Speciation and demographic history of Atlantic eels (Anguilla anguilla and A. rostrata) revealed by mitogenome sequencing

Processes leading to speciation in oceanic environments without obvious physical barriers remain poorly known. European and American eel (Anguilla anguilla and A. rostrata) spawn in partial sympatry in the Sargasso Sea. Larvae are advected by the Gulf Stream and other currents towards the European/North African and North American coasts, respectively. We analyzed 104 mitogenomes from the two species along with mitogenomes of other Anguilla and outgroup species. We estimated divergence time between the two species to identify major events involved in speciation. We also considered two previously stated hypotheses: one where the ancestral species was present in only one continent but was advected across the Atlantic by ocean current changes and another where population declines during Pleistocene glaciations led to increasing vicariance, facilitating speciation. Divergence time was estimated to ∼3.38 Mya, coinciding with the closure of the Panama Gateway that led to reinforcement of the Gulf Stream. This could have advected larvae towards European/North African coasts, in which case American eel would be expected to be the ancestral species. This scenario could, however, not be unequivocally confirmed by analyses of dN/dS, nucleotide diversity and effective population size estimates. Extended bayesian skyline plots showed fluctuations of effective population sizes and declines during glaciations, and thus also lending support to the importance of vicariance during speciation. There was evidence for positive selection at the ATP6 and possibly ND5 genes, indicating a role in speciation. The findings suggest an important role of ocean current changes in speciation of marine organisms.     

Disentangling structural genomic and behavioural barriers in a sea of connectivity

Genetic divergence among populations arises through natural selection or drift and is counteracted by connectivity and gene flow. In sympatric populations, isolating mechanisms are thus needed to limit the homogenizing effects of gene flow to allow for adaptation and speciation. Chromosomal inversions act as an important mechanism maintaining isolating barriers, yet their role in sympatric populations and divergence with gene flow is not entirely understood. Here, we revisit the question of whether inversions play a role in the divergence of connected populations of the marine fish Atlantic cod (Gadus morhua), by exploring a unique data set combining whole‐genome sequencing data and behavioural data obtained with acoustic telemetry. Within a confined fjord environment, we find three genetically differentiated Atlantic cod types belonging to the oceanic North Sea population, the western Baltic population and a local fjord‐type cod. Continuous behavioural tracking over 4 year revealed temporally stable sympatry of these types within the fjord. Despite overall weak genetic differentiation consistent with high levels of gene flow, we detected significant frequency shifts of three previously identified inversions, indicating an adaptive barrier to gene flow. In addition, behavioural data indicated that North Sea cod and individuals homozygous for the LG12 inversion had lower fitness in the fjord environment. However, North Sea and fjord‐type cod also occupy different depths, possibly contributing to prezygotic reproductive isolation and representing a behavioural barrier to gene flow. Our results provide the first insights into a complex interplay of genomic and behavioural isolating barriers in Atlantic cod and establish a new model system towards an understanding of the role of genomic structural variants in adaptation and diversification.     

THE GENETIC ARCHITECTURE OF REPRODUCTIVE ISOLATION DURING SPECIATION‐WITH‐GENE‐FLOW IN LAKE WHITEFISH SPECIES PAIRS ASSESSED BY RAD SEQUENCING

During speciation‐with‐gene‐flow, effective migration varies across the genome as a function of several factors, including proximity of selected loci, recombination rate, strength of selection, and number of selected loci. Genome scans may provide better empirical understanding of the genome‐wide patterns of genetic differentiation, especially if the variance due to the previously mentioned factors is partitioned. In North American lake whitefish (Coregonus clupeaformis), glacial lineages that diverged in allopatry about 60,000 years ago and came into contact 12,000 years ago have independently evolved in several lakes into two sympatric species pairs (a normal benthic and a dwarf limnetic). Variable degrees of reproductive isolation between species pairs across lakes offer a continuum of genetic and phenotypic divergence associated with adaptation to distinct ecological niches. To disentangle the complex array of genetically based barriers that locally reduce the effective migration rate between whitefish species pairs, we compared genome‐wide patterns of divergence across five lakes distributed along this divergence continuum. Using restriction site associated DNA (RAD) sequencing, we combined genetic mapping and population genetics approaches to identify genomic regions resistant to introgression and derive empirical measures of the barrier strength as a function of recombination distance. We found that the size of the genomic islands of differentiation was influenced by the joint effects of linkage disequilibrium maintained by selection on many loci, the strength of ecological niche divergence, as well as demographic characteristics unique to each lake. Partial parallelism in divergent genomic regions likely reflected the combined effects of polygenic adaptation from standing variation and independent changes in the genetic architecture of postzygotic isolation. This study illustrates how integrating genetic mapping and population genomics of multiple sympatric species pairs provide a window on the speciation‐with‐gene‐flow mechanism.     

