680
Depth as a potential driver of genetic structure of Sebastes
mentella across the North Atlantic Ocean
Magnús Örn Stefánsson, Jákup Reinert, Þorsteinn Sigurðsson, Kristján Kristinsson, Kjell Nedreaas,
and Christophe Pampoulie
Stefánsson, M. Ö., Reinert, J., Sigurðsson, Þ., Kristinsson, K., Nedreaas, K., and Pampoulie, C. 2009. Depth as a potential driver of genetic structure
of Sebastes mentella across the North Atlantic Ocean. – ICES Journal of Marine Science, 66: 680 –690.
A primary question that remains to be answered about the fishery for Sebastes mentella is whether it exploits several stocks within the
North Atlantic Ocean. To address this issue, 1240 redfish were collected from several fishing grounds in late 2006 and early 2007 and
genotyped at 12 microsatellite loci. Contemporary allele frequencies were compared with archived data to examine the temporal stability of the genetic structure. The analyses all revealed the presence of three genetically distinguishable clusters, which persisted
throughout the study period, suggesting that the genetic structure detected is genuine. Cluster D included fish from the deep
Irminger Sea and west Faroe Islands, and Cluster I included fish only from the Icelandic shelf. All other fish grouped in a third
cluster (S). Further analyses revealed that the genetic pattern observed was not primarily attributable to isolation by distance, but
rather to depth distribution.
Keywords: microsatellite loci, migration, North Atlantic Ocean, redfish, Sebastes mentella.
Received 6 October 2008; accepted 19 February 2009; advance access publication 24 March 2009.
M. Ö. Stefánsson, Þ. Sigurðsson, K. Kristinsson, and C. Pampoulie: Marine Research Institute, Skúlagata 4, 101 Reykjavı́k, Iceland. J. Reinert: Faroese
Fisheries Laboratory, Nóatún 1, PO Box 3051, FO-110 Tórshavn, Faroe Islands. K. Nedreaas: Institute of Marine Research, PO Box 1870, Nordnes,
5817 Bergen, Norway. Correspondence to C. Pampoulie: tel: þ354 575 2038; fax: þ354 575 2001; e-mail: [email protected].
Introduction
The exploitation of many marine fishery resources exceeds the
limits of sustainable harvesting, and several stocks have declined
drastically during the past 50 years (Christensen et al., 2003;
Morato et al., 2006; FAO, 2007). The depletion and the collapse
of commercial fish stocks have been interpreted as a failure of sustainable fisheries management, which has in turn been suggested
to be linked to the mismatch between management and actual biological units (Stephenson and Kenchington, 2000). Fish stocks are
often managed under the panmixia hypothesis, without consideration of life history, behavioural, or genetic differences among
components of those management units. Such management procedures could have detrimental effects on stock diversity by disproportionately affecting the less productive components of the
stock (Begg et al., 1999; Ward, 2000). Information on population
structure of commercially exploited fish stocks is therefore crucial
for conservation and sustainable management of marine fish
stocks (Hilborn et al., 2003). Genetic markers such as microsatellite loci have consequently been used increasingly to improve stock
identification and to aid fisheries management, with some success
(Rico et al., 1997; Ruzzante et al., 1997; Lage et al., 2004;
Pampoulie et al., 2006, 2008; Gharrett et al., 2007; Westgaard
and Fevolden, 2007; Hyde et al., 2008). Within the genus
Sebastes, however, cryptic species (species distinguished only
with genetic markers) have been exploited as a single management
unit for decades (Danı́elsdóttir et al., 2008; Hyde et al., 2008).
The deep-sea redfish Sebastes mentella (Travin, 1951) is slowgrowing, matures at an age of 10 –15 years and can live for
60 –70 years (Magnússon and Magnússon, 1995). Following
internal fertilization, females carry the fertilized eggs and larvae
internally until they are extruded to undergo a planktonic larval
phase. Sebastes mentella has a wide distribution; it is found
along the Norwegian coast up to Spitsbergen and in the western
Barents Sea, along the northern slope of the North Sea, around
the Faroe Islands, around Iceland, and in the Irminger Sea,
along the Greenland coast and along the North American coast
south to the Flemish Cap, the Grand Banks, and the Gulf of
St Lawrence (Tåning, 1949; Magnússon and Magnússon, 1995).
