Key Message

  • We studied the population genetics structure of Vespa velutina in Mallorca.

  • We identified the same mitochondrial haplotype shared by other European populations.

  • Two independent introductions from two different European regions were detected.

  • Identifying invasive populations pathways may help for invasive species management.

Introduction

Social insects have high invasion capacities (Beggs et al. 2011) mainly due to the ability of single foundress to establish a huge colony and to produce many future queens which in turn offer a large dispersal capacity. Such colonization abilities can be further enhanced by global trade and human mobility (Mikheyev et al. 2009; Arca et al. 2015). The main ecological effects of invasions include the impact on biodiversity (Vilà et al. 2011) and ecosystem functioning (Pejchar and Mooney 2009), with island environments being more vulnerable than continental systems due to their lower complexity, high endemicity rates and less diversity levels (Traveset and Richardson 2006, 2014; Sax and Gaines 2009).

The yellow-legged hornet (Vespa velutina) displays unique biological and life history traits that increase the probability of success in establishing and spreading in new territories (Moller 1996), including excellent dispersal capacities, high reproductive rates, broad diets and habitat ranges, effective predator defenses, and superior competitive skills (Moller 1996; Beggs et al. 2011). Consistently, this species is known to be the first invasive Vespidae predator accidentally introduced from Asia to Europe (Monceau et al. 2014). It was detected in France in 2004 from where the species has successfully spread and stablished in neighboring countries (Laurino et al. 2020). It was first reported from Spain in 2010 (López et al. 2011) where their populations have colonized all north Iberian peninsula arriving to Mallorca island (Balearic Islands, Western Mediterranean, Spain) in 2015 (Leza et al. 2021) probably due to an accidental introduction by humans (Robinet et al. 2019).

The first report of V. velutina in Mallorca leaded to the implementation of an intensive control protocol that allowed the removal of 31 nests and the capture of 16 queens using spring trapping between 2015 and 2018 (Leza et al. 2021). After these interventions, no more nests were found and the species was declared officially eradicated in November 2020 (CAIB 2020) until July 2021 when a new secondary nest was detected and removed in the island. Since then, neither adults nor nest have been detected in the archipelago.

While there is a wealth of studies on different aspects of the biology, ecology and management of V. velutina in Europe, only a few of them have focused on the genetics of the invasion (Arca et al. 2015; Budge et al. 2017; Granato et al. 2019; Jones et al. 2020; Quaresma et al. 2022). These studies have evidenced that the origin of the European invasion by yellow-legged hornets represented a genetic bottleneck event that explains their low levels of genetic diversity, probably as the result of a single queen introduction from southeast China (Arca et al. 2015). Nonetheless, this has not meant any limitation for the species to expand throughout the European continent. Such invasive capacity was also observed in Mallorca, where both the number of reported nests and the colonized area gradually increased since its detection until its eradication, after a potential population bottleneck event (Leza et al. 2021).

In this study, we expand the current knowledge on the invasion genetics of V. velutina in Europe based on mitochondrial (cox1) and nuclear (STRs) DNA analyses of the invading populations in the westernmost Mediterranean archipelago.

Materials and methods

Sampling and DNA extraction

A total of 335 adults of V. velutina were collected in Spain: Asturias (N = 11), Basque Country (N = 20), Catalonia (N = 16), Extremadura (N = 2), Galicia (N = 12) and Mallorca (N = 274) (Fig. 1, Table S1). Among sampled populations, only two individuals could be obtained from Extremadura (Spain) being the first detection of this invasive species in Extremadura. This particular sample was however retained in the analyses to increase the coverage of hornet invasive range. Queen individuals were collected from traps, workers were obtained from both traps and nests, while males were obtained directly from nests. Samples were stored in absolute ethanol and/or at − 20ºC until molecular analysis.

