Geographic extent of introgression in Sebastes mentella and its effect on genetic population structure

Abstract Genetic population structure is often used to identify management units in exploited species, but the extent of genetic differentiation may be inflated by geographic variation in the level of hybridization between species. We identify the genetic population structure of Sebastes mentella and investigate possible introgression within the genus by analyzing 13 microsatellites in 2,562 redfish specimens sampled throughout the North Atlantic. The data support an historical divergence between the “shallow” and “deep” groups, beyond the Irminger Sea where they were described previously. A third group, “slope,” has an extended distribution on the East Greenland Shelf, in addition to earlier findings on the Icelandic slope. Furthermore, S. mentella from the Northeast Arctic and Northwest Atlantic waters are genetically different populations. In both areas, interspecific introgression may influence allele frequency differences among populations. Evidence of introgression was found for almost all the identified Sebastes gene pools, but to a much lower extent than suggested earlier. Greenland waters appear to be a sympatric zone for many of the genetically independent Sebastes groups. This study illustrates that the identified groups maintain their genetic integrity in this region despite introgression.


Sebastes viviparus Krøyer 1845 (Norwegian redfish). Sebastes mentella
is the most economically important species of the genus. It displays a high degree of genetic population diversity across its distributional range. In the Irminger Sea, two distinct groups, "shallow (=shallow pelagic)" (50-550 m) and "deep (=deep pelagic)" (550-800 m), have been reported (Cadrin et al., , 2011Stefansson et al., 2009). These two groups have been discriminated on the basis of morphological characteristics; the "deep" group has brighter red color, stouter appearance and larger size at sexual maturity . They have also been reported to show different rates of parasite infestation . Microsatellite DNA and morphological analyses suggested that these are two incipient species maintained by an ecological isolation mechanism, although evidence for hybridization was found (Stefánsson et al., 2009). The historical divergence between the "shallow" and "deep" groups was further supported by DNA analysis of the mitochondrial control region, microsatellites and the gene coding for the visual-pigment rhodopsin (Shum, Pampoulie, Kristinsson, & Mariani, 2015;Shum, Pampoulie, Sacchi, & Mariani, 2014). Around Iceland, an additional "slope" component of S. mentella has been described (e.g., . However, the genetic connectivity among these groups and their geographical distribution in the North Atlantic is not well understood. Sebastes species coexist in different marine areas and overlap in their depth range, normally between 100 and 400 m (Barsukov, Litvinenko, & Serebryakov, 1984). For instance, S. mentella cooccurs with S. norvegicus and S. fasciatus in the Northwest Atlantic (Roques et al., 2001), S. norvegicus in Greenland waters (ICES 2001), and S. norvegicus and S. viviparus in Icelandic and Norwegian waters (ICES 2001). Studies across the North Atlantic have indicated hybridization of S. mentella with S. norvegicus , S. viviparus (Artamonova et al., 2013) and S. fasciatus (Roques et al., 2001). However, except for the evidence of hybridization in the Northwest Atlantic (Roques et al., 2001), the geographic extent of hybridization has not been well studied, which partly relates to the uncertainty about Sebastes mating grounds. The waters off East Greenland are assumed to be the main nursery area for S. mentella and S. norvegicus juveniles, which are believed to be extruded along the Reykjanes Ridge (Anderson, 1984;Magnusson & Johannesson, 1995). Morphological species identification remains uncertain (particularly for fish <25 cm) in these geographical regions, possibly because of unrecognized hybridization (Johansen, 2003). Hybridization has also been suspected to affect estimates of genetic population structure within species (Artamonova et al., 2013). Specifically, introgression with S. viviparus has been indicated as one of the main reasons for the apparent differentiation between the "deep" and "shallow" groups of S. mentella (Artamonova et al., 2013). The role of hybridization on the genetic population structure is therefore still uncertain.
Despite the importance of Greenland waters as habitat and fisheries area for Sebastes species, the genetic population structure of S. mentella in this region has not yet been described in sufficient detail for fisheries management. For instance, the "slope" group in the East Greenland and Iceland waters are assessed separately by the International Council for the Exploration of the Sea (ICES 2015a) although it is known that the stock identity is uncertain. ICES perceives this as an interim status until the stock structure of S. mentella on the East Greenland slope is better understood. The objectives of this study were: I) to examine the genetic structure of S. mentella in Greenland, Iceland, and Irminger Sea waters, II) to investigate the extent of introgression among co-occurring Sebastes species/gene pools, and III) to assess whether introgression may influence apparent species and population structure in the region.

