Understanding the ecological and evolutionary forces that determine the genetic structure and spread of invasive species is a key component of invasion biology. The bay barnacle, Balanus improvisus (= Amphibalanus improvisus), is one of the most successful aquatic invaders worldwide, and is characterised by broad environmental tolerance. Although the species can spread through natural larval dispersal, human-mediated transport through (primarily) shipping has almost certainly contributed to the current global distribution of this species. Despite its worldwide distribution, little is known about the phylogeography of this species. Here, we characterize the population genetic structure and model dispersal dynamics of the barnacle B. improvisus, and describe how human-mediated spreading via shipping as well as natural larval dispersal may have contributed to observed genetic variation. We used both mitochondrial DNA (cytochrome c oxidase subunit I: COI) and nuclear microsatellites to characterize the genetic structure in 14 populations of B. improvisus on a global and regional scale (Baltic Sea). Genetic diversity was high in most populations, and many haplotypes were shared among populations on a global scale, indicating that long-distance dispersal (presumably through shipping and other anthropogenic activities) has played an important role in shaping the population genetic structure of this cosmopolitan species. We could not clearly confirm prior claims that B. improvisus originates from the western margins of the Atlantic coasts; although there were indications that Argentina could be part of a native region. In addition to dispersal via shipping, we show that natural larval dispersal may play an important role for further colonisation following initial introduction.
Citation: Wrange A-L, Charrier G, Thonig A, Alm Rosenblad M, Blomberg A, Havenhand JN, et al. (2016) The Story of a Hitchhiker: Population Genetic Patterns in the Invasive Barnacle Balanus(Amphibalanus) improvisus Darwin 1854. PLoS ONE 11(1): e0147082. https://doi.org/10.1371/journal.pone.0147082
Editor: Yong-Gang Yao, Kunming Institute of Zoology, Chinese Academy of Sciences, CHINA
Received: May 27, 2015; Accepted: December 27, 2015; Published: January 28, 2016
Copyright: © 2016 Wrange et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data from our study (COI sequences for all individuals and genotype lists based on the microsatellite loci) have been uploaded on the Dryad database webpage (https://datadryad.org/). Accession number: doi:10.5061/dryad.sf2v0. All data can also be accessed directly from the authors.
Funding: This work was performed within Centre for Marine Evolutionary Biology (http://www.cemeb.science.gu.se), which is supported by a Linnaeus grant from the Swedish Research Councils, VR (www.vr.se) and FORMAS (www.formas.se), as well as by Wilhelm and Martina Lundgren Foundation and Helge Ax:son Johnson Foundation. MAR was partly funded by BILS (www.bils.se). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The number of non-native species being introduced to coastal and estuarine habitats has increased dramatically during the past century; largely due to anthropogenic activities including increased shipping intensity and aquaculture trading [1,2]. In many cases, these introductions have resulted in major disturbances of native ecosystems as well as economic losses, and biological invasions are today considered a major contributor to global environmental change and a threat to biodiversity [3–5]. Knowledge of the genetic structure and dispersal dynamics of non-indigenous species is important not only for management of invasive species, but also for improving our understanding of ecological and evolutionary processes related to colonisation of new environments, e.g. the role of genetic diversity, life history traits, environmental tolerance and gene flow via larval dispersal, for successful invasions [6–9].
Dispersal dynamics in the marine realm are generally complex and not well understood . Dispersal of free-drifting larvae depends not only on oceanographic conditions, but also on life history characteristics of the individual species, including larval form, swimming ability, nutritional modes and reproductive timing [11,12]. In general, species with the capacity to disperse over long distances are traditionally expected to have weak population structure since a few migrants may be sufficient to homogenise local genetic variation [13,14]. However, multiple studies have shown surprisingly large genetic differentiation over small geographical scales in a number of marine species with pelagic larval dispersal [15–20]. Genetic differentiation between populations despite high gene flow can e.g. be the result of non-random settlement or selection through post-settlement mortality [21–23]. In addition to natural larval dispersal, human-mediated translocation often increases the complexity of the observed genetic patterns [24–26].
Genetic markers are typically used in tracing origins of populations as well as understanding population structure and dispersal mechanisms [25,27–29]. Sequence-based mitochondrial DNA (mtDNA) has commonly been used to infer demographical history, e.g. founder events and demographic expansions associated with post-glacial colonisation (e.g. [30,31]). Frequency-based markers such as microsatellites and AFLPs, on the other hand, are often used to detect more recent and fine-scale differentiation between populations [32–34] and for studying how ecological factors influence contemporary partitioning of genetic variation . However, genetic structure does not always mirror gene flow, since many populations are not in migration-drift equilibrium [16,36]. Genetic structure is the result of contemporary levels of gene flow superimposed on the traces of phylogeographical and demographical history. Nevertheless, genetic tools can provide relevant baseline information about population structure and potential for local adaptations to evolve. Furthermore, combining genetic data with oceanographic modelling has provided new interesting ways to detect and understand gene flow between populations [37–40].
The acorn barnacle Balanus improvisus (= Amphibalanus improvisus) has invaded shallow coastal environments worldwide and is particularly common in estuaries and low salinity environments where it often forms dense populations [41–43]. There are several factors contributing to the successful introduction and spread of this species. First, B. improvisus is a major fouling organism, attaching to hard surfaces such as ship hulls, thus being transported efficiently over large distances [44–46]. Second, the species has a broad tolerance to many abiotic factors, including salinity [47,48], temperature and pH . Broad environmental tolerance is a common trait for many successful invaders . Finally, as in other barnacles, B. improvisus produces free-swimming larvae, which remain planktonic for several weeks , during which they disperse with oceanographic currents.
Despite being a major fouling organism invading coastal areas worldwide, little is known about the origin, colonisation history and present population genetic structure of this species. B. improvisus is commonly described as originating from America’s Atlantic coast [41,52], however no clear evidence supporting this statement is available. Charles Darwin, who first described the species, recorded populations from the north- and southeast coasts of America, but also from several localities in Europe, including the UK . B. improvisus was first recorded along the Pacific coast of the US in 1853, possibly after human introduction via shipping during the gold rush or oyster imports . In the western Pacific, the first recordings of the species from Japan were not made until the 1950’s . Many of the early recordings of B. improvisus from around the world were made before or around the time that the species was first described by Darwin [54,56,57], leading to confusion about precisely which species were recorded, especially since several other acorn barnacles have similar morphology (e.g. Balanus eburneus, B. crenatus, B. subalbidus) . The timing of these early observations also makes it difficult to estimate historical colonisation patterns based on first observations in the literature.
Although B. improvisus is a highly euryhaline species, it performs best in low salinity [48,58] and has invaded several large brackish seas. It was first observed in the Black Sea in 1899  and when the Volga-Don Canal opened in 1952, it penetrated into the Caspian Sea and became locally dominant within three years . In the Baltic Sea, B. improvisus was first recorded in 1844 from Königsberg (today Kaliningrad)—ten years before the species was officially described by Darwin in 1854 [53,57]. From this putative dispersal centre in the southern Baltic it spread very rapidly and became common, especially in ports [43,56]. There are few records of this species from the Swedish east coast before the 1920’s but by the 1990’s, B. improvisus reached its northernmost limit of distribution at the Northern Quark, at 63° N . The semi-enclosed Baltic Sea is characterized by a steep salinity gradient and has limited connectivity with the adjacent North Sea/Atlantic, which has resulted in both genetic and phenotypic differentiation in many species–and in some cases local adaptations–in the Baltic environment [61,62]. Despite being a common, microhabitat-forming species in the Baltic Sea [43,52], there is limited knowledge about the dispersal dynamics and genetic structure of B. improvisus in the Baltic Sea. Furman [63,64] used allozymes to study population differences in the Baltic Sea but found weak differentiation between populations and no clear geographical structuring (see also ).
Using mitochondrial and microsatellite DNA markers, we addressed the following questions: i) What role might natural larval dispersal versus human-mediated dispersal via shipping play in shaping the population structure? ii) Can the invasive history of B. improvisus be clarified despite multiple dispersal modes and potentially high gene flow?
No specific permits were required for sampling. No specific permissions were required for the locations/activities in this study. The sampling locations are not privately owned or protected in any way. This study did not involve endangered or protected species.
Sample collection and DNA extraction
A total of 454 individuals of the acorn barnacle Balanus improvisus were collected during 2008–2013. The 14 sampling locations were: Coos Bay (Oregon, Pacific coast of USA); Chesapeake Bay (Atlantic USA); North Carolina (Atlantic USA); Mar Chiquita Lagoon (Argentina); Seine estuary in the eastern English Channel (France); Sozopol, Black Sea (Bulgaria); western Caspian Sea (Iran); Tokyo Harbour (Japan) as well as six locations in the Skagerrak-Baltic Sea area; Saltö (Sweden) in the Skagerrak sea, Kiel fjord (Germany), Muuga Bay (Estonia) and Torhamn, Öregrund and Umeå on the Swedish east coast (Table 1, S1 Table). The samples were collected by removing juvenile or adult individuals of B. improvisus from rocks and other hard surfaces that were present in the intertidal/shallow area at each sample site. No samples were taken from boat hulls. The specimens were preserved in 96% ethanol or on FTA cards (Whatman) and stored at -20°C. The morphology of the barnacles in ethanol was checked under a stereomicroscope to confirm the species, according to . DNA extraction was performed using the Nucleospin Tissue Kit (Macherey Nagel) following either the protocol for tissue or FTA cards. Unless the specimens were very small, only the cirri (legs) of the barnacles were used. The DNA concentration and quality was analysed with a Nanodrop (Thermo Scientific). Except for the populations from the eastern USA (NC and CB), around 30 individuals per locality (N = 24–57) were extracted for the subsequent PCR amplifications (Table 1).
Genotyping of mitochondrial COI
Primers for the PCR amplification of the mitochondrial cytochrome c oxidase subunit I gene (COI) were designed based on mitochondrial sequences from the Balanus sequencing project currently carried out within the Linneaus Centre for Marine Evolutionary Biology at University of Gothenburg (www.cemeb.science.gu.se/research/imago-marine-genome-projects). For the PCR, extracted DNA was either diluted to concentrations of 10 ng/μL or used without dilution when the concentration was below 20 ng/μL. PCR reactions were performed in 50μL volume, containing 5 μL of template, 1X RBC buffer with 1.5 mM MgCl2, 1.5μM dNTPs, 0.13 U RBC Taq polymerase (RBC Bioscience) and 0.2 μM each of two of the primers (BiCOI-F-ext: 5'-TCTGAAACTCTTACTTTTGACCG-3', BiCOI-R-ext: 5'-CATTACCTGTTTTAGCTGGTGC-3'; Integrated DNA Technologies). The PCR program consisted of an initial pre-cycle of 3 min at 94°C, 2 min at 55°C and 1 min at 72°C, which was followed by 35 cycles of 30 s at 94°C, 30 s at 55°C and 1 min at 72°C, and finally an extension at 72°C for 7 min. Afterwards, 3.5 μL of the PCR products were visualized on a 1.5% agarose gel stained with GelRed (Invitrogen). This amplification resulted in 850 bp long fragments. However, for the specimens from the eastern USA, Öregrund, and partly Japan and the Black Sea, a nested PCR was necessary using the external primer pair in the first step and an internal primer pair (BiCOI-F int: 5'-TTGATGATACTACATTTCAGGCAG-3', BiCOI-R int: 5'- GATAATCCGAATATCGACGAGG-3') in the second step. This in turn, resulted in a 738 bp long fragment. The PCR products were purified using the Cycle Pure Kit (Omega Biotek), and sequenced by Macrogen Europe, in both directions using the internal primer pair above. The sequences have been deposited in the Dryad Database (https://datadryad.org/) with the accession number DOI: 10.5061/dryad.sf2v0.