Fig. 1
figure 1

Maps of the Iberian Peninsula and Mallorca showing the location of the Vespa velutina individuals sampled from nests (black dots) and from traps (red dots), respectively. Pink-shaded areas indicate the invaded territories in Spain according to Laurino et al. (2020). Maps were obtained with Google Maps (Map data 2020, Google) using the function ‘get_map’ in the package ‘ggmap’ version 3.0.0.902 in R version 3.6.3

Full size image

Genomic DNA was individually extracted and purified from insect legs using the Qiagen DNeasy Blood & Tissue kit (Qiagen) following the manufacturer’s instructions and RNA was removed using 60 μg of RNAse A (Promega).

Mitochondrial DNA sequencing and analysis

Genetic analyses based on mitochondrial cytochrome c oxidase subunit 1 (cox1) gene sequences were conducted on a sub-sample of 103 V. velutina individuals from Asturias (n = 10), Basque Country (n = 15), Catalonia (n = 15), Extremadura (n = 2), Galicia (n = 12), and Mallorca (n = 49). Such analyses included all individuals from mainland Spain, at least one individual from each nest detected in Mallorca and all trapped queens. Therefore, we cover all potential cox1 diversity present on the island. For each specimen, a cox1 gene fragment was PCR-amplified and sequenced using a custom designed V. velutina specific primer-pair Vespa_LCO (ATTCAACAAATCACAAAGATATTGG)/Vespa_HCO (TAAACTTCTGGATGTCCAAAGAATCA) targeting the standard DNA barcoding 658 bp region. Primers were designed based on available V. velutina complete mitochondrial genome sequences in GenBank (accession codes AP017943, AP018460, AP018461, AP018483, AP018484 and KY091645). PCR conditions used 0.2 μM of each primer and 3.5 mM MgCl2 (50 mM) using a standard protocol with an initial denaturation stage at 94 ºC (4 min), followed by 40 cycles with annealing temperature at 48 ºC (30 s), denaturation at 94 °C (30 s) and elongation at 72 °C (60 s), and a final extension step of 10 min. at 72 ºC. PCR products were checked for amplification by electrophoresis on 1% agarose gel stained with ethidium bromide and subsequently purified using the MSB Spin PCRapace kit (Invitek). Sanger sequencing was performed with the same primers as above using the BigDye Terminator Cycle Sequencing kit (Applied Biosystems). Electropherograms were assembled and edited with CodonCode Aligner (CodonCode Corporation) and their respective identity as V. velutina cox1 sequences was assessed by comparing them with those available in GenBank using the BLASTn online tool (https://blast.ncbi.nlm.nih.gov/).

Newly generated cox1 sequences were aligned using MAFFT 7 (Katoh et al. 2018) under default options with all available V. velutina cox1 sequences in GenBank. A median joining haplotype network was computed using the popART 1.7 software (Leigh and Bryant 2015). In addition, the same DNA alignment was also used to infer a maximum likelihood (ML) phylogenetic tree in IQ-TREE v 1.6.12 (Nguyen et al. 2015) by simultaneously searching for the best evolutionary model according to Bayesian Information Criterion (TN + F + G4) and specifying 1000 bootstrap replicates. The resulting tree was visualized with FigTree v1.4.4 (Rambaut 2018).

STR amplification and analysis

All sampled individuals (335) were genotyped for 15 microsatellite loci using oligo combinations developed by previous studies on V. velutina population genetics (Daly et al. 2002; Hasegawa and Takahashi 2002; Arca et al. 2012). Forward primers were labeled with 6-FAM (blue), VIC (green), NED (yellow) and PET (red) dyes and the PCR amplifications were segregated into two multiplex reactions (Table S2). PCR reactions were performed in a total volume of 10 µL including 1 µL of V. velutina template DNA, 5 µL of Multiplex PCR kit (Qiagen), 3.6 µL of primer mix (10 µM) and 0.4 µL of H2O. The PCR conditions consisted of an initial denaturation step at 95 ºC (15 min) followed by 40 cycles at 94 ºC (30 s), 50 ºC (90 s) and 72 ºC (60 s), followed by a final extension stage at 60 ºC (30 min). Fragment analysis was performed on an ABI Prism 3130 DNA Genetic Analyzer (Applied Biosystems, Foster, California, USA), using GeneScan 500LIZ® as internal size standard. Fragment lengths were determined using GENEMAPPER 5.0 (Applied Biosystems) and checked manually.