The Bering Strait connection: dispersal and speciation in boreal macroalgae

A large number of boreal seaweeds have either sibling species or conspecific populations of a single species in the North Pacific and North Atlantic Oceans. This pattern is thought to have arisen from the dispersal between the two oceans through the Arctic Ocean after the opening of the Bering Strait in the mid-to-late Miocene or earliest Pliocene and from subsequent vicariant speciation as the Arctic Ocean froze and Bering Strait closed intermittently during glacial periods. Recent molecular studies of species in all three major seaweed phyla reveal patterns of vicariance. However, a number of lines of evidence point to differences in origins of these clades; some appear to be Pacific in origin whereas others appear to be derived from Atlantic stock. Different origins can be explained by recent stratigraphic finds that push the first Cenozoic opening of the Bering Strait back from 3.1–4.1 to 4.8–7.4 Ma (million years ago). Northern hemisphere ocean circulation models suggest that water flow would have been from the North Atlantic–Arctic south through the Bering Strait prior to the closure of the Panamanian Isthmus c. 3.5 Ma in contrast to the northward flow from the Pacific into the Arctic and North Atlantic, which developed after the closing of the Isthmus. Despite these differences in timing of the two invasions, there are no significant differences in levels of relationships among species with a North Atlantic origin compared with species with a North Pacific origin based on currently available data. More work is required to understand vicariance in seaweeds, especially in deciphering when a speciation event has occurred.     

Complex histories of repeated gene flow in Cameroon crater lake cichlids cast doubt on one of the clearest examples of sympatric speciation

One of the most celebrated examples of sympatric speciation in nature are monophyletic radiations of cichlid fishes endemic to Cameroon crater lakes. However, phylogenetic inference of monophyly may not detect complex colonization histories involving some allopatric isolation, such as double invasions obscured by genome‐wide gene flow. Population genomic approaches are better suited to test hypotheses of sympatric speciation in these cases. Here, we use comprehensive sampling from all four sympatric crater lake cichlid radiations in Cameroon and outgroups across Africa combined with next‐generation sequencing to genotype tens of thousands of SNPs. We find considerable evidence of gene flow between all four radiations and neighboring riverine populations after initial colonization. In a few cases, some sympatric species are more closely related to outgroups than others, consistent with secondary gene flow facilitating their speciation. Our results do not rule out sympatric speciation in Cameroon cichlids, but rather reveal a complex history of speciation with gene flow, including allopatric and sympatric phases, resulting in both reproductively isolated species and incipient species complexes. The best remaining non‐cichlid examples of sympatric speciation all involve assortative mating within microhabitats. We speculate that this feature may be necessary to complete the process of sympatric speciation in nature.     

THE EVOLUTIONARY GENETIC STATUS OF ICELANDIC EELS

The Iceland population of Anguilla eels contains an elevated frequency of fish with vertebral numbers lower than those typical of European localities. Several distinct hypotheses have been advanced to account for these morphologically atypical fish: for example, they could represent (1) genetically “pure” American expatriates, (2) genetically “pure” European types with ontogenetic abnormalities, or (3) hybrids between American and European forms. Here we critically test these and other possibilities by examining the joint distributions of allozyme markers, mitochondrial DNA markers, and vertebral numbers in Icelandic eels. The particular patterns of association among the genetic and morphological traits demonstrate that the Iceland population includes, in low frequency, the products of hybridization between American and European eels. Approximately 2–4% of the gene pool in the Iceland eel population is derived from American eel ancestry. This hybrid zone is highly unusual in the biological world, because the mating events in catadromous eels presumably take place thousands of kilometers from where the hybrids are observed as maturing juveniles. The molecular data, in conjunction with the geographic distributions, strongly suggest that the differences in migrational behavior and morphology between American and European eels include an important additive genetic component. Evolutionary hypotheses are advanced to account for the original separation of North Atlantic eels into American and European populations, and for the presence of hybrids in Iceland.     