Development and genotyping of microsatellites
Three different development approaches were used to obtain microsatellites for Balanus improvisus. First, microsatellites were isolated by the company Genetic Identification Service Inc. (GIS Inc., Chatsworth, USA; http://genetic-id-services.com), which screened four libraries enriched for tetranucleotide (AAAC), (CATC), (TACA), and (TAGA) motifs, following their proprietary protocol. Out of the 69 unique microsatellite-containing clones that were obtained from GIS, primer pairs were successfully designed in 36 clones with a minimum of seven di- or tetranucleotide repeats using PRIMER3  implemented in GENEIOUS PRO v.5.1.7 (; Biomatters Ltd). The primer design was performed according to the following parameters: optimal primer length = 20 bp (range = 18–25 bp), optimal GC content = 50% (range = 30–80%), optimal Tm = 60°C (range = 59–61°C) with a Tm difference not larger than 1°C between the forward and reverse primers from the same pair and product length = 100–400 bp. As a second development approach, microsatellites were identified with Tandem Repeat Finder  from an EST (expressed sequence tags) library, containing both Sanger and 454 sequences from B. improvisus cyprid larvae (unpublished: www.cemeb.science.gu.se/research/imago-marine-genome-projects). Out of 593 microsatellites identified with 2–6 bp motifs, 53 unique di-, tri- and tetranucleotide microsatellites were selected for primer design (>7 repeats, flanking regions >40 bp, perfection >87%), and primers pairs were successfully designed for 36 loci with PRIMER3 implemented in GENEIOUS PRO, following the same parameters as described above. Finally, the third microsatellite development approach involved the identification of microsatellite loci in a B. improvisus (a single individual from Saltö, Sweden) genomic assembly (500 Mbases) made from two 150 bp and 300 bp libraries containing a total of 83 million paired-end (2x100bp) Illumina reads (unpublished: www.cemeb.science.gu.se/research/imago-marine-genome-projects). In this library, 3916 di-, tri- and tetranuclotide microsatellites with a minimum of seven repeats and flanking regions >100 bp were identified with REPEATMASKER 3.3 , and 128 tri- and tetra-nucleotide microsatellites with 16–50 repeats were selected for primer design. Finally, primers were successfully designed for 36 loci with PRIMER3 implemented in GENEIOUS PRO, with the following parameters: optimal primer length = 20 bp (range = 18–25 bp), optimal GC content = 50% (range = 30–80%), optimal Tm = 60°C (range = 58–62°C) with a Tm difference not larger than 2°C between the forward and reverse primers from the same pair and optimal product length = 200 bp (range = 100–300 bp). All the primers were subsequently checked manually. A ‘universal tail’ was added to the 5’-end of each forward primer . Three distinct ‘universal tails’ were used (tail1, tail2 or tail4), as described by Real et al. . A PIG-tail (50-GTTTCTT) was also added to the 5’-end of each reverse primer to avoid ‘plus-A’ PCR artefacts . Some microsatellites were compared with and without tails, with no difference found in amplification success between the two methods. Microsatellite amplifications were carried out in 10-μL reaction volumes, containing 1 μL of template DNA (10–30 ng/μL), 1X PCR buffer (Mg2+ free), 0.2 mM of each dNTP, 1.5–2.5 mM MgCl2 (S2 Table), 0.025 U of RBC Taq polymerase (RBC Bioscience), 0.125 μM of reverse primer, 0.0125 μM of tailed forward primer (Integrated DNA Technologies), 0.125 μM of “universal” primer end labelled with WellRED dye (D2, D3 or D4, Sigma-Aldrich). A touchdown procedure was incorporated in the thermal cycling regime to increase the stringency of the PCRs. The PCR cycling conditions were as follows (alternative step lengths in italic): 94°C for 3 min, Ta1 for 2 min, 72°C for 1 min; (94°C for 30 s, Ta1–1 C for 30(40) s [-1°C per cycle until Ta2], 72°C for 1 min) x 11 cycles; (94°C for 30 s, Ta2 for 30(40) s, 72°C for 1 min) x 23 cycles; 72°C for 5(7/10) min. Details of the PCR protocol, i.e. MgCl2 concentration, number of PCR cycles and type of ‘universal’ tail, are provided in S2 Table. Sizing of PCR products was performed on a Beckman Coulter’s CEQ 8000 automated sequencer where all lanes included a 400-bp ladder. Allele sizes were scored with the software CEQ 8000 Genetic Analysis System (version 8.0.52). In spite of multiple PCR adjustments and primer re-designs, only four microsatellites out of the 108 loci tested were successfully amplified, produced consistent and clearly defined bands and displayed polymorphism: two loci from the GIS library (tetranucleotides B4 and C103), one locus from the EST library (trinucleotide E29), and one from the genomic library (tetranucleotide E42). Genotype lists for these four microsatellite loci have been deposited in the Dryad Database (https://datadryad.org/) with the accession number DOI: 10.5061/dryad.sf2v0).
Analysis of COI sequences
The COI sequences were edited and aligned using GENEIOUS PRO 6.1.6 (http://www.geneious.com, ). Primers and low quality sequence ends were trimmed, and all polymorphic sites were checked by eye on the chromatograms. Thereafter, each forward sequence was aligned with its respective reverse sequence to generate a consensus sequence of 694 bp for every individual. The resulting consensus sequences were aligned again and subsequently used for further analysis.
Molecular diversity indices including haplotype diversity (h, ), nucleotide diversity (π; ), number of polymorphic sites and haplotypes were obtained using ARLEQUIN v.3.5 . The appropriate nucleotide substitution model and the gamma distribution shape parameter for rate heterogeneity among sites were assessed with jMODELTEST2 [76,77]. Results obtained from JMODELTEST2 showed that the best substitution model among those proposed in ARLEQUIN was the Tamura and Nei model with a gamma correction (α = 0.023) for heterogeneity of mutation rates (4th best model choice according to AIC and BIC criteria).
Genealogical relationships among COI haplotypes were examined by reconstructing a haplotype network following the parsimony method of Templeton et al.  implemented in TCS v. 1.21 . However, this network was very unclear since it displayed numerous multiple connections (loops). Therefore, a minimum spanning tree was built in HAPSTAR  to indicate the most likely connections among haplotypes. Thereby all the loops that were observed in TCS were resolved by selecting the most likely connection and deleting other alternative connections, thus converting the parsimony network into a minimum spanning tree. In addition, the phylogenetic relationships of haplotypes were explored with a maximum likelihood phylogenetic tree that was built with PHYML . The HKY substitution model  with a gamma correction α = 1.015 and a proportion of invariant sites p-inv. = 0.691 was selected as the optimal model among those available in PHYML, according to JMODELTEST2. The robustness of branches was tested with 100 bootstraps, and two COI sequences of Balanus eburneus were used as an out-group for this maximum likelihood analysis.
We estimated pair-wise genetic differentiation between populations using the fixation index ΦST  obtained from ARLEQUIN, which includes information on both haplotype frequencies as well as information on molecular distance (using the Tamura-Nei model with gamma correction α = 0.023). Significance was tested with 10,000 permutations and False Discovery Rate (FDR) was calculated to take into account multiple comparisons . Pairwise ΦST were further linearized following  as implemented in ARLEQUIN v.22.214.171.124 and the paired genetic distances between populations (matrix) were visualized using non-metric multidimensional scaling (MDS) performed with metaMDS in the vegan package in R (R Development Core Team, 2006; http://www.R-project.org). A hierarchical analysis of molecular variance (AMOVA)  using the Tamura-Nei model with gamma correction (α = 0.023) was performed using ARLEQUIN to further explore the distribution of the genetic variability (in mtDNA) between groupings in the Baltic Sea that were identified in the MDS. The significance of the observed variances for each hierarchical comparison was tested by 10,000 permutations. The historical demography of B. improvisus was investigated using a mismatch distribution analysis [85,86]. The Harpending’s raggedness (R) test  was used to determine if the observed mismatch distribution was taken from an expanding or stationary population. Tajima’s D  and Fu’s FS  that test for departure from population equilibrium were calculated using ARLEQUIN, and significance was evaluated based on 10,000 simulations. Significance level for Fu’s FS test was set to 0.02 according to recommendation in the ARLEQUIN 3.5 manual.
Analysis of microsatellite data
The software MICRO-CHECKER  was used to investigate the presence of null alleles, stutter and large allele dropout. Since potential null alleles were identified in three out of four loci (B4, E24, E42) the software FreeNA  was used to estimate and correct for null allele frequencies applying the ENA correction method. GENEPOP v. 4.2  was used to calculate number of alleles (NA), observed (HO) and expected (HE) heterozygosity, and to test for deviations from Hardy–Weinberg equilibrium and linkage disequilibrium. FIS was estimated and tested using the probability tests within GENEPOP. Microsatellite allelic richness (Ar) was calculated using the rarefaction method in FSTAT . Population differentiation among samples was estimated using the FST estimator θ , and in FreeNA using ENA correction for null alleles. Statistical significance was calculated for uncorrected data using permutation tests implemented in FSTAT, and False Discovery Rate (FDR) was used to account for multiple comparisons . For null allele corrected data, statistical significance could not be estimated since data contained less than five loci . The BOTTLENECK software (version 1.2.02; ) was applied to investigate population declines, using the TPM model with 70–90% stepwise mutations, a variance of 10–30, and based on 10000 interactions. The Wilcoxon one-tailed test was used to check for significant heterozygosity excess.