The microsatellite dataset generated from the 335 V. velutina individuals was merged with an existing dataset containing genotypes from 417 individuals collected in other invaded territories (mainland Spain: Basque Country = 3 and Galicia = 42, France = 83, Italy = 11, Portugal = 191, and Korea = 8) and also from the native range of the species (Chinese provinces of Yunnan = 20 and Zhejiang/Jiangsu = 30, Indonesia = 21, and Vietnam = 8) (Arca et al. 2015; Quaresma et al. 2022). To enable dataset merging, allele scores of each microsatellite locus were harmonized between laboratories by genotyping 10 DNA samples used by Quaresma et al. (2022), which in turn were previously harmonized with Arca et al. (2015). Although the studies by Budge et al. (2017) and Jones et al. (2020) also generated V. velutina STR data, we opted for the conservative approach of not including them in our analyses since they were not harmonized with other studies.

Observed number of alleles (Na), observed heterozygosity (Ho) and expected heterozygosity (He) were computed per locus and region using ARLEQUIN 3.5.2.2 (Excoffier et al. 2005); whereas, unbiased expected heterozygosity (uHe) was estimated using the formula uHe = He*(2*n_individuals/(2*n_individuals-1)) (Nei 1978) and allelic richness (Ar) was estimated using the R package pegas (Paradis 2010), using the rarefaction method described by Hulbert (1971). In addition, inbreeding coefficient (f) was calculated using the formula f = (uHe-Ho)/uHe, where positive values represent an excess of homozygotes and negative values represent excess of heterozygotes (Li and Horvitz 1953). For each population, the mean and standard deviation of the above-mentioned genetic parameters were calculated to enable comparisons. Genetic differentiation among populations from the invaded and native ranges was estimated based on the pairwise FST values computed in ARLEQUIN 3.5.2.2 and using 10,000 random permutations to assess significance.

Population structure and relationships

The population structure was analyzed using the fast maximum likelihood genetic clustering approach (Beugin et al. 2018). This approach is very similar to the model implemented by STRUCTURE (Pritchard et al. 2000) but allows for much faster estimation of genetic clusters thanks to the use of the Expectation-Maximization (EM) algorithm. We initially investigated the number of clusters by using the k-means algorithm, where the preferred number of clusters was evaluated using the Bayesian information criterion (BIC) score.

A discriminant analysis of principal components (DAPC) was also used to infer population structure, using the adegenet 2.1.6 R package (Jombart 2008). This multivariate analysis combines principal component analysis (PCA) and discriminant analysis to determine the number of genetic clusters in the sample.

The program GeneClass v.2.0 (Piry et al. 2004) was used to assign or to exclude reference populations from either the invaded or the native ranges as possible sources of the Mallorca individuals on the basis of multilocus genotypes by using the standard criterion described by Rannala and Mountain (1997). The Monte Carlo resampling method (Paetkau et al. 2004) was also applied to identify the accurate critical values of exclusion/inclusion by simulating 10,000 genotypes for each population with a threshold probability value of 0.05.

Bottleneck detection

The occurrence of recent genetic bottlenecks in Mallorca populations was explored with the test of heterozygosity excess implemented in BOTTLENECK 1.2.0.2 (Cornuet and Luikart 1996; Piry et al. 1999). Heterozygosity excess is expected in populations that have experienced a significant reduction in size because rare alleles are lost (Cornuet and Luikart 1996). Two mutation models were tested: the infinite allele model (IAM) and the two-phase model (TPM), the latter of which incorporates elements of the IAM and stepwise mutation model (SMM) (variance = 12, SMM = 95%, Piry et al. 1999). The Wilcoxon sign-rank test was used to test for a statistically significant bottleneck (Luikart and Cornuet 1998).