| Sampling
In total, 35 samples (collections of fish) consisting of 2,562 redfish specimens were included in the study (Table 1). Nineteen S. mentella samples were collected during 2011 and 2012 from Greenland waters by commercial fishing vessels and research surveys. All the fish were caught by trawl. Samples were from different seasons (spring and fall) and life stages (juveniles and adults). Norwegian samples were collected from the shelf and pelagic waters in different seasons of the years (Table 1). Sebastes mentella reference samples were included from Canadian, Icelandic (representing the Icelandic shelf component) waters, and Irminger Sea from the EU REDFISH project (1998)(1999)(2000)(2001).
The Irminger Sea reference samples were characterized as "deep" and "shallow" S. mentella based on both morphological characters and sampling depths as described by . We included reference samples of S. norvegicus, S. viviparus, and S. fasciatus from Greenland, Iceland, Norwegian, and Canadian waters (Table 1, Figure 1) to study introgression with S. mentella.
Species identification was conducted on board the ships based on morphological characters (e.g., body size, beak size, eye diameter, direction of spines in the pre-operculum) as suggested by Barsukov et al. (1984). The adults and juveniles were categorized by length (Barsukov et al., 1984) [for instance, in S. mentella, adults ≥29 cm and juveniles = (4-28) cm].

| Population cluster and individual admixture analyses
Population cluster analysis was applied for the 27 samples of S. mentella (cf. genotyped for ten loci (see rationale above). A Bayesian approach, as implemented in STRUCTURE (Pritchard, Stephens, & Donnelly, 2000), was used for clustering of genotypes and estimation of individual admixture proportion in an effort to identify possible hybrids (burn-in = 350,000, MCMC = 500,000, replication = 10 for each K).
The method clusters individuals to minimize Hardy-Weinberg and gametic phase disequilibria between loci within groups. STRUCTURE T A B L E 1 (Continued) may predict fewer than the actual number of clusters in a dataset with hierarchical structure (Kalinowski, 2010), so we ran the program using all samples followed by a cluster by cluster (hierarchical) approach.

Location
The Evanno method (Evanno, Regnaut, & Goudet, 2005) as implemented in STRUCTURE HARVESTER (Earl & vonHoldt, 2012) was used to estimate the number of clusters in each dataset.
The incidence of hybridization was further quantified by additional STRUCTURE analyses in clusters that indicated hybridization. To estimate the individual admixture proportions (Q = genome ancestry fraction) in STRUCTURE, an admixture model was used. Furthermore, a correlated allele frequency model was applied as the differentiation among the clusters was low (as suggested by Nielsen et al., 2003).

| Identification of hybrids
Hybridization of Sebastes in waters close to Greenland (Greenland, Irminger Sea, and Iceland) was compared with hybridization in the Northeast Arctic (Norwegian), and Northwest Atlantic waters (Flemish cap) following the approach of Nielsen et al. (2003). An individual was initially identified as possible hybrid if at least 10% of its genome (Q) originated from other groups (see Randi, 2008). The most pure parental individuals (Q ≥ 0.90) were used as base population for generating in silico "pure" individuals using the program HYBRIDLAB (Nielsen, Bach, & Kotlicki, 2006). Simple mechanical mixing was simulated by generating pure parental genotypes (i.e., randomly drawing alleles from the allele frequency distribution of the "pure population") equal to the numbers of individuals of the estimated clusters. A hybrid swarm (i.e., random mating between species/populations) was generated by random drawing of alleles from the observed allele frequency distribution of the combined species/populations. Differences between these simulated and empirical individual admixture proportions were tested with a Kolmogorov-Smirnov two-sample (K-S) test.