Inference of introduction history using ABC modeling
The introduction history of B. improvisus was explored with approximate Bayesian computation analyses (ABC). ABC analyses of both the four microsatellite markers and mitochondrial sequences of the COI locus were analysed with the program DIYABC v2.1.0 (www1.montpellier.inra.fr/CBGP/diyabc/), described in [96, 97]. We tested four global scenarios of sequential introductions where we contrasted likely scenarios based on known and assumed introductions with possible alternative scenarios. Finally, we tested two more local scenarios for the invasion of B. improvisus in the Baltic Sea. The tested scenarios are shown in S1 Fig. Briefly, scenario 1 represents an introductory history where the east coast of the Americas (here represented by Argentina) is assumed to be the native region for B. improvisus and with one introductory branch to the Pacific coast of America, Japan, the Black Sea and the Caspian Sea, and another introductory branch to Europe (France) and later into the Baltic Sea (Kiel, Torhamn, Estonia, Öregrund and Umeå samples). Scenario 2 instead, assumes the west coast of the Americas to be the native region with one introductory branch to Japan, the Black Sea and the Caspian Sea, and another sequence to the east coast of the Americas then to Europe and into the Baltic Sea. Scenario 3 assumes an early introduction to Europe and subsequent dispersal into the Baltic Sea and later to the Black and the Caspian Seas. Finally, Scenario 4 assumes that B. improvisus originates from Europe. We also simulated two scenarios for the invasion into the Baltic Sea (S1E and S1F Fig). Scenario 1 assumes that the Baltic Sea was colonized from the south to the north through the North Sea, while scenario 2 assumes that the northern part of the Baltic Sea has a Ponto-Caspian origin, as is the case for some introduced species in the Baltic Sea . We here grouped the Kiel, Torhamn and Estonia samples into the group “Baltic Sea south” (BA-S) and the Öregrund and Umeå samples into “Baltic Sea north” (BA-N).
Since there is little information about present and historical demographic parameters, we sampled priors from a wide range of uniform distributions. Effective population sizes for all populations ranged between 10,000 and 100,000 for simulations of COI and between 10,000 and 200,000 for simulations of microsatellites. For the COI we used the mutation model by Tamura-Nei and sampled from a uniform distribution of mean mutation rates between 10−8 and 10−6. For the microsatellite loci the Generalized Stepwise Mutation model was used  with a uniform distribution of mean mutation rates between 10−4 and 10−3. Times of new introductions (population splits) ranged between 10–10,000 generations for all introductions. The genetic variation within and between populations was calculated using a set of summary statistics. For microsatellites we used mean number of alleles and expected heterozygosity within populations, and the pairwise comparisons FST and dμ2 between populations. For the COI sequences we used mean number of haplotypes and mean of pairwise differences within populations, and FST and mean number of pairwise differences between populations. We simulated between 106 to 4∙106 data sets per scenario. The distributions of summary statistics based on the sampled priors were reasonably close to observed data (S2 Fig). The posterior probabilities of the tested scenarios was estimated with a polychotomic weighted logistic regression on the 1% of simulated data sets most similar to the observed data with the proportion of the target scenario as the dependent variable and the differences between observed and simulated data set summary statistics as the independent variables.
Correlating genetic structure to geographic distance and oceanographic connectivity
To test the hypothesis of isolation-by-distance, correlation between pairwise linearized ΦST for COI and FST (with and without null allele correction) for microsatellites , versus geographical distances between locations was analysed using a Mantel test  implemented in R with the function mantel.test in the ncf package (tested with 1000 permutations). Geographical distances were based on the shortest shipping routes between the closest major ports to the sampling sites (calculated at: www.sea-distances.com) and distances obtained in nautical miles were converted into km.
We also tested if genetic differentiation between populations was correlated with oceanographic connectivity in the Baltic Sea. The connectivity between sites was estimated using a biophysical model, where velocity fields from an ocean circulation model were combined with a Lagrangian particle-tracking routine to simulate drift trajectories representing the dispersal of barnacle larvae. Velocity fields were first produced with the BaltiX model, which is a regional Baltic-North Sea configuration of the NEMO ocean model . The spatial resolution is 3.7 km in the horizontal and 56 depth intervals in the vertical ranging from 3 to 22 m. The model has a free surface and allows the grid boxes to stretch and shrink vertically to accurately model the tides without generating empty grid cells at low tide. For a more elaborate description and some validation results see . Dispersal of virtual larvae was simulated with the Lagrangian model TRACMASS , which produced multiple dispersal trajectories using the velocity field data from the BaltiX circulation model. Velocity fields were updated for all grid boxes in the model domain every three hours while the trajectory calculations were done with a 15-minute time step. To mimic expected larval behaviour of barnacle larvae we only allowed dispersal in the top 12 m . We further assumed that spawning occurred from June to August [103–105] and that the pelagic larval duration lasted between 20 and 30 days . The dispersal modelling was repeated for eight years (1995–2002) to account for short-term climate variations due to the North-Atlantic Oscillation cycle . Within each locality consisting of 10 model grid cells (3.7 x 3.7 km2) we released a total of 47040 particles. Dispersal probabilities between the seven localities were calculated by summing all the trajectories starting in locality i having end positions within locality j normalized by the total number of simulated trajectories from locality i. Also the probability to disperse within any locality was calculated. In addition, we calculated the multi-generation connectivity where stepping-stone dispersal was allowed over 16 and 32 single-generation dispersal events, and summed over all possible dispersal routes . We chose to include a broad range of habitats since the distribution of B. improvisus is poorly described in the Baltic Sea, and to reduce the risk of underestimating connectivity. However, we excluded all areas below 100 m, which are anoxic. Tests for correlations between multi-generational connectivity and linearized genetic differentiation (FST /1- FST) were performed using the same Mantel test as described above. The data on linearized FST for genetic differentiation are symmetric between localities. This is generally not the case for oceanographic connectivity and in the study area there are marked asymmetries in water transport. To allow for a Mantel test we made the connectivity matrices symmetric by calculating the minimum connectivity between each pair of localities, which may arguably be best correlated with increasing FST. Minimum connectivity also explained more of the genetic differentiation compared to the maximum, as well as to the arithmetic and geometric means of connectivity.
COI sequence variation and diversity within populations
A 694 bp sequence of the cytochrome oxidase I (COI) was successfully obtained for 448 individuals from 14 populations (Table 1). Sequence comparison revealed 139 distinct haplotypes, defined by 95 polymorphic sites (including 90 transitions, 15 transversions and no indels). Most of the haplotypes (90 out of 139, or 65%) were singletons. Of the remaining 49 haplotypes, 41 were shared among populations and eight were found in a single population (S3 Table). Only one haplotype (BiH001) was shared among all populations except for the Chesapeake Bay (CB). The number of haplotypes per population varied between nine (Black Sea) and 27 (Saltö), not considering North Carolina (NC) and Chesapeake Bay (CB), which had very small sample sizes (Table 1). The NC and CB populations also showed markedly lower haplotype diversity (0.29–0.52) and nucleotide diversity (0.0009–0.0012) compared to the other populations, which may be a result of small sample sizes (Table 1). Therefore, NC and CB populations were only included in sequence-based comparisons (e.g. Figs 1 and 2; where they were grouped as “Atlantic US”) but excluded from frequency-based analyses (pairwise ΦST). Haplotype diversity for the 12 remaining populations ranged between 0.75–0.96. The lowest haplotype diversity out of these 12 populations was found in the northern Baltic populations—Umeå (0.85) and Öregrund (0.75)—as well as in the Caspian Sea (0.83) and Black Sea (BL: 0.82), and the highest was observed in the Estonian (0.96) and Japanese (0.95) populations (Table 1). Nucleotide diversity in the same 12 populations was relatively low and ranged from 0.0025 to 0.0069, with the lowest diversity found in the Black Sea (0.0025) and the highest diversity in Japan (0.0069) and Argentina (0.0063) (Table 1).
The twelve most common haplotypes (represented by five or more individuals) are colour-coded and the “white proportion” contains all the haplotypes represented by five or less individuals including unique haplotypes.
Each node represents one mutational step and the size of each circle represents the frequency of a haplotype. The pie chart colours represent different geographical groupings of populations; PU = Pacific US, AU = Atlantic US (North Carolina and Chesapeake Bay combined), AR = Argentina, FR = France, BS = Baltic South (including Kiel, Torhamn, Estonia and Saltö), BN = Baltic North (Öregrund and Umeå), CA = Central Asia (Caspian Sea and Black Sea) and JP = Japan. Small empty circles (nodes) represent non-sampled haplotypes.
The geographic distribution of the 12 most frequent haplotypes (across the samples) revealed high haplotype diversity in most populations (Fig 1). The most common haplotypes (BiH001 and BiH002) were found in almost all sampled populations (S3 Table). The haplotype BiH001 was found in particularly high proportion within the Caspian Sea, Black Sea and in the northern Baltic Sea (Fig 1). The most common haplotype in the “Atlantic US” (BiH008) was only found in the Caspian Sea and Black Sea, suggesting a possible introduction to these areas from the Atlantic US. However, haplotype diversity was considerably higher within the Caspian Sea and Black Sea, indicating the possibility of multiple sources contributing to the populations found in these regions, although small samples sizes from the Atlantic US populations limits the possibility to draw conclusions on this issue.
The maximum likelihood phylogenetic analysis showed a shallow topology with no support for division of haplotypes into different clades (most bootstrap values were zero; see S3 Fig). The minimum spanning tree revealed a star-like shape with four main haplotypes surrounded by many unique haplotypes, mostly differing by no more than one or two mutation(s) (Fig 2). The star-like shape of the minimum spanning tree was congruent with the high haplotype diversity and low nucleotide diversity described above for most populations, and this result suggests a recent demographic expansion of populations. The hypothesis of a recent demographic expansion in most populations is also supported by the tests of neutral evolution. For example, Tajima’s D showed significant negative values for all populations except the Black Sea population, and Fu’s FS showed significant negative values for all analysed populations (Fig 3). Furthermore, the mismatch distribution analysis displayed unimodal distributions for all populations and the observed mismatch distributions were not significantly different from the expected distributions under a sudden expansion model (R: 0.02–0.07; p>0.05), indicating recent demographic expansions (Fig 3). The only indications of possible bottlenecks, based on haplotype and nucleotide diversity, were observed in the Black Sea, Caspian Sea and northern Baltic, where lower genetic diversity may reflect a founder effect. These areas are all relatively isolated, to which human-mediated dispersal on ships is the most likely route of colonisation. However, the BOTTLENECK tests based on microsatellite loci found no significant bottlenecks (see below).
Observed mismatch distribution (bars) and expected mismatch distributions under the sudden expansion model (solid line) of mtDNA COI sequences for each population. D = Tajima’s D, FS = Fu’s FS, R = raggedness value. Significance is presented as *: P = 0.01–0.05, **: P < 0.01. The populations are sorted according to the pattern of mismatch distribution, with populations displaying more diverged haplotypes (2–4 mismatches) to the left, and populations with less differentiated haplotypes (1–2 mismatches) to the right.
Microsatellite diversity within populations
A total of 443 individuals from 12 populations were successfully genotyped at four microsatellite loci. None of these loci were successfully amplified for the North Carolina (NC) and Chesapeake Bay (CB) populations despite multiple PCR trials. We identified 78 alleles in total (over all four loci), where the mean number of alleles per population ranged from 8 to 13 (S4 Table). No signs of large allele dropout or stutter were detected. There were no indications of linkage disequilibrium between the four loci (p > 0.6). Mean expected heterozygosity (HE) was similar in all populations, ranging from 0.76 in Kiel to 0.83 in Torhamn (Table 1). Allelic richness was highest in France (9.1) and Saltö (9.0) and lowest in the Black Sea (7.5) and Caspian Sea (7.7). All four loci deviated from Hardy Weinberg Equilibrium (HWE) and three out of these four loci (B4, E29 and E42) showed considerable heterozygote deficiency (S2 Table, S4 Table and S5 Table). Their mean null allele frequencies were estimated to 0.39, 0.23 and 0.48, respectively (S5 Table). However, we found similar patterns of genetic differentiation using the locus C103 (almost in HWE), as when using the three loci with high heterozygote deficiency (S6 Table, S4 Fig). Despite indications of bottlenecks based on low nucleotide diversity (COI) in some populations, the BOTTLENECK analyses for microsatellites revealed no significant demographic bottlenecks (identified as excess of heterozygotes) using the Wilcoxon tests (P>0.05) with a FDR correction.