Mating statistics

Both queen and mate genotypes were inferred from worker offspring genetic data using Colony 2.0.1.1 (Wang and Santure 2009) based on the 15 STR loci from Mallorca population. When available, males were also genotyped and known queen alleles were inferred since the haploid genotypes of males are the direct product of queen meiosis. Four trials with assumed genotyping error rates of 0.001, 0.01, 0.05, and 0.1 were performed. The allele frequencies inferred from Mallorca population between 2015 and 2021 were used as input.

Results

Mitochondrial DNA sequencing and analysis

All specimens sampled from Mallorca and mainland Spain (Asturias, Basque Country, Catalonia, Extremadura and Galicia) shared the same cox1 haplotype (GenBank accession codes: ON094523 to ON094624), which was identical to the haplotype shared by all European V. velutina invaders sequenced to date and also by the specimens sampled in Jiangsu and Zhejiang (China) (Haplotype H1, Fig. 2). Moreover, our phylogenetic analysis also revealed a possible new haplotype from China (GenBank: MN716845) (Fig. 2). The possibility that the number of haplotypes is greater than that currently available in GenBank is consistent with the study by Perrard et al. (2014), which describe additional mtDNA haplotypes of V. velutina raising the total number to 25. However, these haplotypes are not available in GenBank and therefore have not been included in our analyses.

Fig. 2
figure 2

a Maximum Likelihood tree (log-likelihood =  − 1942.5559) computed using the TN + F + G4 evolutionary model selected by the ModelFinder algorithm (Kalyaanamoorthy et al. 2017) implemented in the IQ-TREE software. Node numbers represent ultrafast bootstrap support (only values > 75 are shown). Samples contributed by this study are marked with asterisk symbols. b Median joining network of V. velutina cox1 haplotypes inferred in PopART. The pie charts display haplotype frequencies in each sampled region with size proportional to the number of sequences included (see scale at the bottom left). Missing haplotypes and mutational steps are indicated with solid black circles and hatch marks, respectively

Full size image

The phylogenetic analyses based on cox1 sequences showed V. velutina divided into two main clades (Fig. 2a) in agreement with previous reports (Perrard et al. 2014; Takeuchi et al. 2017; Granato et al. 2019). The inferred tree supported the results derived from the network analysis (Fig. 2b) and showed the Spanish samples nested within the European invasion clade.

STR amplification and analysis

The mean (± SE) number of alleles (Na), allelic richness (Ar), and heterozygosities (Ho, uHe) computed for each sampled region of the invaded and native ranges in Europe and Asia are shown in Fig. 3.

Fig. 3
figure 3

Mean and standard error of number of alleles (Na), allelic richness (Ar), heterozygosities (Ho, uHe) and inbreeding coefficient (f) for each studied population from native and invaded range of V. velutina. Numbers following Mallorca codes refer to the year of collection (e.g., Mallorca 15 stands for samples collected in Mallorca in 2015)

Full size image

The 15 microsatellite loci represented 180 different alleles, of which 174 were present in Asia and 64 in Europe. The number of private alleles was 116 and 6 in the native range and Europe, respectively. All loci from Asia were polymorphic, whereas in mainland Europe two loci were monomorphic (D2-142 in Portugal, mainland Spain, Italy and France, and R1-77 in mainland Spain), five loci were monomorphic in Mallorca between 2015 and 2018 (D2-142, D2-185, R1-77 and R1-169) and three loci were monomorphic in Mallorca 2021 (R1-80, D2.142 and R1-77).