| Isolation with migration
The isolation-with-migration (IM) model as implemented in IMa2 (Hey, 2009) was applied to estimate introgression and population  Table 1) demographic parameters for the cluster pairs from Greenland waters. The analyses were conducted only for clusters in which hy- A few preliminary analyses were performed with wider priors.
The estimated effective sample size values of parameter t (time since divergence), swapping rates, autocorrelation value, trend-line plots were evaluated. After a burn-in of 500,000 steps, 2,000,000 more steps were run to save 20,000 genealogies for each pair of clusters.
To ensure convergence for the estimation, 150 chains were used for all of runs with high heating schemes ("−ha0.975 −hb0.75") so that the updates for the chains are accepted at higher rates. Introgression/ bidirectional migration rates (m1 and m2) were estimated as effective number of migrants per generation (2NM = 4N e μ*M/μ/2, independent of mutation rate), population size parameters (Θ = 4N e μ, N e = effective population size), and time since divergence in generations (tμ) without converting for a given mutation rate, and generation time.

| Descriptive statistics
The number of alleles per locus ranged from 13 to 69, and mean allelic richness (N = 17) varied between 3.51 and 20.49. Gene diversities per locus within samples were between 0.14 and 0.969 (mean = 0.77).
Significant heterozygote deficits were observed for all loci (most frequently in Greenland waters,

| Pattern of genetic differentiation
For the S. mentella population structure analysis (27 samples), an initial investigation identified 138 redfish as S. norvegicus, which were removed from the subsequent analyses. As no differentiation was observed between temporal samples from two different S. mentella locations (  (Figure 2 and Figure S1). In many cases, individuals from one sample assigned to different clusters. The clusters were sorted according to the occurrence of the reference samples as "shallow (sample 5)/deep (sample 6)/slope (sample 7)." The "shallow" cluster was the largest, including 80% of the "shallow" sampling group fish (M-Oc, Figure  from commercial vessels were found to be dominated by fish from the "slope" cluster. When the total dataset (i.e., 35 samples) was considered, six clusters were identified by STRUCTURE (Figure 3 and Figure S2). Three of the clusters identified within the S. mentella samples were consistent with the results described above. Two clusters were detected within the S. norvegicus samples, while S. fasciatus and S. viviparus samples were placed into a single cluster. However, with the application of hierarchical cluster approach, three clusters were detected within the S. norvegicus samples and S. fasciatus and S. viviparus samples were separated into two distinct clusters. For the S. norvegicus samples, three clusters were designated as "Norvegicus-A," "Norvegicus-B," and "Giants" (described in Saha et al., 2016). Mechanical mixing, as consistent with the observed deviations from the Hardy-Weinberg equilibrium, was therefore apparent in multiple samples. As no introgression was associated with the "Giants" cluster, the cluster was not used in the downstream analyses. Evidence of hybridization was observed for almost all possible cluster pairs (Figure 3, see details below).
The three-group configuration ("shallow-deep-slope") of S. mentella samples predicted by Bayesian approaches was supported by the AMOVA (F CT = 0.03, P = .000). A much lower, but still significant proportion of the genetic variance could be ascribed to differentiation among samples within groups (F SC = 0.002, P = .000). The S. mentella "shallow" group from Northwest Atlantic (FC) and Northeast Arctic (Nor) were significantly differentiated from each other, and from the Irminger Sea "shallow" group plus "shallow" group samples (except F I G U R E 2 Column charts illustrating individual cluster ancestries obtained from the STRUCTURE analysis (Pritchard et al., 2000) using 13 microsatellites. Three clusters were predicted for the 27 Sebastes mentella samples (cf. Table 1). Sample names are provided at the X-axis, while genome ancestry fractions (Q) are given at the Y-axis two) from Greenland waters (Table 3). In many of the cases, the "deep" groups identified from the Northwest Atlantic, Irminger Sea, Iceland, and Greenland waters were significantly differentiated from one another (Table 4). Finally, the "slope" group identified from the Northeast Arctic waters was significantly differentiated from Greenland waters (Table 5).

| Genetic diversity, Hardy-Weinberg and linkage equilibria within clusters
Genetic diversities, Hardy-Weinberg and linkage equilibria were estimated within the clusters with designated pure and hybrid (Q > 10% from other genomes) individuals identified through the admixture analyses. Clusters were formed with comparable numbers of fish (Table S3) to obtain similar statistical power. The cluster of pure S. fasciatus (Table S3: Fasciatus) showed significant heterozygote deficiency, only before FDR control. All the admixed clusters except for S. mentella and S. fasciatus, and the "deep" and "shallow" groups of S. mentella were out of Hardy-Weinberg equilibrium. However, none of them remained significant except the admixed cluster of S. mentella and S. viviparus ("Shallow" × "Viviparus") after FDR. The admixed clusters had higher gene diversities than one or both of the parental clusters (except "deep-shallow"). A total of 59 (6.39%) tests for deviations from linkage equilibrium within clusters were significant, with more deviations observed in the admixed clusters (eight in Mentella-Fasciatus, six in Norvegicus-A-B and five in "deep-shallow"). In Greenland waters, evidence of significant introgression was observed in all cluster pairs studied using the IM model ( Figure 4).