The species has been historically important commercially. In
the main fishing area (the Irminger Sea), fisheries of the former
Soviet Union started in 1982. Other nations then entered the
fishery, leading to an increase in total landings (Sigurðsson
et al., 2006). Then, from 1992 to 1994, fishing effort shifted out
to greater depth, where larger fish dominated the catch
(Sigurðsson et al., 2006). It is now acknowledged that the
Irminger Sea stock, as well as other stocks (Greenland, Faroe,
and Norway), are overexploited. Although S. mentella have been
known in Norwegian international waters for several years (ICES
2007, 2008; Vinnichenko, 2007), reporting on the distribution
within these waters has been limited. It is thought that abundance
has been low and distribution sporadic owing to the bycatches
being very small and, until recently, only reported during the
extensive herring (Clupea harengus) and blue whiting
(Micromesistius poutassou) fisheries, and research surveys in the
area (KN, pers. obs.). Fishing for S. mentella in Norwegian international waters began in 2004 and effort has been considerable,
# 2009 International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved.
For Permissions, please email: [email protected]
Depth and genetic structure of Sebastes mentella across the North Atlantic
although effort regulations have been in place since 2006 (ICES,
2008).
During the past few decades, one of the main challenges to S.
mentella fishery management has been to assess whether it exploits
only one stock or several stocks (Sigurðsson et al., 2006). This
resulted in part from the morphology-based description of two
forms (oceanic and deep-sea) of pelagic S. mentella within the
Irminger Sea (Magnússon, 1972; Magnússon and Magnússon,
1995). The two forms have somewhat overlapping geographical
distribution, but they differ in their depth distributions (see
Danı́elsdóttir et al., 2008; Stefánsson et al., in press). The
deep-sea form lives deeper and is more common in the northeast
part of the species’ distribution, whereas the oceanic form lives in
shallower water and has a more south –southwest distribution
(ICES, 2001). Recent genetic studies have suggested that the two
forms are reproductively isolated and represent two genetically
distinct populations (Johansen et al., 2000; Stefánsson et al.,
2004, in press; Danı́elsdóttir et al., 2008; Pampoulie and
Danı́elsdóttir, 2008). At a larger geographic scale, a recent study
of microsatellite loci detected three geographically distinct populations in the North Atlantic: eastern (Barents Sea and Norway),
panoceanic (Faroe Islands to the Grand Banks) and western
(Gulf of St Lawrence and the Laurentian Channel). Within the
panoceanic population, genetic differences were small and there
was no isolation by distance among the samples, which were collected shallower than 500 m and over a distance of more than
6000 km (Roques et al., 2002).
The aim of the present study is to assess genetically whether S.
mentella fishing activity exploits one or more stocks within the
North Atlantic. A total of 12 samples of S. mentella was collected
from several fishing grounds in late 2006 and early 2007, and genotyped at 12 microsatellite loci. In addition, archive data from ten
locations were included to examine the temporal stability of the
genetic pattern we observed.
Material and methods
Sampling
Genetic samples were taken from a total of 1240 fish sampled in
late 2006 or early 2007 (Table 1, Figure 1). The temporal stability
of the genetic pattern was examined in two ways. First, archive
samples collected at ten locations in 1995, 1996, 2000, and 2001
were compared with the contemporary samples (2006 –2007;
also referred to as current samples) to examine temporal stability
among and within the clusters detected (Stefánsson et al., 2004).
Second, one of the deep-sea archive samples (d1: collected in the
Irminger Sea in 1995) was processed in all analyses alongside the
contemporary samples.
Genetic analysis
DNA was extracted using chelex (Walsh et al., 1991). Current
samples were screened for variation at 12 microsatellite loci
(SEB9, SEB25, SEB31, SEB33, SEB45, Roques et al., 1999; Sal1,
Sal4, Miller et al., 2000; Smen5, Smen10, Stefánsson et al., in
press; Spi4, Spi10, Gomez-Uchida et al., 2003; Sal3, Miller et al.,
2000). Forward primers were labelled with VIC, NED, FAM, or
PET fluorescent label (Applied Biosystems). Polymerase chain
reaction (PCR) was performed in a reaction volume of 10 ml consisting of 3 ml of 1/10 diluted DNA template, 0.2 ml Teg DNA
polymerase (3 U/ml; product number pol-141, Prokaria Ltd,
Gyflaflöt 5 112 Reykjavı́k, Iceland), 1.0 ml of 10 buffer, 0.8 ml
681
dNTP (10 mM), reverse and forward primers in 0.05– 0.125 ml
(100 mM), and dH2O. PCR consisted of a 4-min denaturation at
948C followed by 30 cycles of 948C denaturing for 50 s, 588C
annealing for 50 s (all primers), and 728C extension for 90 s.