Genetic structure among populations
Population genetic differentiation (pairwise Φst) based on COI data varied from -0.009 (Torhamn-Kiel) to 0.167 (France-Black Sea), with an overall average of 0.039 ±0.005 (SE) (Table 2). Based on COI data, 32 out of 66 population pairs were significantly differentiated from each other (Table 2). France showed significant differentiation from most populations, except from Kiel (KL) and Pacific US (PU). Both Argentina (AR) and Pacific US (PU) were only differentiated from the semi-enclosed, low salinity seas: Caspian Sea (CS), Black Sea (BL) and northern Baltic (UM and OR). The Caspian Sea (CS) and Black Sea (BL) were not significantly differentiated from each other. The Black Sea was differentiated from all other populations (Φst = 0.043–0.192) whereas the Caspian Sea (CS) was differentiated from most other populations, except the southern Baltic (SA, KL and TO) and Japan (JP). Populations in the northern Baltic (UM and OR) were not differentiated from each other (Φst = -0.007, p>0.84). Similarly, the southern Baltic-Skagerrak populations (SA, KL, TO, ES) were not differentiated from each other (Φst <0.003, p>0.3). However, the northern Baltic was differentiated from some populations in the southern Baltic (ES and KL; Φst = 0.027–0.03, p<0.014), and was highly divergent from populations in the Atlantic and Skagerrak (AR, FR and SA, Φst = 0.044–0.127, p<0.0001).
Estimates of microsatellite pairwise FST, after null allele corrections (mean FST = 0.013±0.001 SE), were similar to uncorrected estimates (FST = 0.015±0.001) (S6 Table). Hence, the null allele corrected data were used in the population structure analyses. The microsatellite data showed patterns of pairwise differentiation similar to the mtDNA COI results, albeit with slightly lower differentiation–pairwise FST varied from -0.002 (Black Sea-Caspian Sea) to 0.034 (France-Argentina) with an average of 0.013 ±0.001 (Table 2).
Both the MDS plots, visualising similarity between populations in 2-dimensions based on microsatellites and mtDNA, grouped the Baltic populations together, although the northern Baltic (UM and OR) and southern Baltic (TO, ES, KL) were separated into two clusters only in the mtDNA data (Fig 4). Saltö (SA) grouped together with France based on microsatellites, whereas the COI data placed Saltö (SA) with the other southern Baltic populations (KL, TO, ES), as well as partly with Argentina (AR) and the Pacific populations (PU, JP) (Fig 4). France was clearly different from most populations, especially from the northern Baltic (UM, OR) and Asian populations (CS, BL and JP), probably due to the high frequency of unique COI haplotypes present in this population (Table 1). Both microsatellites and mtDNA COI grouped the Black Sea and Caspian Sea together, where the Black Sea was more differentiated from the rest of the populations.
A) MDS based on mtDNA (ΦST): stress: 0.064; B) MDS based on microsatellites (FST): data including all four loci with null allele correction, stress: 0.11 (stress values over 0.2 suggest that the precision in representing the relationships among populations is limited).
The AMOVA based on COI sequences from the Baltic-Skagerrak populations, corroborated the Φst pattern of differentiation within the Baltic, and showed that 3.7% of the variation was partitioned between the northern Baltic and southern Baltic-Skagerrak (FCT = 0.037; p<0.0001), whereas slightly less variation (2.1%) was explained when separating Saltö from the Baltic populations (FCT = 0.021; P<0.0001) (S7 Table). No variation between populations within these groupings was detected in either analysis (FSC < 0.009, p>0.07) (S7 Table). In contrast, the Fst values based on microsatellite data (Table 2), showed that Saltö (SA) was more differentiated from the other Baltic populations (Fst: 0.008–0.014) than the other Baltic populations were from each other (Fst: -0.004–0.008), indicating a potential dispersal barrier at the entrance to the Baltic Sea. This pattern has been observed for many other Baltic species [61,62].
Inference of invasive history based on modelling
The results from the DIYABC modeling of the potential invasive history of B. improvisus, showed that scenarios 3 and 4 received very weak support based on the posterior probabilities using logistic regression of COI summary statistics (S5A Fig). This may suggest that Europe is an unlikely native area (as in scenario 4), and that the invasion into the Black Sea and the Caspian Sea occurred from the Pacific region rather than from the eastern Atlantic (as in scenarios 1 and 2). The picture was slightly different when scenario probabilities were based on the observed microsatellite loci. Here, scenario 2 received very weak support where the US Pacific was the native area (S5B Fig), while scenario 3 with Argentina as the native area received the strongest support. A possible consensus based on both COI and microsatellite markers, is that scenarios with Argentina (scenarios 1 and 3) representing the native area received the strongest support. However, the rather different results for the COI and microsatellite markers indicate that the invasion structure is weak, again pointing to that mixing through shipping has been important. For the more detailed scenarios of the invasion into the Baltic Sea both markers gave weak support for the hypothesis of a Ponto-Caspian introduction (S6 Fig).
Posterior estimates of the time for introductions generally point to a recent expansion in the range of 100 to 1000 generations. Surprisingly, the COI data produced slightly more recent split times than the microsatellite data. These recent split times are also consistent with a global mixing of genotypes through the increase of shipping activity the past 200 years.
Comparison of gene flow to geographical distance and oceanographic connectivity
On a global scale, there was a significant pattern of isolation by geographical distance (defined as “closest shipping distance”) for microsatellites (R = 0.37, P = 0.005), but not for mitochondrial DNA (R = 0.032, P = 0.49). In the Baltic-Skagerrak region, microsatellites showed no isolation by distance (IBD) (R = 0.30, P = 0.13), and for mitochondrial DNA there was a nearly significant IBD (R = 0.46, P = 0.075) (Fig 5A). Oceanographic connectivity, on the other hand, seemed to better explain patterns of genetic differentiation in the Baltic-Skagerrak region where both microsatellites (R = -0.54, P = 0.008) and mitochondrial DNA (R = -0.68, P = 0.025) revealed a significant correlation between genetic differentiation and (minimum) connectivity between sampling sites (Fig 5B and 5C).
Pairwise genetic distance (ΦST, COI) for populations of B. improvisus in the Baltic Sea plotted as a function of (a) closest geographical shipping distance (km) (Mantel test: R = 0.46, p = 0.075); (b) minimum oceanographic connectivity between sampling sites; (c) standardised and null allele corrected pairwise genetic distance ((FST/1-FST) in microsatellites plotted as a function of minimum oceanographic connectivity between sampling sites.
High levels of genetic diversity in B. improvisus
In this first study investigating global population genetic structure of the widely distributed barnacle B. improvisus, we found high levels of genetic diversity in both mitochondrial and nuclear markers in most populations (h = 0.75–0.96, Ar = 7.5–9.1). These levels are also high when compared to several other invertebrate species with similar invasive backgrounds, such as the tunicate Botrylloides violaceus (h = 0.0–0.61, Ar = 3.0–4.4) , or the globally invasive crab Carcinus maenas (h = 0.3 vs. 0.8; Ar = 3.1 vs. 4.4, for introduced vs. native populations) . However, a recent global phylogenetic study on a closely related barnacle Amphibalanus amphitrite  revealed similar levels of genetic diversity in COI (h = 0.83–0.97).
Previous phylogeographic studies on barnacles have typically investigated regional genetic structure connected to present dispersal patterns [110,111] or questions related to historical post-glacial colonisation events [30,112]. Few genetic studies have aimed to assess the source of origin and colonization history of invasive barnacles (but see [24,113,114]). In the study by Geller et al. , they used mitochondrial and nuclear genes to successfully identify the source of introduction of several Balanus glandula populations along the Pacific US coast. However, the genetic diversity within populations was very low due to the recent invasion (1970’s), and no evidence of multiple sources was detected. In contrast, Chen et al.  found high genetic diversity in COI in the barnacle A. amphitrite, where sequences were divided into three distinct clades (with a 4% divergence between clades) indicating historical isolation between populations or possibly a cryptic species complex. In comparison, we found high genetic diversity within all B. improvisus populations, but could not identify any phylogenetic clades. Furthermore, no clear colonisation routes were identified, probably because high global shipping intensity has resulted in high admixture between populations.
Microsatellite results revealed very high heterozygote deficiencies relative to Hardy-Weinberg expectations (HWE) in three out of four loci (S4 Table, S5 Table), which may be due to several different reasons including technical artefacts such as null alleles (alleles that are not amplified due to e.g. high sequence variability at primer sites) and/or biological reasons such as Wahlund effects, inbreeding and self-fertilization. B. improvisus is a simultaneous hermaphrodite, and self-fertilization has been suggested to occur [115,116], although there are no robust experimental evidence of selfing (only anecdotal observations from field collected samples) and this possibility still requires testing. Furthermore, locus C103 showed no signs of null alleles, suggesting locus-specific issues rather than inbreeding. The heterozygote deficiencies, along with the overall low success in developing polymorphic microsatellites (4 out of 108 tested loci), are possibly due to primer mismatch (resulting in null alleles) associated with the high nucleotide diversity that we observe in the on-going genome sequencing project for B. improvisus (www.cemeb.science.gu.se/research/imago-marine-genome-projects). Heterozygote deficiency is commonly observed in some invertebrate groups [26,27] and the causes and consequences for interpreting genetic structure has been widely discussed [90,117,118]. Although caution should be taken for over-interpreting our results based on microsatellites, especially since only four loci were used, several findings corroborate our results: First, we found similar patterns of genetic differentiation using the locus C103 almost in HWE, as when using the three loci with high heterozygote deficiency (S4 Fig); Second, there were only slight differences in the pattern of genetic differentiation between the null allele corrected data and the non-corrected data, based on all four loci (S6 Table); Finally, the microsatellites and the mtDNA gave overall congruent results, that also correlated with seascape connectivity (Table 2).