| Introgressive hybridization in the genus Sebastes
The highest introgression, which was also the most asymmetric, was observed between the clusters of "shallow" and "deep" S. mentella. Hybridization between the "shallow-slope" pair was just higher than for the "deep-slope" pair. In contrast, symmetric patterns of introgression of much lower levels were estimated between Norvegicus-A-"shallow" Mentella, Norvegicus-B-"shallow" Mentella, and Norvegicus-B-"deep" Mentella.
The coalescent-based analysis suggested that the divergence between "shallow" and "deep" groups was more ancient (Figure 5a) than that between the "shallow" and "slope" groups ( Figure 5b). The estimates of the population size parameter (Θ) were comparable for the three S. mentella clusters ( Figure 6).

| DISCUSSION
Using an unprecedented large number of target and reference samples from different years, seasons, and life stages, the present investigation provided improved population delineation on a larger geographic scale for the S. mentella complex. Our primary results illustrate the presence of three distinct genetic groups within the species across its distribution range. The "shallow" group had a wider distribution than the "deep" and "slope" groups. The occurrence of the "slope" group on the East Greenland Shelf was supported by adult fish collected in both spring and fall, as well as by juveniles. In contrast, the "slope" sample from the Icelandic Shelf was found to be a mixture of fish belonging to both "deep" and "slope" groups. Connectivity in terms of effective number of migrants between the "shallow-slope" pair was found to be greater than connectivity between the "deep-slope" pair. It was also evident that S. mentella from the Northeast Arctic and Northwest Atlantic waters are genetically differentiated. Low, but statistically significant, evidence of introgression was observed among the clusters.