Cycling concluded with a 7-min extension at 728C. Amplified
DNA fragments were separated by an Autosequencer ABI 3730
Genetic Analyser, and were sized according to the Applied
Biosystems GeneScanTM —500LIZTM size standard. Alleles were
scored manually with the Genemapper Analysis Software version
4.0 (Applied Biosystems).
Archive data consisted of allelic information at nine of the
above loci (SEB9, SEB25, SEB31, SEB33, SEB45, Sal1, Sal3,
Smen5, and Smen10). DNA was extracted using phenolchloroform (Sambrook and Russell, 2001) or chelex methods
(Walsh et al., 1991). All loci except SEB25 and SEB31 were amplified separately. PCR and genotyping was performed as described in
Pampoulie and Danı́elsdóttir (2008). Primer concentration varied
from 1.0 mM for SEB9, SEB33, Smen5, Smen10, Sal1, and Sal3 to
0.4 mM for SEB25, 1.6 mM for SEB31, and 0.7 mM for SEB45.
Sample d1 was used to calibrate fragment sizes between current
and archived datasets. Calibration accuracy ranged from 100%
(Sal1, Sal3, and Smen10) through 99.99% (SEB31 smaller allele
and Smen5 smaller allele), 99.98% (SEB9, SEB25, and SEB33
larger allele), and 99.97% (SEB33 smaller allele, SEB45, and
Smen5 larger allele) to 99.95% (SEB31 larger allele).
Data analysis
Genetic variability was assessed using allele frequencies, observed
(HO) and expected (HE) heterozygosity, and mean number of
alleles, calculated in the program FSTAT version 2.9.3 (Goudet,
1995). FSTAT was also used to estimate overall FST (Weir and
Cockerham, 1984) and pairwise FST between samples, to test for
fit to Hardy–Weinberg proportions (HWE), to test for linkage
disequilibrium between loci, to calculate allelic richness (r), and
to test r and HO among genetic clusters. Always, 15 000 permutations were used for significance testing.
Exact tests for significant allelic and genotypic differentiation
between samples were carried out using the Markov chain
method (10 000 dememorization numbers, 100 batches, and
10 000 iterations per batch; Guo and Thompson, 1992) in the
program GENEPOP version 3.4 (Raymond and Rousset, 1995).
Multidimensional scaling analysis (MDS) based on pairwise FST
values was carried out in the R package (Ross and Gentleman,
1996), to visualize relationships among samples.
The program BAPS version 4.13 (Corander et al., 2003, 2004)
was used to cluster groups of fish, where the original samples
were defined as groups. The program was run using the nonspatial model for genetic discontinuities, i.e. population inference
was based on genotypes, and spatial location of the samples was
ignored. The maximum number of clusters was set at 13, equal
to the number of samples. To avoid the risk that the
algorithm could have fallen into a local mode, the program was
run at K ¼ 3, 6, 9, 12, and 13 for five replicates each. Individual
admixture proportions (q) were calculated based on mixture clustering after 1000 iterations. A post hoc analysis was carried out on
the distribution of individual q values (response) from samples
(factor) within different clusters using the Kruskal –Wallis nonparametric analysis of variance (Sokal and Rohlf, 1995).
A hierarchical analysis of molecular variance (AMOVA) was
performed using ARLEQUIN version 3.0 (Excoffier et al., 2005)
to test for spatial structure among ad hoc defined clusters
682
M. Ö. Stefánsson et al.
Table 1. Sample number, month and year of capture, trawl depth range, position, and number of individuals (n) of S. mentella investigated
from the North Atlantic current and archive data.
No.