Origin and invasion history of B. improvisus
It is commonly stated in the literature that B. improvisus originates from the Atlantic coast of North and/or South America (e.g. [41,43,56]), although there is no clear evidence to support this. We were not able to confirm this hypothesis based on our present data; the Atlantic US populations (CB and NC) had low genetic diversity and were highly differentiated from all other populations (see Fig 1), two characteristics that are not expected of an original "source" population . Furthermore, the complete amplification failure of the microsatellite loci for the Atlantic US populations indicates high divergence from the other populations. However, it is likely that the small sample size (N = 14) from the Atlantic US population has biased our data in this region, and it is thus impossible to draw conclusions about the hypothesis of an Atlantic North-American origin of B. improvisus. However, we should note that we sampled and sequenced 128 putative "B. improvisus" individuals from the Chesapeake Bay (CB) and North Carolina (NC), but that most of these were later most likely identified as Balanus subalbidus based on their highly divergent COI sequences and additional sequencing of other individuals that were identified as B. subalbidus based on morphology. B. subalbidus is an estuarine barnacle species with similar morphology to B. improvisus, especially as juveniles . The high frequency of B. subalbidus in our samples indicates that B. subalbidus may be more common in the Chesapeake Bay and Neuse River area (North Carolina), than previously acknowledged . Barnacles are known to cause competitive exclusion [120,121], however, nothing is known about the competitive interactions between B. subalbidus and B. improvisus. If B. improvisus has indeed been partly out-competed by B. subalbidus in low salinities, then the remaining small population sizes of B. improvisus could result in partial isolation and reduced genetic diversity. Furthermore, nothing is known about possible hybridization between B. improvisus and B. subalbidus, where e.g. introgression also could explain the mismatch between the amplification of mtDNA and microsatellites for these samples. Clearly, more studies are required to evaluate this possibility.
In contrast to the north Atlantic US population, there is more support for the Argentinian (southwest Atlantic) population being part of the native region for B. improvisus, based on relatively high haplotype and nucleotide diversity (Table 1), as well as showing no signs of recent expansions (Fig 3). Furthermore, most haplotypes that were found in the Argentinian population were also found in other populations globally (Fig 1). The ABC modelling also suggested that B.improvisus originates from America (S1 Fig, S5 Fig). To confirm this hypothesis, however, additional populations from the area (including both north and south America) must be sampled and compared.
Being a common fouling species on ship hulls, B. improvisus has probably been transported between coasts for many centuries, even before the species first was described by Darwin in 1854. Therefore, the early descriptions of B. improvisus from the Atlantic American coast may not provide evidence of the true origin of B. improvisus. The fact that many introduced populations had a high proportion of private haplotypes (i.e. haplotypes not shared with other populations, including potential native regions), suggests that we may have missed several native populations in our sampling for this study, and thereby the extent of native genetic diversity could be greatly underestimated. For example, the French population displayed a remarkably high proportion (34%) of unique haplotypes (Table 1). Alternatively, the notionally native North-American population(s) may have experienced a recent bottleneck, possibly due to competition with other barnacle species such as B. subalbidus; such a bottleneck may have drastically reduced the native genetic diversity to levels below those of the invasive populations. The spatial and temporal resolution of the genetic and historical data is as yet inadequate to permit a coherent conclusion about the native region.
Evidence of bottlenecks and recent demographic expansions
A common assumption regarding introduced populations has been that they should experience loss of diversity relative to native sources because of founder effects and post-introduction demographic bottlenecks [61,119]. For example, Flight et al.  found higher haplotype diversity and allelic richness in European native populations of the barnacle S. balanoides (h = 0.99, Ar = 15.7) compared to North American (introduced) populations (h = 0.86, Ar = 12.3). However, many recent studies have shown that introduced populations show little reduction in genetic diversity, and that some may even display higher genetic diversity than in the native range, caused by multiple introductions from different source populations [113,122].
In our study, we found high genetic diversity in almost all populations (except “Atlantic US”; h = 0.75–0.96, Ar = 7.5–9.1). This is likely the result of multiple introductions from different sources, with large founding populations and/or potential admixture, e.g. [122,123], exemplified here by fouling organisms such as barnacles, that often form dense populations on ship hulls and are transported between harbours over large distances.
The mismatch distribution analysis as well as the neutrality tests, confirm expectations of a model of recent and sudden expansion of most populations (Fig 3). High haplotype diversity and low nucleotide diversity among haplotypes is another residual effect of recent evolutionary history. The only indications of possible bottlenecks, based on haplotype and nucleotide diversity, were observed in the Black Sea, Caspian Sea and northern Baltic, where lower genetic diversity may reflect a founder effect. These areas are all relatively isolated, to which human-mediated dispersal on ships is the most likely route of colonisation. However, based on the BOTTLENECK analyses for microsatellites, no significant demographic bottlenecks (identified as excess of heterozygotes) were detected using the Wilcoxon tests (P>0.05) with a FDR correction.
The main haplotype in the Atlantic US samples (BiH008) was also found in the Caspian Sea and Black Sea populations, supporting the idea that at least one introduction from Atlantic US to these regions has occurred (Fig 2). This introduction pathway has also been suggested for other species, e.g. Mnemiopsis leyidi . However, the Caspian Sea and Black Sea also contained many other haplotypes not represented in the Atlantic US suggesting multiple introductions from different areas (notwithstanding the potential sampling error in the Atlantic US populations mentioned previously).
The role of natural larval dispersal versus human-mediated dispersal
Colonisation history of fouling organisms is often complex, since human-mediated dispersal via maritime activities has occurred for several centuries. Especially during the past century, shipping intensity has increased dramatically, facilitating new invasions of non-indigenous species as well as potentially modifying population genetic structure in already established species . Previous studies on other invasive invertebrate species have shown that patterns of long distance dispersal are correlated with human-mediated vectors [109,124].
The high genetic diversity that we observed in most populations including many shared haplotypes on a global scale, suggests that shipping and other anthropogenic activities have played an important role in shaping the current population genetic structure of B. improvisus. An interesting pattern, related to historical shipping, is the unexpectedly high similarity between Argentina (AR) and Saltö (SA) (Φst = -0.003). From the late 19th century to early 20th century there was intensive shipping activity between these regions: the small town Krokstrand (close to Saltö) in Sweden was a key port from which granite was shipped to several parts of the world during many decades, including Buenos Aires in Argentina where it was used to pave the streets of the city .
Another common source of introductions of non-native species is indirectly through aquaculture activities . The high diversity and large proportion of private haplotypes observed in the French population in our study could partly be a result of multiple introductions from different source populations due to aquaculture. Although B. improvisus was observed in French waters already in 1872 , they were most likely also introduced along with the import of oysters (Crassostrea gigas) to French oyster farms in the 1960’s . Other examples of species that were introduced to France through aquaculture are the widespread brown alga Sargassum muticum and the mollusc Crepidula fornicata [26,129]. Similar to our study, Dupont et al.  found high genetic diversity in introduced C. fornicata populations along the French coast, and suggested that these populations derived from several genetically diverse, but poorly differentiated source populations.
On a global scale, there was a pattern of isolation by geographical distance (defined as closest shipping distance) for microsatellites, but not for mitochondrial DNA. One possible explanation for this pattern is that most CO1 haplotypes were already present when global shipping began to translocate populations between places, thus resulting in a re-shuffling of “old” haplotypes that may poorly reflect geographic distance. Microsatellites on the other hand may reflect contemporary processes of gene flow and random drift; while it is also possible that new alleles can arise through mutations to produce an “isolation by distance” signal, it is unlikely that new mutations have caused this pattern (cf. ), given the short time for human-induced introduction (<200 years in most cases). In this context, it is interesting that Flight et al.  found similar differences between marker types for the barnacle S. balanoides.
Oceanographic connectivity seemed to better explain patterns of genetic differentiation at a regional scale (in the Baltic-Skagerrak region) where both microsatellites and mitochondrial DNA revealed a significant correlation with (minimum) connectivity between sampling sites (Fig 5B and 5C). The lack of genetic differentiation between two northern Baltic populations (UM and OR) as well as the low differentiation between the southern Baltic populations supports the idea that larval dispersal also plays an important role for further spreading of the species within the Baltic Sea (Table 2). Several studies on invertebrates with pelagic larval stages (including barnacles) have shown the importance of oceanographic connectivity for determining population structure [111,131,132]. However, oceanographic connectivity probably acts in synergy with human-mediated dispersal. As an example, the differentiation of the northern Baltic populations (OR, UM) from the southern Baltic populations (KL, TO, ES) may be a combination of restricted larval dispersal as well as independent colonisation events via shipping from different source populations.
Notably, the AMOVA comparing groups of populations within the Baltic Sea (S7 Table) indicated that there is a stronger differentiation between the northern and southern Baltic-Skagerrak than between the Baltic populations and Saltö (Skagerrak). This is interesting, since the entrance to the Baltic Sea (Danish Belt/Öresund) is often considered a main breakpoint in genetic differentiation in many species, clearly separating Baltic and Atlantic populations . However, selection for local adaptations to different environmental conditions (especially the salinity gradient in the Baltic), in addition to restricted gene flow due to oceanographic currents, may be an explanation for this pattern . This could thus partly explain the different results observed in our study, since we used assumedly neutral markers (at least the microsatellites), reflecting mostly neutral processes including drift and gene flow, rather than selection. The role of natural selection caused by the strong environmental gradients in the Baltic Sea was recently investigated , however no populations from the northern Baltic Sea populations (latitude >60°N, salinity <5psu) were included, hence it remains to be studied in B. improvisus.
Ecological and evolutionary implications for B. improvisus
That B. improvisus is a cosmopolitan species, successfully colonising many parts of the world, is not surprising. The species displays several life history traits that are associated with successful invaders including broad environmental tolerance (both to natural factors such as salinity, temperature and desiccation, but also to pollution; [9,48,133,134]. Broad tolerance may be a result of phenotypic plasticity, where an organism can maintain or adjust their phenotype when the environment changes, without genetic changes necessarily occurring . An interesting comparison is the closely related barnacle A. amphitrite that has a narrower environmental tolerance  and displays a more structured population genetic pattern with several distinct clades , compared to B. improvisus.
Furthermore, B. improvisus displays multiple modes of dispersal (both through free-swimming larvae and fouling), has an extended spawning season and a low habitat selectivity at settlement , all of which facilitate colonisation of new areas at different scales. However, the species is often restricted to low salinity environments such as estuaries and coastal lagoons and harbours, suggesting that it may have a limited ability to successfully compete with other sessile invertebrates in high salinity environments. Estuaries are generally considered to be sources of diversification by adaptive divergence and spatial isolation . Fluctuations in many abiotic factors, combined with isolation of habitat from other estuarine areas, often contributes to genetic structuring, enhancing effects of micro-evolutionary processes including drift, selection and gene flow . However, in B. improvisus, intensive human-mediated gene flow probably counteracts these processes, resulting in the pattern of high diversity and little differentiation even on a global scale, as we found in this study. This high diversity and gene flow from multiple sources may have implications for the ability of B. improvisus to adapt to local conditions, compete with native species, spread to new areas as well as cope with future environmental change.
In conclusion, we found high genetic diversity in most populations, including both putatively native and introduced regions, indicating high gene flow between populations via shipping and other anthropogenic activities. Furthermore, mitochondrial and nuclear markers revealed similar patterns. Our results could not clearly identify the geographic origin of B. improvisus, but suggest that Argentina may at least be part of the native region. Further sampling of multiple populations in this area is required to clarify the pattern. Evidence of bottlenecks associated with founder events were only observed in the Caspian and Black Sea and in the northern Baltic Sea, although most populations showed genetic patterns consistent with recent expansion events. Oceanographic connectivity, rather than geographic distance, seems to explain regional genetic structure within the Baltic Sea, highlighting the importance of larval dispersal for further natural spreading of introduced species. Knowledge about different types of dispersal dynamics of non-native species is crucial to our understanding of both evolutionary aspects of colonisation as well as future management of biological invasions. The implications of the broad environmental tolerance and high dispersal capacity of B. improvisus for its ability to adapt to local conditions and tolerate future environmental changes remains to be elucidated.
