Genetic variation reveals complex population structuring of Tomicus piniperda L. (Coleoptera, Scolytidae) in the UK: Implications for management of this important pest

Tomicus piniperda is a common pest of pine trees responsible for significant economic damage. Although the impact of T. piniperda on forestry is likely to increase in the future due to climate change, little is known about its dispersal within the United Kingdom (UK) or between continental Europe and the UK. This study addressed these knowledge gaps using mitochondrial and microsatellite DNA markers. Mitochondrial DNA revealed high levels of genetic diversity as previously reported for this species across Europe. Nuclear DNA diversity revealed two geographically incoherent groups (Scotland, Wales, Belgium and the Netherlands vs. England and Switzerland). These genetic patterns likely reflect a combination of historical gene flow, anthropogenic introduction and restricted contemporary gene flow between the UK and Europe. Significant structure was found among UK samples suggesting T. piniperda might not disperse extensively between pine stands. This differs from widespread gene flow across continental Europe and may be due to the low percentage of land covered by forests in the UK in addition to the high fragmentation of its forest cover. Genetic variation was higher among the Scottish samples compared to those from more southern clines, suggesting the potential proximity of Scotland to a Pleistocene glacial refuge for T. piniperda.


INTRODUCTION
Bark beetles are the most damaging pests of conifer forests worldwide (Janes & Batista, 2016), and their impact is predicted to increase in the future due to climate change and the growth of international trade (Avtzis & Lakatos, 2021).These pests are easily moved around and spread to new areas as a result of national and international trade with the main introduction pathways being the movement of wood packaging material, logs, processed woods and containers (Meurisse et al., 2019).Tomicus piniperda L. (Coleoptera, Scolytidae) is the member of its genus with the widest geographic distribution, covering the Palearctic area from western Europe (Portugal) to east Asia (Japan) (Lieutier et al., 2015).It was also found in North America where it is invasive and where it has spread from Ohio (USA) to several other states and to Canada (Lieutier et al., 2015).It is a cold-tolerant species that cannot develop in warm and dry climates (Horn et al., 2012).
T. piniperda is a very common and important pest, including in the United Kingdom (UK) where it is responsible for significant economic damage (Lieutier et al., 2015;Wainhouse & Inward, 2016).T. piniperda can attack a variety of conifer species but Scots pine (Pinus sylvestris) is its main host throughout its range (Hanson, 1937;Lieutier et al., 2015).Scots pine is one of only three native conifers in the UK, together with yew and juniper (Taxus baccata and Juniperus communis) and the only native species used for commercial forestry, representing 17% of the trees grown (Edlin, 1970;Forest Research, 2021).
T. piniperda has high dispersal abilities, most likely a result of the rarity of suitable breeding material in the landscape (Kerdelhué et al., 2006), with adults believed to cover several kilometres during their spring flights (Forsse, 1989).Larvae develop in weakened or freshly dead trees as the species cannot survive the defences of healthy conifers (Kerdelhué et al., 2006;Lieutier, 2002).However, T. piniperda can attack and breed in trees that are weakened by other factors such as insect-induced defoliation or drought and so increase tree mortality (Wainhouse & Inward, 2016).In the UK, T. piniperda has contributed to the death of trees that were heavily defoliated by the pine looper moth Bupalus piniaria (Forest Research, 2023;Wainhouse & Inward, 2016).After emerging from breeding galleries within trees, young adult T. piniperda feed on healthy and vigorous pine shoots in order to complete maturation, which can lead to significant tree damage when the beetle population is high (Hanson, 1937;Lieutier et al., 2015;Wainhouse & Inward, 2016).Substantial growth loss is reported when maturation feeding occurs in young plantations or when a large proportion of shoots is destroyed (Hanson, 1937;Lieutier et al., 2015;Wainhouse & Inward, 2016).T. piniperda is also responsible for indirect damage when it vectors other pests, for example, blue stain fungi (Lieutier et al., 2015;Solheim et al., 1993) and the pitch canker disease pathogen (Fusarium circinatum; Bezos et al., 2015).As for other forest pest species, climate change is predicted to increase the damage caused by T. piniperda to Scots pine trees (Wainhouse & Inward, 2016).