Population structure and relationships

There was considerable divergence among populations: Pairwise FST ranged from 0.0001 to 0.6366 and all values were statistically significant (p value < 0.05), except between Mallorca 2015 and 2016 which was not different from zero (p value = 0.43) (Fig. 4). Generally, higher FST values corresponded to comparisons between native and invasive populations (excepting the pair Korea—Zhejiang/Jiangsu). Regarding the Mallorca populations, all pairwise comparisons between samples from 2015 to 2018 yielded low FST values, in contrast with the higher FST values retrieved from the analyses between them and Mallorca 2021 (range: 0.4378–0.465). According to FST values, the closest European population to the group conformed by Mallorca 2015–2018 was the Italian one (range: 0.1791–0.3392), while Mallorca 2021 was retrieved as closely related to V. velutina individuals from Catalonia (0.1822).

Fig. 4
figure 4

Multi-locus estimates of pairwise FST genetic distances between V. velutina populations from both invaded and native ranges (lower diagonal) and between Mallorca populations and Mainland Spain populations (top right). Numbers following Mallorca codes refer to the year of collection (e.g., Mallorca 15 stands for samples collected in Mallorca in 2015)

Full size image

DAPC analyses clearly separated Asian and European populations (Fig. 5). Within the latter, the French populations were retrieved as the closest samples to the Asian ones (Fig. 5a). Consistent with FST estimates, Mallorca populations between 2015 and 2018 were located near to the Italian samples, and the Mallorca population 2021 near to V. velutina individuals from mainland Spain (Fig. 5b). The DAPC plot explained 93% of the total variance.

Fig. 5
figure 5

Scatterplot of the discriminant analysis of principal components (DAPC) of a Asian populations, France, and b magnified view of other European countries. Individuals genotypes are represented in dots, while populations are distinguished by colors and 95% inertia ellipses. The DAPC plot explained 93% of the total variance

Full size image

A pattern of genetic heterogeneity within the European range was detected by the Maximum Likelihood genetic clustering using the expectation–maximization algorithm implemented in the adegenet package (Fig. 6). At K = 2, there was a clear separation between native and France populations from the rest of the invasive populations in Europe. Further partitioning of variation (K from 3 to 5) revealed different genetic structures in Europe (Fig. 6). K = 5 showed that the invasion of Mallorca between 2015 and 2018 had a differentiated genetic structure (mean Q-value ± SE = 99.99 ± 0.0006%) that partially includes the Italian population (mean Q-value ± SE = 40.66 ± 12.47%). In the other hand, the Mallorca population from 2021 (mean Q-value ± SE = 99.99 ± 0.0000007%) clustered with samples from mainland Spain (mean Q-value ± SE = 67.22 ± 4.42%). In addition, Portuguese (mean Q-value ± SE = 97.87 ± 1.16%), French (mean Q-value ± SE = 99.99 ± 0.00003%) and Asian populations (mean Q-value ± SE = 99.99 ± 0.00007%) conformed three distinct clusters. Likewise, DAPC analyses identified the Asian, French, Atlantic (Portugal + Spain), Mediterranean (Spain + Italy) and Mallorca as differentiated genetic entities (Fig. 5).

Fig. 6
figure 6

Genetic structure of V. velutina. Populations admixture ancestry components are shown from K = 2 to K = 5 for the 752 individuals. The lowest Bayesian information criterion (BIC) was obtained at K = 5. Populations are delimited by black borders

Full size image

Notably, both DAPC (Fig. 5) and admixture (Fig. 6) analyses supported the genomic distinctiveness of Mallorca population between 2015 and 2018 from all other V. velutina populations.

One individual from Mallorca 2015 was detected by GENECLASS as first-generation migrant (p-value = 0.004) with a putative origin in Italy (log(L) =  − 6.804). Likewise, an individual from Mallorca 2017 was also detected as first-generation migrant (p-value < 0.001) with a putative origin located in mainland Spain (log(L) =  − 8.926).