Habitat
Current data
1
Irminger Sea archivea
2
Icelandic shelf west
3
Irminger Sea deep northeast
4
Irminger Sea deep northeast
5
Irminger Sea shallow southwest
6
7
Irminger Sea shallow southwest
Faroe Islands west
8
Faroe Islands east
9
Faroe Islands east
10
11
Norwegian shelf
Barents Sea
12
Norwegian international waters
13
Norwegian international waters
Archive data
d1
Irminger Sea deepa
d2
Irminger Sea deep
d3
Irminger Sea deep
d4
Irminger Sea deep
i1
Icelandic shelf
i2
Icelandic shelf
s1
Irminger Sea shallow
s2
Irminger Sea shallow
s3
Irminger Sea shallow
s4
Irminger Sea shallow
Month
Year
Gear
Trawl depth range
Latitude (88 N)
Longitude
n
7
10
4
4
7
8
7
1
1
1
1
1
1
1
1
10
1
1
9
10
10
11
11
9
9
9
9
10
11
1995
2006
2006
2006
2006
2006
2006
2007
2007
2007
2007
2007
2007
2007
2007
2006
2007
2007
2006
2006
2006
2006
2006
2006
2006
2006
2006
2006
2006
Pe
Dm
Pe
Pe
Pe
Pe
Pe
Dm
Dm
Dm
Dm
Dm
Dm
Dm
Dm
Dm
Dm
Dm
Pe
Pe
Pe
Pe
Pe
Pe
Pe
Pe
Pe
Pe
Pe
740
444 –518
695
786
284
284
494
589 –644
629 –606
689 –609
704 –658
388 –401
398 –361
366 –433
426 –438
430
400
620
440
320
360
300
340
350
375
415
445
385
300
58.57
65.17
61.45
61.44
55.38
55.39
56.10
60.59
60.45
61.11
60.45
61.59
62.09
62.21
62.27
66.14
71.56
70.21
72.06
72.19
72.19
72.19
72.01
71.38
72.05
72.01
72.21
72.27
72.19
236.47
227.26
229.12
229.11
243.00
243.00
245.09
29.32
29.57
29.27
29.57
24.19
23.57
24.13
24.33
6.20
16.32
17.15
5.07
9.46
9.39
9.39
9.30
6.18
5.16
5.17
5.15
9.13
9.39
94
188
93
93
76
18
84
9
15
19
14
40
24
86
48
91
26
44
18
19
19
14
20
16
18
20
20
7
7
7
10
4
6
3
3
7
10
10
6
1995
1996
2000
2001
2000
2001
1995
1996
2000
2001
Pe
Pe
Pe
Pe
Dm
Dm
Pe
Pe
Pe
Pe
740
540
695
780
375 –384
384 –457
230 –260
225 –420
449
200 –250
58.57
59.43
61.30
61.77
64.12
64.00
58.60
59.75
55.03
57.98
235.47
233.75
228.08
227.90
212.92
213.67
236.42
232.20
245.15
239.05
94
54
25
83
49
51
94
100
36
95
Gear, fishing gear used, either demersal (Dm) or pelagic (Pe) trawls.
a
Sample number 1 and d1 represent the archive individuals processed with current samples.
(named clusters D, S, and I) and temporal stability within each
cluster.
To examine how the two different spatial dimensions, depth
and geographical distance, could explain the genetic variation,
we used the partial Mantel test (e.g. Sokal and Rohlf, 1995)
carried out in the software IBD 1.2 (Bohonak, 2002). Matrices
of genetic distances [using FST/(12FST), e.g. Raymond and
Rousset, 1995], depth (logarithm of the differences in catch
depth, m), and distance (logarithm of the shortest geographical
distance, km) were constructed for pairwise comparison and
tests were based on 10 000 randomizations. To eliminate the
effect of geographical distance on the depth variable, we applied
distance-based redundancy analysis (dbRDA; Anderson, 2001)
with geographical distance as a covariate. The method performed
a multivariate regression on the distance response matrix (see
above) and p-values were obtained after 9999 permutations. For
all multiple tests, the Bonferroni corrections were applied to
p-values (Rice, 1989).
Results
Genetic variability of the current samples
Genetic variability and tests for HWE across loci for each sample
are presented in Table 2. Allelic richness (r) and observed heterozygosity (HO) were higher in samples collected at the Irminger Sea
deep zone (,500 m). Genotypic proportions were out of HWE in
only 8 of 156 exact tests. None of these values were attributable to
particular loci or populations, and all lost significance after the
Bonferroni correction (Table 2).