| Genetic population structure of Sebastes mentella
The three groups of S. mentella observed in this study are in agreement with earlier findings (for review, see . The observation that the "shallow" group includes fish from all 27 geographical samples reveals the widest geographical distribution of this genetic group. Likewise, fish from inside and outside the Irminger Sea were assigned to the "deep" group. These observations are congruent with earlier findings (Shum et al., 2015). The F ST estimate between these two distinct genetic groups in the present investigation (F ST = 0.03) was higher than that estimated by Stefansson et al. (2009, F ST = 0.009) but almost identical to that of Shum et al. (2014, F ST = 0.031). Estimates of the demographic history by the IM model revealed that the divergence between these two groups was more ancient (tμ = 0.11) than that between "Giants" and "Norvegicus-B" (tμ = 0.06), supporting a deep evolutionary divergence between "shallow" and "deep" groups as suggested by Stefánsson et al. (2009), Shum et al. (2014. The "slope" group was mainly observed in East Greenland in catches from the commercial fleets. These samples consisted of adult fish caught during the larval extrusion period (spring) and of juveniles collected in autumn. The finding of juvenile fish is not surprising, because Greenland waters are suggested to be nursery grounds for all S. mentella and S. norvegicus groups of the region (Anderson, 1984;Magnusson & Johannesson, 1995). However, the presence of adult "slope" individuals suggests that East Greenland waters may also act as an important area of distribution for the adults of this group, which has not previously been recognized.
The finding that 29% of the individuals from the Icelandic Shelf sample were assigned to the "slope" group, but the remaining individuals clustered with "deep," indicates sympatric occurrence of different groups on the Icelandic Shelf. Previously, it has been indicated that the Icelandic Shelf was the main distributional area for the "slope" group . However, the results presented herein suggest that this distribution extends further into East Greenland waters. At present, the "slope" group S. mentella on the Greenland and Iceland Shelf waters are assessed separately by ICES. In light of the results presented here, this should be reconsidered as stock dynamics appear to be linked across continental slopes, and the effect of fishing intensity on either shelf could affect the entire stock unit. A similar scenario has also been described for Atlantic cod where the East Greenland and Iceland slopes are inhabited by the same populations , and for Greenland halibut in the entire region that is considered as a single unit (ICES 2015b).
The genetic differentiation observed for the "deep-slope" pair was the largest among the three S. mentella genetic groups (Table 2).
Congruent with this observation, the F ST estimates suggested the closest connectivity for the "shallow-slope" pair, which was further   (2003) and Daníelsdóttir et al. (2008) indicated closer relationship for "slopedeep" pair. The discrepancy might be associated with the types of marker applied, because markers can be of varied mode of inheritance, function, and statistical properties. Moreover, the allozyme studies considered each locus separately, rendering low power for estimating connectivity among the genetic groups.
Based on the results from the hierarchical variance analysis (AMOVA), there was evidence of statistically significant genetic variance among samples within groups. This result was also supported by the many significant pairwise F ST estimates between samples within groups. Most importantly, the identification of significant differentiation between the Northeast Arctic sample and other samples within the "shallow" group (Table 3)  in the Northeast Arctic is also supported by the finding of a separate larval extrusion and nursery ground in this region (e.g., . The finding that S. mentella from the Northwest Atlantic were T A B L E 3 Pairwise F ST values (Weir & Cockerham 1984) within Sebastes mentella "shallow" group (cf. Figure 2) N significantly differentiated from other S. mentella samples (Tables 3   and 4) supports the idea of a "Western stock" as proposed by . The highest substructuring was found within the "deep" group (Tables 3-5), indicating higher habitat segregation for the group in comparison with "shallow" and "slope." This observation is consistent with Shum et al. (2015) who hypothesized based on their mtDNA data that female "deep" S. mentella may exhibit some degree of philopatry.
Both the "deep" and "slope" groups were mainly concentrated in the central North Atlantic. Nevertheless, a significant number of fish from the Northwest Atlantic assigned to "deep" group, and some fish from Northeast Arctic ascribed to the "slope" group. The "shallow" group was distributed across the North Atlantic with evidence of variance among samples within the group. Such trans-Atlantic genetic structuring have been indicated for the entire S. mentella complex , but is genetically confirmed only for the "shallow" group (Shum et al., 2015), which could be associated with the less extensive sampling in earlier studies. However, trans-Atlantic patterns of genetic population structure have been observed in other species with comparable genetic structuring such as Atlantic cod (Bradbury et al., 2013;Hemmer-Hansen et al., 2013).   (Scribner, 1993), as seen when comparing clusters with pure and admixed individuals (Table S3). The reason why the apparently nonadmixed S. fasciatus cluster (Figure 3) still displayed such deviations could be due to the inclusion of undetected hybrids with other clusters not included in the baseline such as S. fasciatus-S. norvegicus. This was supported by the detection of admixed individuals in the simulation including Norwegian S. norvegicus as baseline. In contrast, a few apparently admixed clusters did not display such deviations, which could be associated with the low power of those tests (i.e., as suggested by Nielsen et al., 2003). Another indication of introgression is the elevated levels of polymorphism observed, which is expected for mixed gene pools (e.g., Roques et al., 2001).