Assessing patterns of hybridization between North Atlantic eels using diagnostic single-nucleotide polymorphisms

The two North Atlantic eel species, the European eel (Anguilla anguilla) and the American eel (Anguilla rostrata), spawn in partial sympatry in the Sargasso Sea, providing ample opportunity to interbreed. In this study, we used a RAD (Restriction site Associated DNA) sequencing approach to identify species-specific diagnostic single-nucleotide polymorphisms (SNPs) and design a low-density array that combined with screening of a diagnostic mitochondrial DNA marker. Eels from Iceland (N=159) and from the neighboring Faroe Islands (N=29) were genotyped, along with 94 larvae (49 European and 45 American eel) collected in the Sargasso Sea. Our SNP survey showed that the majority of Icelandic eels are pure European eels but there is also an important contribution of individuals of admixed ancestry (10.7%). Although most of the hybrids were identified as F1 hybrids from European eel female × American eel male crosses, backcrosses were also detected, including a first-generation backcross (F1 hybrid × pure European eel) and three individuals identified as second-generation backcrosses originating from American eel × F1 hybrid backcrosses interbreeding with pure European eels. In comparison, no hybrids were observed in the Faroe Islands, the closest bodies of land to Iceland. It is possible that hybrids show an intermediate migratory behaviour between the two parental species that ultimately brings hybrid larvae to the shores of Iceland, situated roughly halfway between the Sargasso Sea and Europe. Only two hybrids were observed among Sargasso Sea larvae, both backcrosses, but no F1 hybrids, that points to temporal variation in the occurrence of hybridization.     

Extended phylogeny of Aquilegia: the biogeographical and ecological patterns of two simultaneous but contrasting radiations

Studies of the North American columbines (Aquilegia, Ranunculaceae) have supported the view that adaptive radiations in animal-pollinated plants proceed through pollinator specialisation and floral differentiation. However, although the diversity of pollinators and floral morphology is much lower in Europe and Asia than in North America, the number of columbine species is similar in the three continents. This supports the hypothesis that habitat and pollinator specialisation have contributed differently to the radiation of columbines in different continents. To establish the basic background to test this hypothesis, we expanded the molecular phylogeny of the genus to include a representative set of species from each continent. Our results suggest that the diversity of the genus is the result of two independent events of radiation, one involving Asiatic and North American species and the other involving Asiatic and European species. The ancestors of both lineages probably occupied the mountains of south-central Siberia. North American and European columbines are monophyletic within their respective lineages. The genus originated between 6.18 and 6.57 million years (Myr) ago, with the main pulses of diversification starting around 3 Myr ago both in Europe (1.25–3.96 Myr ago) and North America (1.42–5.01 Myr ago). The type of habitat occupied shifted more often in the Euroasiatic lineage, while pollination vectors shifted more often in the Asiatic-North American lineage. Moreover, while allopatric speciation predominated in the European lineage, sympatric speciation acted in the North American one. In conclusion, the radiation of columbines in Europe and North America involved similar rates of diversification and took place simultaneously and independently. However, the ecological drivers of radiation were different: geographic isolation and shifts in habitat use were more important in Europe while reproductive isolation linked to shifts in pollinator specialisation additionally acted in North America.     

Cryptic speciation and genetic structure of widely distributed brittle stars (Ophiuroidea) in Europe

Pérez‐Portela, R., Almada, V. & Turon, X. (2012). Cryptic speciation and genetic structure of widely distributed brittle stars (Ophiuroidea) in Europe. —Zoologica Scripta, 00, 000–000. The development of molecular techniques has led to the detection of numerous cases of cryptic speciation within widely distributed marine invertebrate species and important taxonomic revisions in all the major marine taxa. In this study, we analysed a controversial marine species complex in the genus Ophiothrix, a widespread taxon in European waters traditionally assigned to two nominal species, Ophiothrix fragilis and O. quinquemaculata. These species are important components of the rocky shores and deep marine benthos along the North Atlantic and Mediterranean littoral. Their status (including variants of both species) has remained contentious due to overlapping variability in morphological characters. In this study, we analysed the genetic and morphological differences of Ophiothrix lineages along the Atlantic and Mediterranean coasts. We also assessed population genetic structure in the Atlantic and Mediterranean basins by sequencing two mitochondrial genes, the 16S rRNA gene and COI gene, of 221 specimens from 13 locations. Phylogenetic analyses demonstrated the existence of two genetically distinct lineages, attributable to two different species although unrelated to previous taxonomic distinctions. Morphological differences could also be detected between these lineages. Samples from the Northeast Atlantic and one from the deep Mediterranean grouped within Lineage I, whereas Lineage II pooled together the southern Atlantic and rocky shallow Mediterranean samples. In the northern region of the Iberian Peninsula and at a deep locality in the Mediterranean, both lineages overlap. Speciation processes likely happened during the Mio–Pliocene transition (about 4.8–7.5 million years ago), when marine‐level oscillations led to the blockage of major marine corridors in Europe and promoted genetic isolation by vicariance. Secondary contact between lineages following sea‐level increases and recolonization during the refilling of the Mediterranean after the Miocene salinity crisis could explain the present‐day distribution of genetic variability. No barriers to gene flow along the Atlanto‐Mediterranean area were detected for Lineage II, and the lack of genetic structure could be caused by a mixture of several factors, such as wide dispersal potential, recent demographic expansion and large population size.     