S1 Fig. Schematic figures of modeled scenarios of introductory history for Balanus improvisus.
The first four scenarios (a, b) explore the most likely origin of B. improvisus on a global scale while the last two scenarios focus on two alternative colonization routes for the Baltic Sea. Abbreviations of population samples are: AR–Argentina, FR–France, SA–Saltö, BA–Baltic Sea, BL–Black Sea, CS–Caspian Sea, PU–Pacific US, JP–Japan, BA-S–southern Baltic Sea (KL, Torhamn, Estonia), BA-N–northern Baltic Sea (Öregrund, Umeå). See S1 Table for more geographic information.
S2 Fig. PCA on summary statistics from the simulated datasets of the four global scenarios and the observed data (yellow) based on; (a) COI sequences and (b) microsatellite loci.
S3 Fig. Maximum likelihood phylogenetic tree based on all B. improvisus individuals in the study.
The robustness of branches was tested with 100 bootstraps, and a COI sequence of Balanus eburneus was used as an out-group for this maximum likelihood analysis.
S4 Fig. Multi-dimensional scaling plot (MDS) based on pairwise FST (not corrected for null alleles) for the different microsatellite loci.
A) locus C103 separately; B) B4, E29 and E42 pooled. Although the loci B4, E29 and E42 showed high heterozygote deficiency, they display similar patterns to locus C103.
S5 Fig. Posterior probability of the four ABC modeling scenarios for invasion of B. improvisus on a global scale.
The figures are based on samples of (a) COI sequences, and (b) microsatellite loci.
S6 Fig. Posterior probability of the two scenarios for introduction of B. improvisus into the Baltic Sea.
The figures are based on samples of (a) COI sequences and (b) microsatellite loci.
S1 Table. Information about the spatial and temporal sampling of DNA from Balanus improvisus.
S2 Table. Microsatellite primer information.
The table contains locus name, repeat motif, primer sequence, universal tail name, MgCl2 (mM) = magnesium chloride concentration, Tm = melting temperature, Ta1 and Ta2 = annealing temperatures (see PCR protocols in methods); number of PCR cycles, Allele size range (bp); N = number of samples used to test microsatellites; Na = total number of alleles; Ho = observed heterozygosity; He = expected heterozygosity; FIS = inbreeding coefficient; FST uncorr. = uncorrected average genetic distance between populations; FST corr. = null allele corrected average pairwise genetic distance between populations. Also included are the different tails and dyes added to the microsatellite primers (T1-D2, T2-D2, T2-D4).
S3 Table. Haplotype list with haplotype frequencies (based on COI) for all B. improvisus populations.
The populations are: Argentina (AR), France (FR), Saltö (SA), Kiel (KL), Torhamn (TO), Estonia (ES), Öregrund (OR), Umeå (UM), Black Sea (BL), Caspian Sea (CS), Japan (JP) and Pacific US (PU).
S4 Table. Summary of microsatellite loci results for each population tested.
Sample size (N), expected (He) and observed (Ho) heterozygosities, deviations from HWE (FIS) according to Weir & Cockerham (1984), number of alleles (A) and number of alleles per locus (Ar). Populations: Argentina (AR), France (FR), Saltö (SA), Kiel (KL), Torhamn (TO), Estonia (ES), Öregrund (OR), Umeå (UM), Black Sea (BL), Caspian Sea (CS), Japan (JP) and Pacific US (PU).
S5 Table. Estimated null allele frequencies for microsatellites.
The EM algorithm (Dempster et al. 1977) was used for each locus and B. improvisus population separately.
S6 Table. Pairwise FST between B. improvisus populations based on four microsatellite loci.
Below diagonal: Pairwise FST without correction for null alleles using ENA correction, FDR corrected significant differentiation in bold (FDR adjusted α0.05 = 0.0341); above diagonal: null-allele corrected pairwise FST between population. NB: p-values cannot be obtained in FreeNA when <5 loci.
S7 Table. Analysis of molecular variance (AMOVA) results on B. improvisus
Based on mtDNA (COI) data for populations in the Baltic Sea-Skagerrak region: UM = Umeå, OR = Öregrund, TO = Torhamn, KL = Kiel, ES = Estonia, SA = Saltö.
This work was supported by a Linnaeus-grant from the Swedish Research Councils, VR and Formas (http://www.cemeb.science.gu.se), as well as from Stiftelsen Wilhelm och Marina Lundgrens Vetenskapsfond. A special thanks to our collaborators worldwide for their help with collecting barnacles for this study: Toshi Yamaguchi, Richard Emlet, Daniel Rittschof, Eric Holm, Richard Osman, Eduardo Spivak, Alireza Sari, Ventzi Karamfilow, Tiia Möller, Christian Pansch and Daniel Johansson, as well as to Zuzanna Zadrodzka and Anna-Karin Ring for assistance in the lab.
Conceived and designed the experiments: ALW GC PRJ CA. Performed the experiments: ALW GC AT. Analyzed the data: ALW GC AT MAR CA. Contributed reagents/materials/analysis tools: ALW GC MAR CA. Wrote the paper: ALW GC AB JNH PRJ CA.
- 1. Seebens H, Gastner MT, Blasius B. The risk of marine bioinvasion caused by global shipping. Ecol Lett. 2013;16(6):782–90. pmid:23611311
- 2. Banks NC, Paini DR, Bayliss KL, Hodda M. The role of global trade and transport network topology in the human-mediated dispersal of alien species. Ecol Lett. 2015;18(2):188–99. pmid:25529499
- 3. Molnar JL, Gamboa RL, Revenga C, Spalding MD. Assessing the global threat of invasive species to marine biodiversity. Front Ecol Environ. 2008;6(9):485–92.
- 4. Ricciardi A. Are modern biological invasions an unprecedented form of global change? Conserv Biol. 2007;21(2):329–36. pmid:17391183
- 5. Ruiz GM, Fofonoff P, Hines AH, Grosholz ED. Non-indigenous species as stressors in estuarine and marine communities: Assessing invasion impacts and interactions. Limnol Oceanogr. 1999;44(3):950–72.
- 6. Lee CE. Evolutionary genetics of invasive species. Trends Ecol Evol. 2002;17(8):386–91.
- 7. Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With KA, et al. The population biology of invasive species. Annu Rev Ecol Syst. 2001;32:305–32.
- 8. Sax DF, Stachowicz JJ, Brown JH, Bruno JF, Dawson MN, Gaines SD, et al. Ecological and evolutionary insights from species invasions. Trends Ecol Evol. 2007;22(9):465–71. pmid:17640765
- 9. Tepolt CK. Adaptation in marine invasion: a genetic perspective. Biol Invas. 2015;17:887–903.
- 10. Weersing K, Toonen RJ. Population genetics, larval dispersal, and connectivity in marine systems. Mar Ecol-Prog Ser. 2009;393:1–12.
- 11. Cowen RK, Sponaugle S. Larval dispersal and marine population connectivity. Ann Rev Mar Sci 1. Palo Alto: Annual Reviews; 2009. p. 443–66. pmid:21141044
- 12. Levin LA. Recent progress in understanding larval dispersal: new directions and digressions. Integr Comp Biol. 2006;46(3):282–97. pmid:21672742
- 13. Hartl DL, Clark AG. Principles of population genetics. 4th ed. Sunderland MA, USA: Sinauer Associates; 2007.
- 14. Palumbi SR. Genetic divergence, reproductive isolation and marine speciation. Annu Rev Ecol Syst. 1994;25:547–72.
- 15. Barber PH, Palumbi SR, Erdmann MV, Moosa MK. Sharp genetic breaks among populations of Haptosquilla pulchella (Stomatopoda) indicate limits to larval transport: patterns, causes, and consequences. Mol Ecol. 2002;11(4):659–74. pmid:11972755
- 16. Hellberg ME. Gene flow and isolation among populations of marine animals. Annu Rev Ecol Evol Syst. 40. Palo Alto: Annual Reviews; 2009. p. 291–310.
- 17. Kelly RP, Palumbi SR. Genetic structure among 50 species of the Northeastern Pacific rocky intertidal community. PLoS One. 2010;5(1):e8594. pmid:20062807
- 18. Knutsen H, Jorde PE, Andre C, Stenseth NC. Fine-scaled geographical population structuring in a highly mobile marine species: the Atlantic cod. Mol Ecol. 2003;12(2):385–94. pmid:12535089
- 19. Taylor MS, Hellberg ME. Genetic evidence for local retention of pelagic larvae in a Caribbean reef fish. Science. 2003;299(5603):107–9. pmid:12511651
- 20. Todd CD, Lambert WJ, Thorpe JP. The genetic structure of intertidal populations of two species of nudibranch molluscs with planktotrophic and pelagic lecithotrophic larval stages: are pelagic larvae "for" dispersal? J Exp Mar Biol Ecol. 1998;228(1):1–28.
- 21. Johannesson K, Johannesson B, Lundgren U. Strong natural selection causes microscale allozyme variation in a marine snail. Proc Natl Acad Sci U S A. 1995;92(7):2602–6. pmid:7708692
- 22. Schmidt PS, Rand DM. Adaptive maintenance of genetic polymorphism in an intertidal barnacle: Habitat- and life-stage-specific survivorship of Mpi genotypes. Evolution. 2001;55(7):1336–44. pmid:11525458
- 23. Veliz D, Duchesne P, Bourget E, Bernatchez L. Genetic evidence for kin aggregation in the intertidal acorn barnacle (Semibalanus balanoides). Mol Ecol. 2006;15(13):4193–202. pmid:17054512
- 24. Chen H-N, Tsang LM, Chong VC, Chan BKK. Worldwide genetic differentiation in the common fouling barnacle, Amphibalanus amphitrite. Biofouling. 2014;30:1067–78. pmid:25343722
- 25. Darling JA, Bagley MJ, Roman J, Tepolt CK, Geller JB. Genetic patterns across multiple introductions of the globally invasive crab genus Carcinus. Mol Ecol. 2008;17(23):4992–5007. pmid:19120987
- 26. Dupont L, Jollivet D, Viard F. High genetic diversity and ephemeral drift effects in a successful introduced mollusc (Crepidula fornicata: Gastropoda). Mar Ecol-Prog Ser. 2003;253:183–95.
- 27. Astanei I, Gosling E, Wilson J, Powell E. Genetic variability and phylogeography of the invasive zebra mussel, Dreissena polymorpha (Pallas). Mol Ecol. 2005;14(6):1655–66. pmid:15836640
- 28. Reusch TBH, Bolte S, Sparwel M, Moss AG, Javidpour J. Microsatellites reveal origin and genetic diversity of Eurasian invasions by one of the world's most notorious marine invader, Mnemiopsis leidyi (Ctenophora). Mol Ecol. 2010;19(13):2690–9. pmid:20561193
- 29. Rius M, Pascual M, Turon X. Phylogeography of the widespread marine invader Microcosmus squamiger (Ascidiacea) reveals high genetic diversity of introduced populations and non-independent colonizations. Divers Distrib. 2008;14(5):818–28.