Knowledge of pest dispersal is necessary to understand the scale at which to implement management, surveillance plans and to assess the efficacy of management techniques (Mazzi & Dorn, 2012;Miller & Sappington, 2017).Genetic markers are powerful tools that allow the direct and indirect study of dispersal (Broquet & Petit, 2009).Population studies of T. piniperda across continental Europe have revealed high levels of genetic variation, high gene flow across large geographical areas (Horn et al., 2009;Kerdelhué et al., 2006;Ritzerow et al., 2004) and signals of persistence in, and expansion from, multiple glacial refugia (Horn et al., 2009;Ritzerow et al., 2004).None of these studies included samples from the UK, so it is not clear if there is connectivity of T. piniperda populations between continental Europe and the UK.Furthermore, little is known about the dispersal of the beetle between Scots pine stands in the UK (Lieutier et al., 2015) and this information could help control efforts by the British forestry sector.
Accordingly, the aim of this study was to evaluate the genetic structure of T. piniperda within the UK as well as connectivity between UK and European conspecific populations.Mitochondrial DNA analysis was employed to permit direct linking of data with that from previous studies (Horn et al., 2009).In addition, nuclear microsatellite DNA markers were employed as they may be more sensitive to recent population processes due to their more rapid rate of evolution.Three specific questions were addressed.Firstly, is the English Channel a barrier to dispersal, as found in other insect species (Minot & Husté, 2022;Thomas et al., 2021)?Secondly, is dispersal in the UK limited by environmental factors such as host distribution or physical barriers or is there long-distance dispersal within the country?Thirdly, is genetic diversity in the UK lower than in European populations in accordance with the classic southern richness northern purity model of post-glacial colonization of northern Europe (Hewitt, 2000)?The results of the present study will inform the management and surveillance of T. piniperda and provide insights on possible patterns of dispersal of other bark beetles with similar dispersal abilities, size and body mass.

Insect samples collection
T. piniperda from the UK (England, Wales and Scotland-Table 1 and   Figure 1) were collected between 2019 and 2021, either as immatures (i.e., larvae or pupae) or as adults.In Wales, one larva per Tomicus-like gallery was collected in each gallery found beneath the bark of Scots pine logs cut and left in billet piles in the forest or from freshly felled trees.
Larvae were preserved in 99% ethanol until further analyses and specimen identification were confirmed with DNA barcoding before any further investigations.In England, Wales and Scotland, adult beetles were caught in Scots pine forests using Wita ® Prall Cross-vane panel traps baited with ethanol and α-pinene.All individuals were stored in 99% ethanol until identification based on morphological characteristics and further analyses.Beetles from Europe (Belgium, the Netherlands and Switzerland-Table 1 and Figure 1) were provided by European experts to obtain samples from the continental side of the English Channel and to assess genetic diversity within populations and connectivity between populations on either side.Larva specimen identification was confirmed with DNA barcoding before any further investigations, and adult identification was checked with morphological keys before any further analyses.

DNA extraction
Total genomic DNA from entire larvae or adults was extracted following a Cetyl Trimethylammonium Bromide (CTAB)-chloroform/isoamyl alcohol method (Winnepenninckx, 1993).All samples were incubated overnight at 37 C in CTAB buffer with Proteinase K. Additionally, genomic DNA was extracted in a different way for the beetles loaned by the Royal Belgian Institute of Natural Sciences museum.For those samples, DNA was extracted from only two legs per beetle in order to preserve the samples.E.Z.N.A. ® Insect DNA Kit (Omega Bio-tek) was used following the recommended protocol except for the following steps: (1) the sample preparation step was done by bead beating the legs into the extraction buffer instead of using liquid nitrogen and (2) the centrifugation time of step 5 (following the addition of chloroform:isoamyl alcohol [24:1]) was increased to 15 min instead of 2 min (Omega Bio-tek, 2023).