Geographic variation in the time‐frequency characteristics of high‐frequency whistles produced by killer whales (Orcinus orca)

Investigating intraspecific variation in acoustic signals can indicate the extent of isolation and divergence between populations and adaptations to local environments. Here we analyze the variation in killer whale high‐frequency (>17 kHz) whistles recorded off Norway, Iceland, and in the North Pacific. We used a combination of methods including multivariate comparisons of spectral and temporal parameters and categorization of contours to types. Our results show that spectral and temporal characteristics of high‐frequency whistles recorded in the North Pacific show significant differences from whistles recorded in the Northeast Atlantic, being generally stereotyped, lower in frequency, and slightly longer in duration. Most high‐frequency whistles from the North Pacific were downsweeps, whereas this was one of the least common types recorded in the Northeast Atlantic. The repertoire of whistles recorded in Norway was similar to Iceland, but whistles produced in Norway had significantly lower maximum frequency and frequency range. Most methods were able to discriminate between whistles of the North Pacific and the Northeast Atlantic, but were unable to consistently distinguish whistles from Iceland and Norway. This suggests that macro‐ and microgeographic differences in high‐frequency whistles of killer whales may reflect historical geographic isolation between ocean basins and more recent divergence between adjacent populations.     

Genome‐wide single‐generation signatures of local selection in the panmictic European eel

Next‐generation sequencing and the collection of genome‐wide data allow identifying adaptive variation and footprints of directional selection. Using a large SNP data set from 259 RAD‐sequenced European eel individuals (glass eels) from eight locations between 34 and 64^oN, we examined the patterns of genome‐wide genetic diversity across locations. We tested for local selection by searching for increased population differentiation using F_ST‐based outlier tests and by testing for significant associations between allele frequencies and environmental variables. The overall low genetic differentiation found (F_ST = 0.0007) indicates that most of the genome is homogenized by gene flow, providing further evidence for genomic panmixia in the European eel. The lack of genetic substructuring was consistent at both nuclear and mitochondrial SNPs. Using an extensive number of diagnostic SNPs, results showed a low occurrence of hybrids between European and American eel, mainly limited to Iceland (5.9%), although individuals with signatures of introgression several generations back in time were found in mainland Europe. Despite panmixia, a small set of SNPs showed high genetic differentiation consistent with single‐generation signatures of spatially varying selection acting on glass eels. After screening 50 354 SNPs, a total of 754 potentially locally selected SNPs were identified. Candidate genes for local selection constituted a wide array of functions, including calcium signalling, neuroactive ligand–receptor interaction and circadian rhythm. Remarkably, one of the candidate genes identified is PERIOD, possibly related to differences in local photoperiod associated with the >30° difference in latitude between locations. Genes under selection were spread across the genome, and there were no large regions of increased differentiation as expected when selection occurs within just a single generation due to panmixia. This supports the conclusion that most of the genome is homogenized by gene flow that removes any effects of diversifying selection from each new generation.     

Hidden histories of gene flow in highland birds revealed with genomic markers

Genomic studies are revealing that divergence and speciation are marked by gene flow, but it is not clear whether gene flow has played a prominent role during the generation of biodiversity in species‐rich regions of the world where vicariance is assumed to be the principal mode by which new species form. We revisit a well‐studied organismal system in the Mexican Highlands, Aphelocoma jays, to test for gene flow among Mexican sierras. Prior results from mitochondrial DNA (mtDNA) largely conformed to the standard model of allopatric divergence, although there was also evidence for more obscure histories of gene flow in a small sample of nuclear markers. We tested for these ‘hidden histories’ using genomic markers known as ultraconserved elements (UCEs) in concert with phylogenies, clustering algorithms and newer introgression tests specifically designed to detect ancient gene flow (e.g. ABBA/BABA tests). Results based on 4303 UCE loci and 2500 informative SNPs are consistent with varying degrees of gene flow among highland areas. In some cases, gene flow has been extensive and recent (although perhaps not ongoing today), whereas in other cases there is only a trace signature of ancient gene flow among species that diverged as long as 5 million years ago. These results show how a species complex thought to be a model for vicariance can reveal a more reticulate history when a broader portion of the genome is queried. As more organisms are studied with genomic data, we predict that speciation‐with‐bouts‐of‐gene‐flow will turn out to be a common mode of speciation.     