- 30. Campo D, Molares J, Garcia L, Fernandez-Rueda P, Garcia-Gonzalez C, Garcia-Vazquez E. Phylogeography of the European stalked barnacle (Pollicipes pollicipes): identification of glacial refugia. Mar Biol. 2010;157(1):147–56.
- 31. Remerie T, Vierstraete A, Weekers PHH, Vanfleteren JR, Vanreusel A. Phylogeography of an estuarine mysid, Neomysis integer (Crustacea, Mysida), along the north-east Atlantic coasts. J Biogeogr. 2009;36(1):39–54.
- 32. Garroway CJ, Radersma R, Sepil I, Santure AW, De Cauwer I, Slate J, et al. Fine-scale genetic structure in a wild bird population: the role of limited dispersal and environmentally based selection as causal factors. Evolution. 2013;67(12):3488–500. pmid:24299402
- 33. Hoffman JI, Clarke A, Clark MS, Fretwell P, Peck LS. Unexpected fine-scale population structure in a broadcast-spawning Antarctic marine mollusc. PLoS One. 2012;7(3).
- 34. Knutsen H, Olsen EM, Jorde PE, Espeland SH, Andre C, Stenseth NC. Are low but statistically significant levels of genetic differentiation in marine fishes 'biologically meaningful'? A case study of coastal Atlantic cod. Mol Ecol. 2011;20(4):768–83. pmid:21199035
- 35. Gonzalez EG, Zardoya R. Relative role of life-history traits and historical factors in shaping genetic population structure of sardines (Sardina pilchardus). BMC Evol Biol. 2007;7.
- 36. Duvernell DD, Lindmeier JB, Faust KE, Whitehead A. Relative influences of historical and contemporary forces shaping the distribution of genetic variation in the Atlantic killifish, Fundulus heteroclitus. Mol Ecol. 2008;17(5):1344–60. pmid:18302693
- 37. Galindo HM, Olson DB, Palumbi SR. Seascape genetics: a coupled oceanographic-genetic model predicts population structure of Caribbean corals. Curr Biol. 2006;16(16):1622–6. pmid:16920623
- 38. Godhe A, Egardt J, Kleinhans D, Sundqvist L, Hordoir R, Jonsson PR. Seascape analysis reveals regional gene flow patterns among populations of a marine planktonic diatom. Proc R Soc Lond Ser B-Biol Sci. 2013;280(1773).
- 39. Serra IA, Innocenti AM, Di Maida G, Calvo S, Migliaccio M, Zambianchi E, et al. Genetic structure in the Mediterranean seagrass Posidonia oceanica: disentangling past vicariance events from contemporary patterns of gene flow. Mol Ecol. 2010;19(3):557–68. pmid:20051010
- 40. Teacher AGF, Andre C, Jonsson PR, Merila J. Oceanographic connectivity and environmental correlates of genetic structuring in Atlantic herring in the Baltic Sea. Evol Appl. 2013;6(3):549–67. pmid:23745145
- 41. De Rivera CE, Steves BP, Fofonoff PW, Hines AH, Ruiz GM. Potential for high-latitude marine invasions along western North America. Divers Distrib. 2011;17(6):1198–209.
- 42. Kennedy VS, DiCosimo J. Subtidal distribution of barnacles (Cirripedia, Balanidae) in Chesapeake Bay, Maryland. Estuaries. 1983;6(2):95–101.
- 43. Leppakoski E, Gollasch S, Gruszka P, Ojaveer H, Olenin S, Panov V. The Baltic—a sea of invaders. Can J Fish Aquat Sci. 2002;59(7):1175–88.
- 44. Carlton JT, Newman WA, Pitombo FB. Barnacle invasions: introduced, cryptogenic, and range expanding Cirripedia of North and South America. In: Galil BS, Clark PF, Carlton JT, editors. In the wrong place—alien marine crustaceans: Distribution, biology and impacts. Invading Nature—Springer Series in Invasion Ecology. Dordrecht: Springer; 2011. p. 716.
- 45. Foster BA, Willan RC. Foreign barnacles transported to New Zealand on an oil platform. N Z J Mar Freshwater Res. 1979;13(1):143–9.
- 46. Gollasch S. The importance of ship hull fouling as a vector of species introductions into the North Sea. Biofouling. 2002;18(2):105–21.
- 47. Fyhn HJ. Holeuryhalinity and its mechanisms in a cirriped crustacean, Balanus improvisus. Comp Biochem Physiol A-Physiol. 1976;53(1):19–30.
- 48. Wrange AL, André C, Lundh T, Lind U, Blomberg A, Jonsson PR, et al. Importance of plasticity and local adaptation for coping with changing salinity in coastal areas: a test case with barnacles in the Baltic Sea. BMC Evol Biol. 2014;14.
- 49. Pansch C, Nasrolahi A, Appelhans YS, Wahl M. Tolerance of juvenile barnacles (Amphibalanus improvisus) to warming and elevated pCO2. Mar Biol. 2013;160(8):2023–35.
- 50. Lenz M, da Gama BAP, Gerner NV, Gobin J, Groner F, Harry A, et al. Non-native marine invertebrates are more tolerant towards environmental stress than taxonomically related native species: Results from a globally replicated study. Environ Res. 2011;111(7):943–52. pmid:21632049
- 51. Corell H, Moksnes PO, Engqvist A, Doos K, Jonsson PR. Depth distribution of larvae critically affects their dispersal and the efficiency of marine protected areas. Mar Ecol-Prog Ser. 2012;467:29–46.
- 52. Leppäkoski E, Olenin S. Non-native species and rates of spread: Lessons from the brackish Baltic Sea. Biol Invasions. 2000;2(2):151–63.
- 53. Darwin C. A monograph on the sub-class Cirripedia: The Balanidae (or sessile cirrepedes); the Verrucidae. London: The Ray Society; 1854.
- 54. Carlton JT, Zullo VA. Early records of the barnacle Balanus improvisus Darwin from the Pacific coast of North America. California Academy of Sciences. 1969;75:1–6.
- 55. Kawahara T. Invasion into Japanese waters by the European barnacle Balanus improvisus Darwin. Nature. 1963;198:301.
- 56. Barnes H, Barnes M. A note on Balanus improvisus in Finnish waters. Societas Scientiarum Fennica Commentationes Biologicae. 1962;XXIV(5):3–11.
- 57. Gislén T. Till kännedom om invandringen och utbredningen av Balanus improvisus vid svenska kuster. Fauna and Flora. 1950;45:32–7.
- 58. Turpayeva YP, Simkina RG. The reaction of the Black Sea Balanus improvisus (Darwin) to reduced salinities. Transactions of the Institute of Oceanography. 1961;XLIX:205–23.
- 59. Gomoiu MT, Skolka M. Changements recents dans la biodiversité de la Mer Noire dus aux immigrants. Danube Delta–Black Sea system under global changes impact Geo-Eco-Marina, Romanian Centre of Marine Geology and Geoecology 1996. p. 34–47.
- 60. Grigorovich IA, Therriault TW, MacIsaac HJ. History of aquatic invertebrate invasions in the Caspian Sea. Biol Invasions. 2003;5(1–2):103–15.
- 61. Johannesson K, André C. Life on the margin: genetic isolation and diversity loss in a peripheral marine ecosystem, the Baltic Sea. Mol Ecol. 2006;15(8):2013–29. pmid:16780421
- 62. Johannesson K, Smolarz K, Grahn M, Andre C. The future of Baltic Sea populations: local extinction or evolutionary rescue? Ambio. 2011;40(2):179–90. pmid:21446396
- 63. Furman ER. Enzyme genetic variation in Balanus improvisus Darwin (Crustacea: Cirripedia) in the Baltic Sea. Ophelia. 1989;30(1):35–45.
- 64. Furman ER. Geographical variation of Balanus improvisus in biochemical and morphometric characters. J Mar Biol Assoc UK. 1990;70(4):721–40.
- 65. Zullo VA. Marine flora and fauna of the Northeastern United States. Arthropoda: Cirripedia. USA: 1979.
- 66. Rozen S, Skaletsky H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol. 2000;132:365–86. pmid:10547847
- 67. Drummond AJ. Geneious. v.6.1 ed: Biomatters; 2010. p. Available at http://www.geneious.com.
- 68. Benson G. Tandem repeats finder: a program to analyze DNA sequences. Nucleic Acids Res. 1999;27(2):573–80. pmid:9862982
- 69. Smit AFA, Hubley R, Green P. RepeatMasker Open-3.0. http://www.repeatmasker.org1996-2010. Available from: http://www.repeatmasker.org.
- 70. Schuelke M. An economic method for the fluorescent labeling of PCR fragments. Nat Biotechnol. 2000;18(2):233–4. pmid:10657137
- 71. Real KM, Schmidt DJ, Hughes JM. Mogurnda adspersa microsatellite markers: multiplexing and multi-tailed primer tagging. Conserv Genet Res. 2009;1(1):411–4.
- 72. Brownstein MJ, Carpten JD, Smith JR. Modulation of non-templated nucleotide addition by taq DNA polymerase: Primer modifications that facilitate genotyping. Biotechniques. 1996;20(6):1004–10. pmid:8780871
- 73. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, et al. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012;28:1647–9. pmid:22543367
- 74. Nei M. Molecular evolutionary genetics: Columbia University Press, New York & Guildford, England; 1987. i–x, 1–512 p.
- 75. Excoffier L, Lischer HEL. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour. 2010;10(3):564–7. pmid:21565059
- 76. Darriba D, Taboada GL, Doallo R, Posada D. jModelTest 2: more models, new heuristics and parallel computing. Nat Methods. 2012;9(8):772–.
- 77. Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol. 2003;52(5):696–704. pmid:14530136
- 78. Templeton AR, Crandall KA, Sing CF. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. 3. Cladogram estimation. Genetics. 1992;132(2):619–33. pmid:1385266
- 79. Clément M, Posada D, Crandall KA. TCS: a computer program to estimate gene genealogies. Mol Ecol. 2000;9(10):1657–9. pmid:11050560
- 80. Teacher AGF, Griffiths DJ. HapStar: automated haplotype network layout and visualization. Mol Ecol Resour. 2011;11(1):151–3. pmid:21429113
- 81. Hasegawa M, Kishino H, Yano TA. Dating of the human ape splitting by a molecular clock of mitochondrial DNA. J Mol Evol. 1985;22(2):160–74. pmid:3934395
- 82. Excoffier L, Smouse PE, Quattro JM. Analysis of molecular variance inferred from metric distances among DNA haplotypes–application to human mitochondrial DNA restriction data. Genetics. 1992;131(2):479–91. pmid:1644282
- 83. Benjamini Y, Hochberg Y. Controlling the false discovery rate—a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 1995;57(1):289–300.