Mitochondrial DNA amplification and analysis
A 950 base pair (bp) fragment of the mitochondrial cytochrome oxidase I and II genes was amplified using T. piniperda-specific primers used in previous studies by Kerdelhué et al. (2002) and Horn et al. (2009).This permitted the incorporation of data from across a wider geographical range, by allowing integration with sequences from the wider study of European populations by Horn et al. (2009).
T A B L E 1 Tomicus piniperda sampling locations, sample size and type of sample used.(Hall, 1999;Thompson et al., 1994).

Levels of genetic variability and diversity within sampling sites
were assessed with ARLEQUINv3.5.2.2 (Excoffier & Lischer, 2010).S1), and haplotypic relationships were inferred using a Median-joining network constructed using PopART (Leigh & Bryant, 2015).Additionally, a map showing the amplified COI-COII region haplotype frequencies per sampling site for T. piniperda was drawn in R studio using the package ggplot2 (Wickham, 2016).

Microsatellite amplification and analysis
All individuals sampled by the present study were genotyped at five microsatellite DNA loci to assess multi-locus nuclear genetic diversity within and between samples, using primers developed by Kerdelhué et al. (2003).All microsatellite loci were amplified in 10 μL PCR volumes with 0.5 μmol of each primer, 5 μL of 2XBioMix (Bioline, UK), 2 μL of ddH 2 O and 2 μL of template DNA.For four loci (TP-CT2-8F, TP-CT1-4F, TP-CT2-5H and TP-CT1-8B), PCR cycling conditions were 95 C for 5 min, followed by 55 cycles of 95 C for 30 s, 1 min at 50 C as annealing temperature and 1 min 30 s at 72 C, followed by a final extension stage at 72 C for 5 min.The remaining microsatellite locus (TP-CT2-5F) was amplified in the same PCR mix but with a slightly different thermocycling profile (35 cycles instead of 55 cycles).
All amplicons were separated on an Applied Biosystems 3730, and alleles were inferred manually using Peak Scanner software (Applied Biosystems).
The program GENALEX was used to assess genetic variation within sampling sites: number of alleles (N A ), observed heterozygosity (H O ) and expected heterozygosity (H E ; Peakall & Smouse, 2006).Allelic richness (A R ), rarefied to seven individuals, was assessed as a measure of the number of alleles independent of sample size with FSTAT 2.9.4 with n fixed as the smallest number of individuals typed for a locus and sampling site, hence allowing to compare this quantity between different sample sizes (Goudet, 1995).Genotype frequency conformance to Hardy-Weinberg expectations (HWE) was tested using the program Genepop (Raymond & Rousset, 1995;Rousset, 2008).Frequency of null alleles (NA, any allele at a microsatellite locus that consistently fails to amplify) was estimated with the program FreeNA (Chapuis & Estoup, 2006).F-statistics (Wright, 1978) with significance assessed following 3000 permutations in the software FSTAT 2.9.4 (Goudet, 1995).Global and pairwise F ST values were also recalculated using the null allele correction in FreeNA, over 10,000 replicates (Chapuis & Estoup, 2006).
The individual-based Bayesian clustering method implemented in STRUCTURE V2.3.4.(Pritchard et al., 2000) was used to estimate the most probable number of genetic groups within the data and assign probabilities of individuals to groups.STRUCTURE analysis was run with and without LOCPRIOR, with and without permitting admixture, with correlated allele frequencies and for a range of K values (1-7) with 20 runs for each value of K. Simulations were run over 500,000 Markov Chain Monte Carlo MCMC iterations with the first 100,000 repetitions discarded as burn-in.Structure K harvester (Earl & von Holdt, 2012) was used to determine the optimal value of K.The webtool CLUMPAK was used to cluster the runs and determine an outcome (Kopelman et al., 2015).Additionally, a principal component analysis (PCA) of microsatellite genotypes was performed using Adegenet package of R program (Jombart, 2010) to study the genetic relationships among samples without any preliminary assumptions.The program POWSIM (Ryman & Palm, 2006) was used to assess the statistical power when testing for genetic differentiation (Type I error, false significant results probability) by running the analysis for F ST = 0.000 and with the same sample sizes employed here.