| Introgression within the genus
The extent of introgression was greater among the three groups of S. mentella than among other clusters (Figure 4 and Figures S3-S16).
However, the magnitude observed in our study was less than those reported for S. mentella-S. norvegicus , or for S. mentella-S. viviparus (Artamonova et al., 2013 cf. Figure S7). A low level of introgression within the redfish from these regions has been reported by Saha et al. (2016) and Schmidt (2005).
The discrepancies with other studies might be associated with differences in statistical power among different studies. Incomplete or relaxed reproductive barriers among closely related species may provide opportunities for hybridization when in sympatry (Barton & Hewitt, 1989). Speciation within the genus Sebastes is a recent event (Briggs, 1995), and thus, these species may have relaxed reproductive barriers allowing hybridization to some extent (e.g., Roques et al., 2001).
Nevertheless, Sebastes are ovoviviparous and display particular mating behavior during copulation (Helvey, 1982;Kendall, 1991) which implies that large-scale introgression may require more than simple sympatric existence, as evident in the Gulf of St. Lawrence (Roques et al., 2001). (Weir & Cockerham 1984) within Sebastes mentella "slope" group (cf. Figure 2 The pattern of introgression between "shallow" and "deep" groups in Greenland waters appeared asymmetric (Figure 4). Similar observations have been made in other studies of Sebastes, and the roles of selection and differential abundance have been considered (Roques et al., 2001;Seeb, 1998). Considering the estimates of population size parameters in Greenland waters (Figure 6), more introgression would be expected from the "deep" group toward the "shallow," because smaller populations are more likely to be introgressed by more abundant populations over time (Arnold, Hamrick, & Bennett, 1993). However, the opposite was observed. The estimates of the population size were comparable with overlapping confidence intervals. Therefore, other factors than population size, such as selection, may be more likely as drivers of the asymmetric pattern of introgression. The spatial and temporal overlap between the groups during mating may also not be reflected by the population sizes per se.  (Roques et al., 2001).
Heterozygote deficiencies were apparent in some loci for several samples, consistent with observations using some of these loci in other studies (e.g., Roques et al., 2001;Schmidt, 2005). No null alleles or short allele dominances were detected. Furthermore, sampling was not always in the mating area or during the mating period, and morphological identification of these species is uncertain (Barsukov et al., 1984). Therefore, mechanical mixing was considered as likely explanation of the observed heterozygote deficiencies.
Although the present investigation provides evidence of introgression between the identified Sebastes gene pools in sympatry, it may be difficult in some cases to distinguish the signal of introgression from noise. STRUCTURE has a tendency to misclassify "pure" individuals as "hybrids" when low differentiation is observed between the gene pools (Bohling, Adams, & Waits, 2013). As discussed earlier, introgression was supported by other genetic analyses such as IM results and the magnitude of introgression observed in our study is less than those reported in other investigations (e.g., Artamonova et al., 2013;Stefánsson et al., 2009). It is evident that hybridization was not of similar magnitude throughout the North Atlantic. For instance, interspecific hybridization appeared to be larger in Norwegian waters, whereas very low magnitude of introgression was observed in other areas (e.g., Figures S6, S7, S10 and S12-S14). Inclusion of more loci might have provided more power for the hybridization analyses. However, the number of loci used in our hybridization analyses was larger than that used by Roques et al. (2001)  The generally wide confidence intervals for our demographic estimates may reflect that the data do not allow precise estimation of the demographic history of these species and populations. Only one pair of clusters was considered at a time, thereby ignoring the role of other contemporary clusters (i.e., "ghost populations"). This approach may limit the applicability of the IM method for providing realistic estimates of introgression and demographic parameters. However, no severe bias has been observed in other comparable studies (Chan et al., 2013;Jacobsen & Omland, 2012;Won & Hey, 2005), so possible bias in our findings is believed to be minor and have little impact on our main conclusions.
The SM model is only option for IM analyses of microsatellite data, but microsatellites probably do not follow this model. Strasburg and Rieseberg (2009)  have consequently exceeded biological advice by more than three times due to autonomous quota setting by each fishing nation (ICES 2015a). The further clarification of stock identity for S. mentella in the present study is anticipated to lead to a more robust advice which again likely will ensure a sustainable fishery by a common fishery management.

| CONCLUSIONS
We provide the first genetic investigation of S. mentella throughout its range, including Greenland waters. The identification of all three S. mentella groups in Greenland waters supports the interpretation of this region as a nursery area and population mixing zone. Fish from the "shallow" group were identified across the North Atlantic, but divided into three populations: Northeast Arctic, Irminger Sea/ Greenland, and Northwest Atlantic. Therefore, we clarify the genetic stock identity of S. mentella from the Northeast Arctic, which was previously disputed. The "deep" group S. mentella were identified in the Irminger Sea and Greenland waters, but some were also found in the Northwest Atlantic. The "slope" fish from the Icelandic Shelf and Greenland waters were not genetically differentiated, which suggests genetic connectivity of "slope" fish in East Greenland-Iceland Shelves.
In Greenland waters, the "slope" group was the main target for the commercial fleets in spring. Although genetic heterogeneity was evident, low-to-moderate extent of gene flow was observed across the North Atlantic, implying incomplete reproductive isolation for Sebastes clusters possibly due to their close evolutionary relationship.
The genetic differentiation between S. mentella groups were less than that between the other Sebastes clusters. Our findings mostly support the existing management practice of S. mentella throughout the North Atlantic (Cadrin et al., , 2011, which strengthen the population genetic basis of stock boundaries. The consideration of all stock identities for these species into the management practices will help to manage sustainable utilization of this important fishery resource.