Since FST values, DAPC and genetic structure analyses suggest two different introductions from mainland Europe in Mallorca, invasive populations sampled in the island were assigned by GENECLASS to two different regions: Mallorca samples between 2015 and 2018 were assigned to Italy (score = 100%), and those from 2021 were assigned to mainland Spain, specifically Catalonia (score = 100%).

Bottleneck detection

The decrease in genetic diversity observed in the invasive populations may be indicative of a bottleneck event (Arca et al. 2015). In this regard, the heterozygosity excess indicating a genetic bottleneck was tested in Mallorca populations between 2015 and 2021. The results were consistent with a bottleneck episode according to the significant value of heterozygosity excess estimated under both IAM and TPM models (p value < 0.05).

Mating statistics

Depending on the assumed genotyping error rate, the total inferred number of fathers varied from 1 to 10 over the Mallorca nests studied. On the one hand, all nest from 2015 to 2018 (first invasion) presented different patrilines resulting in a mean number of observed mattings per nest of 3.94 (range 2–6.5). On the other hand, a single patriline was detected in the only nest found in 2021 (second invasion). Diploid males were detected in two nests collected in 2017 and 2021, respectively.

Discussion

Our results are compatible with a scenario in which V. velutina populations in both mainland Spain and the Balearic archipelago could have derived from the spreading southward of the V. velutina population initially established in France, rather than from multiple independent introductions from the native range. This view is supported by the identification of a single mitochondrial haplotype shared by all our samples and those previously reported from the remainder of the invaded territories in Europe (Arca et al. 2015; Budge et al. 2017; Granato et al. 2019; Husemann et al. 2020; Jones et al. 2020; Quaresma et al. 2022). Furthermore, the European origin of the invasions of both mainland Spain and the Balearic archipelago is reinforced by nuclear STR data revealing the existence of a shared allelic pool and similar genetic diversity levels (Na and Ar) (Arca et al. 2015; Budge et al. 2017; Jones et al. 2020; Quaresma et al. 2022). Indeed, the FST values reported here indicate that Mallorca populations are closely related to representatives from Italy (samples Mallorca 2015–2018) and Catalonia (single nest found in Mallorca in 2021), whereas mainland Spain colonies show closer genetic affinities with those from Italy and Portugal.

The invasion of Mallorca Island by V. velutina is compatible with a bottleneck event. The estimated excess of observed heterozygosity (Ho) and reduced inbreeding coefficient (f) can lead to a decrease in terms of both adaptive potential and fitness of a founding population, therefore compromising its establishment and spread across new ranges (Willi et al. 2006). However, there exists evidence that V. velutina populations can establish successfully even after a severe genetic bottleneck resulting from the introduction of very few or even a single female (Arca et al. 2015). The case of the invasion of Mallorca reported here reinforce this view, since even after a potential second bottleneck the invasive population of V. velutina kept growing after the first detection despite eradication efforts (Leza et al. 2021). In fact, the increase in observed heterozygosity detected on consecutive populations between 2015 and 2018 could be explained by the control and eradication efforts applied in the island (Leza et al. 2021). The eradication plan focused on entire nest removal before male emergences, therefore preventing reproduction and lowering the genetic diversity levels of the population and increasing the bottleneck event and founder effect on the next generations.

Genetic diversity of founders is critical for the adaptative response to new environments (Lockwood et al. 2005; Drake and Lodge 2006). In this regard, the polyandrous mating system of V. velutina allowed the introduction of an important fraction of the genetic diversity of the native population into Europe (Arca et al. 2015). Here we contribute additional genetic evidence of the polyandrous behavior of this invasive species based on the inference of up to 10 potential males fecundating a single queen. High allelic diversity in the founding individuals coupled with other extraordinary life history traits common to social Hymenoptera (Moller 1996; Beggs et al. 2011) probably helped spread of this invasive species after a new bottleneck event in Mallorca. According to our genetic results and ecological niche models developed specifically for the island (Herrera et al. 2023), if an early detection and eradication plan would not have been implemented (Leza et al. 2021) the species probably would have colonized and stablished in part of the island, as predicted by Robinet et al. (2019).