Mitochondrial diversity within and between sampling sites
A partial region of the cytochrome oxidase I and II genes was success- Locations of haplotypes from both studies are presented in Table S2.
mtDNA structure among samples The map of haplotype frequency (Figure 3) represents the geographical repartition of the 17 haplotypes found in this study.The most common haplotype of the de novo collected samples (1A) was found in 59 European and UK individuals from all sampling sites except Belgium (Figures 3 and 4).Three other haplotypes (2H-2I, 3B and 3E) were also found in 10 or more individuals (10, 11 and 11, respectively).
While 3E was found in samples from both Europe and the UK, 2H-2I and 3B were only found in UK samples including Wales, England and Scotland.The remaining 13 haplotypes were found in fewer than 10 individuals (ranging from 1 to 6 individuals).Of these less common F I G U R E 2 Network of Tomicus piniperda mtDNA cytochrome oxidase I and II haplotypes found by the present study combined with those found by Horn et al. (2009).The disc is colour coded by location.The number of marks on the branch represents the number of mutations between haplotypes.Due to the slightly shorter sequence used here than by Horn et al., 2009 (763 vs. 797 bp), some haplotypes from Horn et al. (2009) were collapsed into a single haplotype.

Microsatellite diversity
In total, 5 loci were screened across 150 individuals and 6 sample locations.Fourteen out of 30 loci by site comparisons displayed significant (at p < 0.05) deviations from HWE of genotype frequencies (Table 4).
Estimates of null allele frequencies are presented in Table S3.Values across loci of sample size (N), mean number of allele (NA), allelic richness (A R ), mean observed heterozygosity (H O ) and mean expected heterozygosity (H E ) are presented in Table 5. Overall values of genetic differentiation (F ST ) across the sample set were significant across all loci (F ST without null allele correction = 0.047; F ST with null allele correction = 0.049).Pairwise multi-locus F ST were similar when estimated with and without null allele correction (range 0.000-0.100;

T. piniperda population history
In their Europe-wide phylogeographic study, Horn et al. (2009) reported a complex phylogeographic structure (36 haplotypes) wherein four genetic groups with varying distributions were identified Overall, the distribution of common and private haplotypes suggests that the UK samples exhibit a more recent co-ancestry with the Main-European group rather than the Central-European group.
A striking feature of both the microsatellite and mtDNA data was the generally higher levels of genetic variability reported for the Scottish population than for others.The high haplotype diversity in Scotland contrasts with the general model of southern richness and northern purity observed in north-temperate species colonizing postglacial northern Europe (Hewitt, 2000(Hewitt, , 2004)).Two non-mutually exclusive processes could generate the pattern observed here: (i) genetic erosion among the more southern populations studied here and/or (ii) colonization of Scotland from a northern refuge.Genetic erosion is unlikely the reason for the low genetic diversity found in Europe.First, haplotype and ND found in T. piniperda are similar to values reported for other bark beetle species in southern Europe (e.g., Ips typographus, Dendroctonus micans, Pityogenes chalcographus: Bertheau et al., 2012;Mayer et al., 2015), suggesting that there is no evidence of pervasive genetic loss in our most southern populations.
Second, although level of genetic erosion would be more prominent in mtDNA, no strong signal of genetic erosion was found in nuclear data either, with the genetic diversity of both European and UK samples being relatively high (Table 5).
Across multiple studies, a glacial refuge was identified for Scots pine (P.sylvestris) in northern Europe (Scandinavia, T oth et al., 2017).
