the low diversity AV and AVU populations which represent a rather isolated area of ca. 4 ha surrounded by P. sylvestris stands. The diversity indexes of AV were relatively lower, even though the individuals were sampled over a greater area than in most of the remaining populations. Presumably, the AV population originates from a local seed source. To track the origin of the P. mugo in Denmark and Lithuania, a study with common DNA markers is needed to compare the genetic composition of the introduced populations with that of at variable parts of the species range. Suggestions for gene conservation For gene conservation of the P. mugo ecotype in Nerija, our results suggest first to capture the among-taxa diversity by establishing one conservation unit for each ssp. uncinata and ssp. mugo. For ssp. uncinata which does not form stands in Nerija, 40–50 single individuals could be sampled for an ex situ conservation unit such as clonal archive. For ssp. mugo, the R2 stand with the widely spaced naturally regenerating trees could be a good candidate for an in situ genetic reserve, which needs no further investment but occasional measures to promote its natural regeneration. For the most widespread ssp. rotundata, the low population differentiation and the absence of a clear geographical structure would argue for one large in situ genetic reserve. However, the tendency for the genetic separation between the tall monocormic morphotype (e.g. RN12, RV11 in Fig. 2) and the shorter polycormic morphotype of ssp. rotundata (e.g. RD26, RD34, UAV13 in Fig. 2) advocates for two genetic reserves in the plantations with dominance of either of these two morphotypes. A good candidate for the short polycormic ssp. rotundata could be the northern S2 population with high frequency of private haplotypes, whereas, for the tall monocormic ssp. rotundata, one of the southern populations N1 or P1 could be considered. It is safer to establish the genetic reserves in large forest tracts to avoid sampling stands originating from low diversity local collections (as AV in our case). Concluding remarks Based on the genetic marker data, we have to reject the hypothesis that the P. mugo population in Nerija originates from a single seed source and is of low genetic diversity. The genetic diversity of the P. mugo material in Nerija is high as a whole and high in most its range in Nerija. Low population differentiation together with high haplotypic diversity and high number of shared haplotypes in most of the populations indicates that the major part of the P. mugo material originates from a number of geographically and genetically related sources of high genetic Eur J Forest Res (2013) 132:137–150 diversity. On the other hand, a notable share of private haplotypes in most of the populations shows that these sources are less likely to be within Nerija, because otherwise, a much lower number of private haplotypes is expected. The high frequency of private haplotypes in the northernmost plantations leaves a possibility for minor introductions from other genetically distinct sources. There also may exist small plantations established with local seed, as indicated by the shortage of the private haplotypes in the AV plantation. Another observation is that the genetic diversity of the Pinus mugo material in Kursiu Nerija is structured based on the sub-species and rather than on geography. The large number of shared haplotypes provides a strong evidence for a gene flow among the P. mugo taxa. The locus pcp30277 is a good candidate for a cheap species-specific marker distinguishing the paternity between P. sylvestris and P. mugo. The implications for gene conservation are first to capture the sub-species variation and then identify morphotype-based ssp. rotundata conservation areas in different parts of Nerija. Acknowledgments This research was funded by a grant (LEK-11/ 2010) from the Research Council of Lithuania. The authors also want to express appreciation to the staff of Kursiu Nerija National Park for the field guidance. References Anderson E, Hubricht L (1938) Hybridization in Tradescantia 111. The evidence for introgressive hybridization. Am J Bot 25:396–402 Boratyńska K, Boratyński A (2007) Taxonomic differences among closely related pines Pinus sylvestris, P. mugo, P. uncinata, P. rotundata and P. uliginosa as revealed in needle sclerenchyma cells. Flora 202:555–569 Boratyński A, Boratyńska K, Lewandowski A (2003) Evidence of the possibility of natural reciprocal crosses between Pinus sylvestris and P. uliginosa based on the phenology of reproductive organs. Flora 198:377–388 Christensen KI (1987a) Taxonomic revision of the Pinus mugo complex and P. 9 rhaetica (P. mugo 9 sylvestris) (Pinaceae). Nord J Bot 7:383–408 Christensen KI (1987b) A morphometric study of the Pinus mugo Turra complex and its natural hybridization with P. sylvestris L. (Pinaceae). Feddes Repert 98:623–635 Christensen KI, Dar GH (1997) A morphometric analysis of spontaneous and artificial hybrids of Pinus mugo x sylvestris (Pinacae). Nord J Bot 17:77–86 Danusevičius D, Marozas V, Brazaitis G, Petrokas R (2012) Spontaneous hybridization between Pinus mugo (Turra) and Pinus sylvestris (L.) at the Lithuanian sea-side: a morphological survey. Sci World J 11. doi:10.1100/2012/172407 Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12(1):13–15 Dzialuk A, Muchewicz E, Boratynski A, Montserrat