Palaeobiogeography of the North Pacific toothed mysticetes (Cetacea,Aetiocetidae): a key to Oligocene cetacean distributional patterns

Biogeographical distributional patterns of cetaceans reflect dispersal events and colonization of the oceans from their ancestral area in the ancient Sea of Tethys ~53 Ma. Likewise, they reveal several vicariance events throughout the evolutionary history of this group. However, our understanding of how these processes took place and what biogeographical scenarios occurred among the different groups of cetaceans through time is limited. Consequently, this work focuses on explaining the distributional patterns of the well‐known North Pacific toothed mysticetes, Aetiocetidae, through the power of retrodiction offered by track analysis (panbiogeography) and cladistic biogeography, using the approach of evolutionary biogeography. Our results show that the distributional patterns of Aetiocetidae explain their endemism in the North Pacific, as well as indicating that their hypothetical ancestor probably colonized the Pacific from the Atlantic Ocean by a dispersal event (founder effect) via the Central American Seaway. Furthermore, their biogeographical history shows that the adaptive radiation (cladogenesis) of Aetiocetidae is result of peripatric speciation followed by sympatric speciation within a heterogeneous environment. Finally, the biogeographical framework of Aetiocetidae further supports the relevant role that the Pacific Ocean has played in the evolution of Oligocene cetaceans as a geographical area that promoted endemism, dispersal and colonization. At more local scales, environmental conditions further promoted increased diversity and disparity amongst Mysticeti.     

Natural hybrids in Atlantic eels (Anguilla anguilla, A. rostrata): evidence for successful reproduction and fluctuating abundance in space and time

The outcome of natural hybridization is highly variable and depends on the nonexclusive effects of both pre- and post-mating reproductive barriers. The objective of this study was to address three specific questions regarding the dynamics of hybridization between the American and European eels (Anguilla rostrata and Anguilla anguilla). Using 373 AFLP loci, 1127 eels were genotyped, representing different life stages from both continents, as well as multiple Icelandic locations. We first evaluated the extent of hybridization and tested for the occurrence of hybrids beyond the first generation. Second, we tested whether hybrids were randomly distributed across continents and among Icelandic sampling sites. Third, we tested for a difference in the proportion of hybrids between glass eel and yellow eel stages in Iceland. Our results provided evidence for (i) an overall hybrid proportion of 15.5% in Iceland, with values ranging from 6.7% to 100% depending on life stages and locations; (ii) the existence of hybrids beyond the first generation; (iii) a nonrandom geographic distribution of hybrids in the North Atlantic; and (iv) a higher proportion of first and later generation hybrids in yellow eels compared to glass eels, as well as a significant latitudinal gradient in the proportion of hybrids in Icelandic freshwater. We propose that the combined effect of both differential survival of hybrids and variation in hybridization rate through time best explain these patterns. We discuss the possibility that climate change, which is impacting many environmental features in the North Atlantic, may have a determinant effect on the outcome of natural hybridization in Atlantic eels.     

Genomic divergence between the migratory and stationary ecotypes of Atlantic cod

Atlantic cod displays a range of phenotypic and genotypic variations, which includes the differentiation into coastal stationary and offshore migratory types of cod that co‐occur in several parts of its distribution range and are often sympatric on the spawning grounds. Differentiation of these ecotypes may involve both historical separation and adaptation to ecologically distinct environments, the genetic basis of which is now beginning to be unravelled. Genomic analyses based on recent sequencing advances are able to document genomic divergence in more detail and may facilitate the exploration of causes and consequences of genome‐wide patterns. We examined genomic divergence between the stationary and migratory types of cod in the Northeast Atlantic, using next‐generation sequencing of pooled DNA from each of two population samples. Sequence data was mapped to the published cod genome sequence, arranged in more than 6000 scaffolds (611 Mb). We identified 25 divergent scaffolds (26 Mb) with a higher than average gene density, against a backdrop of overall moderate genomic differentiation. Previous findings of localized genomic divergence in three linkage groups were confirmed, including a large (15 Mb) genomic region, which seems to be uniquely involved in the divergence of migratory and stationary cod. The results of the pooled sequencing approach support and extend recent findings based on single‐nucleotide polymorphism markers and suggest a high degree of reproductive isolation between stationary and migratory cod in the North‐east Atlantic.     