T. piniperda has an obligate relationship with pine species and is thought to have shared glacial refugia with its host tree species (Horn et al., 2009).Therefore, a refugium in Scandinavia, in close proximity to and connected to Scotland could explain the high genetic diversity of T. piniperda found in Scotland.Alternatively, a northern glacial refuge for T. piniperda could have been in Ireland or Scotland, as suggested for Scots pine (Prus-Głowacki et al., 2012;Smout, 2014;T oth et al., 2017).Irish Scots pine is hypothesized to be the origin of the suggests that this might be a particularly important route for species to arrive in the UK, especially those exhibiting traits associated with greater dispersal potential via atmospheric events (Kazachkova et al., 2008;Minot & Husté, 2022;Siljamo et al., 2020).In this regard, the geographically incoherent genetic clustering of the two English The patchy genetic structure among the Scottish, Welsh and English samples indicates that T. piniperda might not disperse extensively between stands in the UK (at small spatial scales).This result is unexpected given that the species has a high dispersal potential and reported high gene flow in other areas (Faccoli et al., 2005;Forsse, 1989;Kerdelhué et al., 2006).The low Type I error rate revealed by POWSIM and congruence of patterns reported by group and individual-based analyses collectively indicate that this result should not be dismissed as statistical noise.The history of woodlands and forestry practices in the UK is an important consideration when comparing these results to other systems where an absence of genetic differentiation has been detected.Woodlands in the UK have experienced considerable loss and fragmentation through a long history of human activity (Bozzano et al., 2011;Watts, 2006) and Scots pine stands that are usually small and scattered (Hanson, 1937).This fragmentation could serve to restrict connectivity among stands to a greater extent than in other geographical regions and/or contribute to a metapopulation structure wherein connectivity is episodic.Distinguishing between population isolation and metapopulation structure would benefit from temporal replicate sampling as well as analysis of a greater number of loci through genome-wide Single nucleotide polymorphism (SNP) analysis (da Fonseca et al., 2016;Suchan et al., 2016).
Polymerase chain reactions (PCRs) were performed in 10 μL reaction volumes with 0.5 μmol of each primer, 5 μL of 2 X BioMix (Bioline, UK), 2 μL of ddH 2 O and 2 μL of template DNA.PCR cycling conditions were 95 C for 5 min, followed by 35 cycles of 95 C for 30 s, 1 min at 50 C as annealing temperature and 1 min 30 s at 72 C, followed by a final extension stage at 72 C for 5 min.PCR products obtained were purified using SureClean (Bioline, UK) and then processed by Sanger Sequencing with the forward and reverse primers using AB Big Dye technology on an Applied Biosystems 3730.All the electropherograms obtained were edited visually with Chromas and aligned using the CLUSTALW program implemented in BIOEDIT Number of polymorphic sites (S), haplotype diversity (or gene diversity, H) and nucleotide diversity (ND) was calculated for the different sampling sites.Genetic differentiation between pairs of sampling sites was assessed using Φ ST and exact tests of haplotype frequency homogeneity, with significance assessed by 10,000 permutations.Sequences were aligned with data from across Europe collected by Horn et al. (2009, see accession numbers in Table fully amplified and sequenced for 115 individuals from 7 populations (Scotland, Wales, South of England, North of England, Switzerland, Belgium and The Netherlands; Table2).After alignment of a 772-bp fragment, 16 polymorphic base positions and a total of 17 different haplotypes were found among the samples.The analysis revealed an overall haplotype diversity (H) of 0.712 (SD 0.042) and an overall ND of 0.003 (SD 0.004).Between 1 and 8, haplotypes were found per population.Overall, haplotype diversity (H) within locations ranged from 0.000 (Belgium) to 0.840 (Scotland).ND was lowest in Belgium (ND = 0.000) and highest in North of England (ND = 0.005).Correspondence between the 17 haplotypes found in the present study (discussed in the paragraph above) and the 36 haplotypes revealed byHorn et al. (2009) study of T. piniperda across continental European locations is presented in Figure2.Sequences were trimmed to 763 to be aligned with data fromHorn et al. (2009).This yielded a total of 44 haplotypes.Indeed, due to the slightly shorter sequence used here than byHorn et al., 2009 (763 vs. 797  bp, because of the low quality of some of our samples), some haplotypes fromHorn et al. (2009) were collapsed into a single haplotype.Specifically, the haplotypes 2H-2I and 1Z-2G were collapsed into single haplotypes.Although 7 of our 17 haplotypes corresponded withHorn et al. (2009) haplotypes, most of our haplotypes (10 of 17) had not been recorded previously.The most common haplotype (1A, n = 59) in our study was also the most common haplotype inHorn et al. (2009) and that was found to be distributed throughout Europe but found mostly outside the Iberian Peninsula and Italy.On the contrary, the second most common haplotype (1B) inHorn et al. (2009) was not found in our samples, and the second and third most common haplotypes (3B and 3E) in the present study were not found byHorn et al. (2009).