Bottleneck events, inbreeding, limited gene flow and genetic drift reduce sex allele diversity (Darrouzet et al. 2015). As well as other Hymenoptera species, V. velutina sex is determined by a multi-allelic locus in which females are diploid (heterozygous genotype) and males are haploid (hemizygous genotype) (Heimpel and De Boer 2008). When allelic diversity is low at the sex locus, there is a higher chance that the queen will produce homozygous eggs, which will develop into diploid males (Darrouzet et al. 2015; Quaresma et al. 2022) that are usually sterile and do not contribute to colony tasks. In this study we report two nests hosting diploid males (7.69% of all sampled nests) after a bottleneck event. This value is considerably lower than the ratios reported by studies conducted in United Kingdom and France, where diploid males represented 75 and 48% of all nests genotyped, respectively (Darrouzet et al. 2015; Budge et al. 2017).

Our case study represents the first time that the yellow-legged hornet reaches a Mediterranean island after its arrival to Europe in 2004 (Monceau et al. 2014). In fact, the genetic results presented in this study shows that this species has reached Mallorca in two different moments from two different European regions, respectively: Italy (2015) and Mainland Spain (2021). Therefore, all nests identified and eliminated in Mallorca during the eradication efforts can be attributed to a single introduction that was initially detected in 2015. Fortunately, thanks to an observant citizen, the second introduction in 2021 was swiftly detected, enabling the competent administration to promptly locate the nest and sent it to the University of the Balearic Islands for subsequent analyses. Given that the nest hosted male specimens, two possible scenarios were confronted: There was already a reproductive caste established indicating that the species was already dispersing, or the male specimens were sterile. Each of these scenarios could have distinct implications for the management tasks associated with controlling this invasive species. Consequently, a genetic analysis was conducted on male hornets using 15 STRs microsatellite markers revealing that all specimens were diploid, and therefore sterile (Heimpel and De Boer 2008; Darrouzet et al. 2015). The employed methodology provided valuable insights into the potential dispersal of this species, enabling us to predict whether the competent administration should be prepared to detect and eradicate new nests in the upcoming years in the way described by (Leza et al. 2021).

Robinet et al. (2019) explored the human-mediated dispersal of V. velutina concluding that the Mediterranean islands could not be naturally colonized, and only an anthropogenic introduction could favor the colonization of such territories. Mallorca island is 176 km away from the mainland nearest point, so the most probably introduction could be through shipping traffic. The genetic structure analyses performed here support the pathways of this species and assign French populations to a cluster that is more closely related to Asian populations than to the rest of European populations, in contrast with other genetic studies (Arca et al. 2015; Budge et al. 2017; Quaresma et al. 2022). Interestingly, the individuals sampled in Mallorca clustered into a well differentiated position from the rest of European or Asian clusters, which is consistent with the bottleneck event inferred by our analyses. This result may indicate that the diversity loss of invasive populations of V. velutina may vary across geographic range as documented for numerous animal and plant invasions (Forsman 2014).

Our findings reveal the pathways used for this species to reach new isolated territories as the Mediterranean islands. Such a threat situation requires the development of effective management and prevention plans, which can greatly benefit from knowing both the origin and the genetic structure of the invading populations.

Author’s contributions

CH, ML and JAJ-R conceived the study; CH and ML provided Vespa velutina samples; CH, JF and JAJ-R designed the experiments; CH, JF and JAJ-R performed the experiments and analyzed the data; CH and JAJ-R performed phylogenetic analyses, analyzed the data and created the figures; CH wrote the original draft, and all authors commented and contributed to the final manuscript. All authors read and approved the final manuscript.