- 84. Slatkin M. A measure of population subdivision based on microsatellite allele frequencies. Genetics. 1995;139(3):1463.
- 85. Rogers AR. Genetic evidence for a Pleistocene population explosion. Evolution. 1995;49(4):608–15.
- 86. Rogers AR, Harpending H. Population growth makes waves in the distribution of pairwise genetic differences. Mol Biol Evol. 1992;9(3):552–69. pmid:1316531
- 87. Harpending HC. Signature of ancient population growth in a low resolution mitochondrial DNA mismatch distribution. Hum Biol. 1994;66(4):591–600. pmid:8088750
- 88. Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics. 1989;123(3):585–95. pmid:2513255
- 89. Fu YX. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics. 1997;147(2):915–25. pmid:9335623
- 90. Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes. 2004;4(3):535–8.
- 91. Chapuis MP, Estoup A. Microsatellite null alleles and estimation of population differentiation. Mol Biol Evol. 2007;24(3):621–31. pmid:17150975
- 92. Rousset F. GENEPOP ' 007: a complete re-implementation of the GENEPOP software for Windows and Linux. Mol Ecol Resour. 2008;8(1):103–6. pmid:21585727
- 93. Goudet J. FSTAT (Version 1.2): A computer program to calculate F-statistics. J Hered. 1995;86(6):485–6.
- 94. Weir BS, Cockerham CC. Estimating F-statistics for the analysis of population structure. Evolution. 1984;38(6):1358–70.
- 95. Piry S, Luikart G, Cornuet J. BOTTLENECK: A computer program for detecting recent reductions in the effective population size using allele frequency data. J Hered. 1999;90(4):502–3.
- 96. Cornuet JM, Ravigné V, Estoup A. Inference on population history and model checking using DNA sequence and microsatellite data with the sofware DIYABC (v1.0). BMC Bioinformatics. 2010;11:401. pmid:20667077
- 97. Cornuet JM, Pudlo P, Veyssier J, Dehne-Garcia A, Gautier M, Leblois R, et al. DIYABC v2.0: a software to make approximate Bayesian computation inferences about population history using single nucleotide polymorphism, DNA sequence and microsatellite data. Bioinformatics. 2014;30:1187–9.
- 98. Estoup A, Jarne P, Cornuet JM. Homoplasy and mutation model at microsatellite loci and their consequences for population genetics analysis. Mol Ecol. 2002;11(9):1591–604. pmid:12207711
- 99. Mantel N. Detection of disease clustering and a generalized regression approach. Cancer Res. 1967;27(2P1):209–20.
- 100. Madec G. Nemo ocean engine, version 3.3. Technical report. Technical report. Paris, France: Institut Pierre-Simon Laplace (IPSL), 2010.
- 101. Hordoir R, Dieterich C, Basu C, Dietze H, Meier HEM. Freshwater outflow of the Baltic Sea and transport in the Norwegian current: A statistical correlation analysis based on a numerical experiment. Cont Shelf Res. 2013;64(0):1–9.
- 102. Döös K. Interocean exchange of water masses. J Geophys Res: Oceans. 1995;100(C7):13499–514.
- 103. Berntsson KM, Jonsson PR. Temporal and spatial patterns in recruitment and succession of a temperate marine fouling assemblage: A comparison of static panels and boat hulls during the boating season. Biofouling. 2003;19(3):187–95. pmid:14619287
- 104. Blom S-E. Balanus improvisus Darwin on the west coast of Sweden. Zoologiska Bidrag från Uppsala. 1965;37(1):59–76.
- 105. Blom S-E, Nyholm K-G. Settling times of Balanus balanoides (L.) Balanus crenatus Brug. and Balanus improvisus Darwin on the west coast of Sweden. Zoologiska Bidrag från Uppsala. 1961;33:149–55.
- 106. Anil AC, Chiba K, Okamoto K, Kurokura H. Influence of temperature and salinity on larval development of Balanus amphitrite—implications in fouling ecology. Mar Ecol-Prog Ser. 1995;118(1–3):159–66.
- 107. Hurrell JW, Deser C. North Atlantic climate variability: The role of the North Atlantic Oscillation. J Mar Syst. 2009;78(1):28–41.
- 108. White C, Selkoe KA, Watson J, Siegel DA, Zacherl DC, Toonen RJ. Ocean currents help explain population genetic structure. Proc R Soc Lond Ser B-Biol Sci. 2010;277:1685–94.
- 109. Bock DG, Zhan A, Lejeusne C, MacIsaac HJ, Cristescu ME. Looking at both sides of the invasion: patterns of colonization in the violet tunicate Botrylloides violaceus. Mol Ecol. 2011;20(3):503–16. pmid:21199029
- 110. Cheang CC, Tsang LM, Ng WC, Williams GA, Chu KH, Chan BKK. Phylogeography of the cold-water barnacle Chthamalus challengeri in the north-western Pacific: effect of past population expansion and contemporary gene flow. J Biogeogr. 2012;39(10):1819–35.
- 111. Sotka EE, Wares JP, Barth JA, Grosberg RK, Palumbi SR. Strong genetic clines and geographical variation in gene flow in the rocky intertidal barnacle Balanus glandula. Mol Ecol. 2004;13(8):2143–56. pmid:15245390
- 112. Flight PA, O’Brien MA, Schmidt PS, Rand DM. Genetic structure and the north american postglacial expansion of the barnacle, Semibalanus balanoides. J Hered. 2012;103(2):153–65. pmid:21885571
- 113. Geller J, Sotka EE, Kado R, Palumbi SR, Schwindt E. Sources of invasions of a northeastern Pacific acorn barnacle, Balanus glandula, in Japan and Argentina. Mar Ecol-Prog Ser. 2008;358:211–8.
- 114. Yamaguchi T, Prabowo RE, Ohshiro Y, Shimono T, Jones D, Kawai H, et al. The introduction to Japan of the Titan barnacle, (Megabalanus coccopoma) (Darwin, 1854) (Cirripedia: Balanomorpha) and the role of shipping in its translocation. Biofouling. 2009;25(4):325–33. pmid:19219673
- 115. Furman ER, Crisp DJ. Biometrical changes during growth of isolated individuals of Balanus improvisus. J Mar Biol Assoc UK. 1989;69:511–21.
- 116. Furman ER, Yule AB. Self-fertilisation in Balanus improvisus (Darwin). J Exp Mar Biol Ecol. 1990;144:235–9.
- 117. Brownlow RJ, Dawson DA, Horsburgh GJ, Bell JJ, Fish JD. A method for genotype validation and primer assessment in heterozygote-deficient species, as demonstrated in the prosobranch mollusc Hydrobia ulvae. BMC Genetics. 2008;9(55).
- 118. DeWoody JA, Nason JD, Hipkins VD. Mitigating scoring errors in microsatellite data from wild populations. Mol Ecol Notes. 2006;6(4):951–7.
- 119. Dlugosch KM, Parker IM. Founding events in species invasions: genetic variation, adaptive evolution, and the role of multiple introductions. Mol Ecol. 2008;17(1):431–49. pmid:17908213
- 120. Dungan ML. Competition and the morphology, ecology, and evolution of acorn barnacles: an experimental test. Paleobiology. 1985;11(2):165–73.
- 121. Stanley SM, Newman WA. Competitive exclusion in evolutionary time—the case of the acorn barnacles. Paleobiology. 1980;6(2):173–83.
- 122. Roman J, Darling JA. Paradox lost: genetic diversity and the success of aquatic invasions. Trends Ecol Evol. 2007;22(9):454–64. pmid:17673331
- 123. Wares JP, Hughes AR, Grosberg RK. Mechanisms that drive evolutionary change. Insights from species introductions and invasions. Sax DF, Stachowicz JJ, Gaines SD, editors: Sinauer Associates Inc.; 2005. 229–57 p.
- 124. Zhan A, Perepelizin PV, Ghabooli S, Paolucci E, Sylvester F, Sardiña P, et al. Scale-dependent post-establishment spread and genetic diversity in an invading mollusc in South America. Divers Distrib. 2012;18(10):1042–55.
- 125. Karlsson U. Krokstrand 1870–1940: skärvor från en stenhuggarepok. Strömstad, Sweden. 1979.
- 126. Katsanevakis S, Zenetos A, Belchior C, Cardoso AC. Invading European seas: assessing pathways of introduction of marine aliens. Ocean Coast Manag. 2013;76:64–74.
- 127. Goulletquer P, Bachelet G, Sauriau PG, Noel P. Open Atlantic coast of Europe—a century of introduced species into French waters. Leppakoski E, Gollasch S, Olenin S, editors: Kluwer Academic Publishers; 2002. 276–90 p.
- 128. Gruet Y, Heral M, Robert JM. Premières observations sur l'introduction de la faune associée au naissain d'huÎtres Japonaises Crassostrea gigas (Thunberg), importé sur la côte Atlantique Francaise. Cahiers de Biologie Marine. 1976;17(2):173–84.
- 129. Cheang CC, Chu KH, Fujita D, Yoshida G, Hiraoka M, Critchley A, et al. Low genetic variability of Sargassum muticum (Phaeophyeae) revealed by a global analysis of native and introduced populations. J Phycol. 2010;46(6):1063–74.
- 130. Limborg MT, Pedersen JS, Hemmer-Hansen J, Tomkiewicz J, Bekkevold D. Genetic population structure of European sprat Sprattus sprattus: differentiation across a steep environmental gradient in a small pelagic fish. Mar Ecol-Prog Ser. 2009;379:213–24.
- 131. York KL, Blacket MJ, Appleton BR. The Bassian Isthmus and the major ocean currents of southeast Australia influence the phylogeography and population structure of a southern Australian intertidal barnacle Catomerus polymerus (Darwin). Mol Ecol. 2008;17(8):1948–61. pmid:18363669
- 132. Zakas C, Binford J, Navarrete SA, Wares JP. Restricted gene flow in Chilean barnacles reflects an oceanographic and biogeographic transition zone. Mar Ecol-Prog Ser. 2009;394:165–77.
- 133. Crooks JA, Chang AL, Ruiz GM. Aquatic pollution increases the relative success of invasive species. Biol Invasions. 2011;13(1):165–76.
- 134. Rainbow PS, Fialkowski W, Sokolowski A, Smith BD, Wolowicz M. Geographical and seasonal variation of trace metal bioavailabilities in the Gulf of Gdansk, Baltic Sea using mussels (Mytilus trossulus) and barnacles (Balanus improvisus) as biomonitors. Mar Biol. 2004;144(2):271–86.
- 135. Pigliucci M, Murren CJ, Schlichting CD. Phenotypic plasticity and evolution by genetic assimilation. J Exp Biol. 2006;209(12):2362–7.
- 136. Qiu JW, Qian PY. Tolerance of the barnacle Balanus amphitrite amphitrite to salinity and temperature stress: effects of previous experience. Mar Ecol-Prog Ser. 1999;188:123–32.
- 137. Bilton DT, Paula J, Bishop JDD. Dispersal, genetic differentiation and speciation in estuarine organisms. Estuar Coast Shelf Sci. 2002;55(6):937–52.