Figures S2 and 3.In the individual PCA, three first axes accounted for only 2% of all variation (Figure6).Axes one and three did not separate the individuals genotyped into cluster representing of sampling sites.On the other hand, the second PCA axis revealed a weak difference between Switzerland and the North of England versus the rest of the

F
I G U R E 5 STRUCTURE barplot showing the probability of each Tomicus piniperda individual assigning to either of the two genetic clusters (indicated by blue and orange).(a) With LOCPRIOR but without permitting admixture; (b) with LOCPRIOR and admixture.BEL, Belgium; nENG, North of England; NET, the Netherlands; SCO, Scotland; sENG, South of England; SWI, Switzerland; WAL, Wales.number of markers, replicated patterns could be detected in our data across several group-and individual-based methods.MtDNA analysis allowed direct integration with the data from Horn et al. (2009) and revealed a total of 44 haplotypes.Ten of these haplotypes were only found in our samples, 27 haplotypes were only found in Horn et al.(2009)  and the remaining 7 haplotypes were found in both datasets when combined.Simulation analysis revealed a low Type I error rate for the microsatellite data, which reported numerically small but statistically significant F ST between most sites.The nuclear data partitioned the samples into two groups in the Bayesian clustering analysis: one comprising the Scottish, Welsh and North-European samples (Belgium and the Netherlands) and the other comprising the two English and Swiss samples.A salient feature of both the mtDNA and microsatellite variation was the generally higher level of variation detected among the Scottish samples compared to those from more southern sites (England, Wales and Europe).Collectively, both marker sets add to evidence of high levels of genetic diversity and regional endemism in this species confirming the results obtained byHorn et al. (2009).
(i.e., Iberian, Pyrenean, Central-European and Main-European) and used to infer Pleistocene refugia locations and post-glacial colonization dynamics.Generally, haplotype diversity in the four European groups was comparable to the haplotype diversity found in the UK samples of the present study.The Iberian group inHorn et al. (2009) was characterized by numerous (n = 13) low frequency endemic Principal components analysis (PCA) of five microsatellite loci genotypes from European and British Tomicus piniperda.(a) Histogram of eigenvector; (b) plot of eigenvector 1 versus 2; (c) plot of eigenvector 1 versus 3; (d) plot of eigenvector 2 versus 3. haplotypes and a general absence of haplotypes found at high frequency elsewhere.Based on this,Horn et al. (2009) suggested that this Iberian refugial group is a southern edge relict that had not contributed to the recolonization of Northern Europe, with the Pyrenees acting as a barrier to dispersal and hindering migration events.Here, the UK samples reported both the predominance of haplotype 1A, which was very uncommon in the Iberian group and the general absence of most Iberian satellite haplotypes (1L, 1O, 1R, 1U, 1Y, 2B, 2C, 2D, 2J, 2K).These results support the prediction that the Iberian refuge/region has not been a contributor to the UK populations of T. piniperda.Horn et al. (2009) reported that the complex phylogeographic structure in Europe was shaped by the movement of individuals during interglacial periods leading to some haplotype sharing among groups emanating from different refugial areas at the end of the Last Glacial Maximum.They proposed that outside of Iberia and the Pyrenees, samples could be classified into two groups.First, a Central-European group spanning Germany, Poland, Russia (Kaliningrad) and an unusual Italian population.Second, a Main-European group covering all other sampling sites from the west (France) to the east (e.g., Russia, Estonia) and the North (Scandinavia-Norway, Sweden and Finland).Haplotype 1A was the most abundant in results fromHorn et al. (2009) and dominated the Main-European group.This haplotype was also the most abundant in our UK samples (n = 59).Additionally, haplotype 1A was only found at very low frequency in the Iberian and Central-European group and was not found at all in the Pyrenean group.Moreover, two haplotypes (1M, 1P) that were prevalent in the Central-European group were absent in our UK samples (with the exception of a single 1M found in Scotland), and two lower frequency haplotypes (1H and 1Q) typical of the Central-European group were also absent from the UK samples.