PSMC (pairwise sequentially Markovian coalescent) analysis of RAD (restriction site associated DNA) sequencing data

The pairwise sequentially Markovian coalescent (PSMC) method uses the genome sequence of a single individual to estimate demographic history covering a time span of thousands of generations. Although originally designed for whole‐genome data, we here use simulations to investigate its applicability to reference genome‐aligned restriction site associated DNA (RAD) data. We find that RAD data can potentially be used for PSMC analysis, but at present with limitations. The key factor is the proportion (p) of the genome that the RAD data covers. In our simulations, a proportion of 10% can still retain a substantial amount of coalescent information, whereas for 1% estimation becomes unreliable. The performance depends strongly on mutation rate (μ) and recombination rate (r) and is proportional to μ*p/r. When the value of this term is low, increasing the amount of data and number of iterations helps restoring the power of the estimation. We subsequently analyse one whole‐genome‐sequenced and 17 RAD‐sequenced three‐spined sticklebacks (Gasterosteus aculeatus) from a lake in Greenland. The whole‐genome sequence suggests a relatively recent expansion and decline within ca. 4000–40 000 generations ago, possibly reflecting postglacial expansion and founding of the lake population. RAD data, where chromosomes from 10 individuals are combined, identify a similar pattern. Our study provides guidance about the use of PSMC analysis and suggests measures that can improve its utility for RAD data. Finally, the study shows that RAD loci in general contain coalescent information that can be used for developing more targeted methods.     

Biogeography of Asterias: North Atlantic climate change and speciation

Fossil evidence suggests that the seastar genus Asterias arrived in the North Atlantic during the trans-Arctic interchange around 3.5 Ma. Previous genetic and morphological studies of the two species found in the Atlantic today suggested two possible scenarios for the speciation of A. rubens and A. forbesi. Through phylogenetic and population genetic analysis of data from a portion of the cytochrome oxidase I mitochondrial gene and a fragment of the ribosomal internal transcribed spacer region, I show that the formation of the Labrador Current 3.0 Ma was probably responsible for the initial vicariance of North Atlantic Asterias populations. Subsequent adaptive evolution in A. forbesi was then possible in isolation from the European species A. rubens. The contact zone between these two species formed recently, possibly due to a Holocene founding event of A. rubens in New England and the Canadian Maritimes.     

The North Atlantic subpolar gyre and the marine migration of Atlantic salmon Salmo salar: the ‘Merry‐Go‐Round’ hypothesis

One model for marine migration of Atlantic salmon Salmo salar proposes that North American and southern European stocks (<62° N) move directly to feeding grounds off west Greenland, then overwinter in the Labrador Sea, whereas northern European stocks (>62° N) utilize the Norwegian Sea. An alternate model proposes that both North American and European stocks migrate in the North Atlantic Subpolar Gyre (NASpG) where S. salar enter the NASpG on their respective sides of the Atlantic, and travel counterclockwise within the NASpG until returning to natal rivers. A review of data accumulated during the last 50 years suggests a gyre model is most probable. Freshwater parr metamorphose into smolts which have morphological, physiological and behavioural adaptations of epipelagic, marine fishes. Former high‐seas fisheries were seasonally sequential and moved in the direction of NASpG currents, and catches were highest along the main axis of the NASpG. Marking and discrimination studies indicate mixed continental origin feeding aggregations on both sides of the Atlantic. Marked North American smolts were captured off Norway, the Faroe Islands, east and west Greenland, and adults tagged at the Faroes were recovered in Canadian rivers. Marked European smolts were recovered off Newfoundland and Labrador, west and east Greenland, and adults tagged in the Labrador Sea were captured in European rivers. High Caesium‐137 (¹³⁷Cs) levels in S. salar returning to a Quebec river suggested 62·3% had fed at or east of Iceland, whereas levels in 1 sea‐winter (SW) Atlantic Canada returnees indicated 24·7% had fed east of the Faroes. Lower levels of ¹³⁷ Cs in returning 1SW Irish fish suggest much of their growth occurred in the western Atlantic. These data suggest marine migration of S. salar follows a gyre model and is similar to other open‐ocean migrations of epipelagic fishes.