JM, Boratynska K, Burczyk J (2009) Genetic variation of Pinus uncinata 149 (Pinaceae) in the Pyrenees determined with cpSSR markers. Plant Syst Evol 277:197–205 Eliades N-G, Eliades DG (2009) User’s manual for HAPLOTYPE ANALYSIS: software for analysis of haplotypes data. Forest genetics and forest tree breeding, Georg-Augst University Goettingen, Germany. URL http://www.uni-goettingen.de/en/134935.html Excoffier L, Lischer H (2010) ARLEQUIN VER 3.5.1.2 USER MANUAL (2010) Computational and molecular population genetics lab (CMPG). Institute of Ecology and Evolution University of Berne, Baltzerstrasse 6, 3012 Bern, Switzerland Goldstein DB, Linares AR, Cavallisforza LL, Feldman MW (1995) An evaluation of genetic distances for use with microsatellite loci. Genetics 139:463–471 Hamernik J, Musil I (2007) The Pinus mugo complex—its structuring and general overview of the used nomenclature. J For Sci 53:253–266 Heiser CB (1949) Natural hybridization with particular reference to introgression. Bot Rev (Lanc) 15:645–687 Heiser CB (1973) Introgression re-examined. Bot Rev (Lanc) 39:347–366 Heuertz M, Teufel J, Gonzalez-Martınez SC, Soto A, Fady B, Alia R, Vendramin GG (2010) Geography determines genetic relationships between species of mountain pine (Pinus mugo complex) in western Europe. J Biogeogr 37:541–556 Jørgensen H (2006) NOBANIS—Invasive Alien Species Fact Sheet— Pinus mugo. From: Online database of the North European and Baltic Network on Invasive Alien Species—NOBANIS. www.nobanis.org Kormutak A, Ostrolucka M, Vookova B, Pretova A (2005) Artificial hybridization of Pinus sylvestris L. and Pinus mugo Turra. Acta Biol Cracoviensia Ser Bot 47:129–134 Kormutak A, Demankova B, Gmory D (2008) Spontaneous hybridization between Pinus sylvestris L. and P. mugo Turra in Slovakia. Silvae Genet 57(2):76–82 Lewandowski A, Boratynski A, Mejnartowicz L (2000) Allozyme investigations on the genetic differentiation between closely related pines–Pinus sylvestris L., P. mugo Turra, P. uncinata Ramond ex DC. and P. uliginosa Neuman (Pinaceae). Plant Syst Evol 221:15–24 Marcet E (1967) Über den Nachweis spontaner Hybriden von Pinus mugo Turra und Pinus sylvestris L. Aufgrund von Nadelmerkmalen. Berichte Schweizer Botanischen Gesellschaft 77:314–361 Mar-Möller C (1965) Vore skovträer og deres dyrkning. Copenhagen Miller MP (1997) Tools for population genetic analyses (TFPGA) version 1.3. A WindowsÒ program for the analysis of allozyme and molecular population genetic data. Department of Biological Sciences, Box 5640, Northern Arizona University, Flagstaff, AZ 86011-5640 Monteleone I, Ferrazzini D, Belletti P (2006) Effectiveness of neutral RAPD markers to detect genetic divergence between the subspecies uncinata and mugo of Pinus mugo Turra. Silva Fennica 40(3):391–406 Morgante M, Felice N, Vendramin GG (1998) Analysis of hypervariable chloroplast microsatellites. In: A Karp (ed) Pinus halepensis reveals a dramatic bottleneck. Molecular tools for screening biodiversity: plants and animals. Chapman and Hall, London, pp 402–412 Mousadik A, Petit RJ (1996) High level of genetic differentiation for allelic richness among populations of the argan tree [Argania spinosa (L) Skeels] endemic to Morocco. Theor Appl Genet 92:832–839 Müller PE (1887) Om Bjergfyrren (Pinus montana, Mill.). Et Forsög i anvendt Plantegeografi. Copenhagen Naydenov K, Senneville S, Beaulieu J, Tremblay F, Bousquet J (2007) Glacial vicariance in Eurasia: mitochondrial DNA evidence from Scots pine for a complex heritage involving genetically distinct refugia at mid-northern latitudes and in Asia Minor. BMC Evol Biol 7:233. doi:10.1186/1471-2148-7-233 123 150 Neale DB, Sederoff RR (1989) Paternal inheritance of chloroplast DNA and maternal inheritance of mitochondrial DNA in loblolly pine. Theor Appl Genet 77:212–216 Nei M (1973) Analysis of gene diversity in subdivided populations. Proc Natl Acad Sci USA 70:3321–3323 Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6:288–295 Petit RJ, El Mousadik A, Pons O (1998) Identifying populations for conservation on the basis of genetic markers. Conserv Biol 12:844–855 Pons O, Petit RJ (1996) Measuring and testing genetic differentiation with ordered versus unordered alleles. Genetics 144:1237–1245 Reynolds J, Weir BS, Cockerham CC (1983) Estimation of the coancestry coefficient: basis for a short-term genetic distance. Genetics 105:767–779 Robledo-Arnuncio JJ, Collada C, Alıa R, Gil L (2005) Genetic structure of mountane isolates of Pinus sylvestris L. In a Mediterranean refugial area. J Biogeogr 32:595–605 123 Eur J Forest Res (2013) 132:137–150 Schmid J (2000) DNA- und Isoenzym-Polymorphismen in Populationen des Berg-Kiefern-Komplex, Pinus mugo Turra s.l. PhD Thesis, Albert-Ludwigs-University, Freiburg Strakauskaite N (2004) Klaipeda and the Curonian spit guide. R. Paknio publishing, Klaipeda, p 80. ISBN 9986-830-87-7 Vendramin GG, Lelli L, Rossi P, Morgante M (1996) A set of primers for the amplification of 20 chloroplast microsatellites in Pinaceae. Mol Ecol 5:595–598 Wachowiak W, Prus-Głowacki W (2008) Hybridisation processes in sympatric populations of pines Pinus sylvestris L., P. mugo Turra and P. uliginosa Neumann. Plant Syst Evol 271:29–40 Wachowiak W, Ba˛czkiewicz A, Celiński K, Prus-Głowacki W (2004) Species-specific chloroplast DNA polymorphism in the trnVrbcL region in Pinus sylvestris and P. mugo. Dendrobiology 51:67–72