sites with the Swiss sample, while the remaining Scottish and Welsh samples clustered with Belgium and the Netherlands directs consideration as to how this pattern is generated and maintained.While the clustering of Scotland and Wales with Belgium and the Netherlands could be attributed to historical gene flow and potential genetic inertia, other processes are required to explain the genetic similarity among the England-Switzerland group and the differentiation of the English from the closer, North-European samples.Although not supported by the mtDNA data, the pattern is fully or partially detected by several individual and group-based analysis of the nuclear data and therefore cannot be dismissed as noise.Building on data from other studies, we propose that this may reflect a combination of factors.The study byKerdelhué et al. (2006) reported a sample from the west side of the Alps as genetically different from four out of six of the other sampled sites in France.Horn et al. (2009) uncovered a peculiar group of haplotypes in northern Italy (SON-IT), near the border with Switzerland, that was distinguished from the other nearby Italian and French samples.This specific group of haplotypes was clustered with the Central-European group (i.e., Germany, Poland and Kaliningrad) although being geographically separated from it.These findings add support to the view that Switzerland may harbour a genetically distinct population though this requires confirmation by more geographically extensive nuclear genotyping of the species across Europe as the haplotypes found in that locality are different from the haplotypes found in Switzerland in the present study.The similarity with the English group may therefore reflect the retention of ancestral polymorphism and complex colonization dynamics.Alternatively, the pattern may be generated by anthropogenic movement of individuals.In relation to the proposed anthropogenic transport between England and Switzerland, Germany, Poland and Russia are countries known for their production and exportation of wood and wood products (OEC, 2021).Consequently, anthropogenic movement of T. piniperda between Central-Europe, the UK and Switzerland may have occurred by international trade despite phytosanitary measures.While the genetic difference between England and north-western continental Europe may reflect anthropogenic introduction into England, contemporary dispersal and survival between these areas would be expected to homogenize population structures.Therefore, there might not be extensive dispersal of T. piniperda individuals across the English Channel.
Haplotype and nucleotide diversity for Tomicus piniperda populations.
3D, 3F, 3G, 3H, 3I and 3J) only found in UK individuals (Scotland, Wales and England).Exact tests revealed significant differences in haplotype frequencies between most populations.In contrast, Φ ST tests reported fewer significant values(i.e., between Scotland and   Wales, Switzerland and Wales and Switzerland and South of England)between populations with no clear correlation to any geographical configuration (Table3).T A B L E 2

Table 6
ST values all above 0.03.
). Patterns of pairwise F ST significance were also similar for the adjusted and unadjusted data (Table6).Scotland was significantly different from all other sampling sites but Belgium and the Netherlands with most pairwise F ST values being above 0.05.Wales was significantly different from Switzerland and Scotland (before ENA) in addition to South of England (after ENA) with most pairwise F was significantly different from North of England and Switzerland (before and after ENA, pairwise F ST values above 0.07).Finally, the Netherlands was different from England and Switzerland (before and after ENA, pairwise F ST values above 0.05).The Bayesian clustering method in STRUCTURE performed for the five loci combined and using LocPrior suggested a structure with Summary table for all five microsatellites loci.
T A B L E 4 p-Values from single locus-by-sample tests of conformance to Hardy-Weinberg equilibrium genotype proportions.Note: Values are across loci.Standard error given in parentheses.Abbreviations: AR, allelic richness; HE, mean expected heterozygosity; HO, mean observed heterozygosity; N, sample size; NA, mean number of alleles.T A B L E 6 Pairwise F ST values based on five microsatellite loci.Note: Statistically significant values in bold.Above diagonal are values obtained after correction for null alleles using the ENA method.Below diagonal are values from uncorrected data (not using ENA).