Genetic differentiation among the current samples
Bayesian-based cluster analysis revealed the presence of three distinct gene pools of S. mentella in the data (Table 3). Both archive
Depth and genetic structure of Sebastes mentella across the North Atlantic
683
Figure 1. Locations of 13 samples of S. mentella from habitats in the Irminger Sea deep zone (.500 m), archive (1), current (3, 4), Icelandic
shelf west (2); Irminger Sea shallow zone southwest (,500 m; 5, 6); Faroe Islands west (7) and east (8, 9); Norwegian shelf (10), Barents Sea
(11), and Norwegian international waters (12, 13); and locations of archived data from the deep zone (d1– d4). The bathymetry at 1000 m is
indicated as dotted lines.
and current samples from the Irminger Sea deep zone and the
Faroe Islands west grouped in one cluster (Cluster D). Samples
from the Icelandic shelf west grouped separately (Cluster I), and
the third cluster (S) consisted of samples from the Irminger Sea
shallow zone (.500 m), Faroe Islands east, Norwegian shelf,
Barents Sea, and Norwegian international waters.
The mean value of individual admixture proportions (q)
showed clear segregation among clusters (Table 4). The distribution of individual q-values from samples within clusters was
significantly different in all clusters (Cluster D, Kruskal –Wallis
H12, 1239 ¼ 714.2, p , 0.0001; Cluster I, H12, 1239 ¼ 461.6,
p , 0.0001; Cluster S, H12, 1239 ¼ 805.8, p , 0.0001).
Tests for homogeneity of allele frequencies showed highly significant (p , 0.00001) allele frequency variation (average FST ¼
0.0019, p , 0.00001) for all loci except Sal4 (p , 0.0005). In all,
44 out of 78 pairwise comparisons of FST values between
samples were significant (Table 5). Significant values (p ,
0.00001) were only obtained between samples that grouped in separate clusters (Tables 3 and 4).
MDS analysis based on pairwise FST showed a clear difference
between samples of the different clusters (Figure 2a). Both dimensions drawn showed a comparable level of variation. Variation
among samples from Clusters D and S was mostly explained
along dimension 1, whereas variation along dimension 2 explained
most of the variation among Clusters I and S. Variation among
Clusters D and I was demonstrated among both dimensions.
The AMOVA based on the detected clusters showed that a small
but highly significant portion (3%, p , 0.00001) of the variation
could be attributed to the among-cluster component (Table 6),
supporting the existence of three gene pools. A negative variance
component for among samples within clusters indicated the
absence of genetic structure within clusters. A graphic representation of the geographical locations of Clusters D, I, and S is
shown in Figure 3.
Tests for allelic richness (r) and observed heterozygosity (HO)
showed that both measures were significantly higher for Cluster
D (r ¼ 14.8, s.d. ¼ 7.44, p , 0.005; HO ¼ 0.78, s.d. ¼ 0.208, p ,
0.01) than for Clusters I (r ¼ 12.9, s.d. ¼ 6.80; HO ¼ 0.75, s.d. ¼
0.208) and S (r ¼ 12.3, s.d. ¼ 5.79; HO ¼ 0.76, s.d. ¼ 0.169).
Neither measure differed between Clusters I and S (r, p . 0.8;
HO, p . 0.5). The distribution of individual admixture proportions (Figure 4) revealed that fish assigned (50%) to
Cluster D were collected deeper (.600 m) than those assigned
to either Cluster I (,500 m) or S (,500 m).
Temporal stability of the pattern observed
Pairwise comparisons of FST among current and archive samples
for Clusters D, I, and S showed that none of the archive samples
were significantly different from current ones from the same
cluster (data not shown). The MDS clearly illustrated that
archive samples grouped with current ones from the same
cluster (Figure 2b). In addition, AMOVA showed that no
portion of the variation was attributable to the among-years variance component, indicating temporal stability within all clusters
(Table 6). In support of these findings, the archive sample from
the Irminger Sea deep zone that was processed with current
684
Table 2. Number of alleles (A), allelic richness (r)a, sample size (n), expected (HE) and observed (HO) heterozygosity, and tests for deviation from HWD (p-value) are listed.
Locus
No.
1
2
3
4
5
6
8
SEB9
14
13.2
88
0.830
0.852
0.775
11
8.4
188
0.686
0.638
0.059
14
11.4
93
0.828
0.785
0.144
13
10.7
93
0.811
0.688
0.002
12
10.305
93
0.753
0.742
0.427
13
10.7
84
0.663
0.690
0.824
12
10.7
57
0.775
0.754
0.395
12
10.9
SEB25
20
18.5
84
0.823
0.786
0.177
21
15.7
187
0.869
0.925
0.997
21
18.0
93
0.854
0.828
0.248
22
18.2
93
0.843
0.806
0.168
16
12.599
93
0.825
0.849
0.789
14
10.9
84
0.827
0.786
0.187
18
16.4
57
0.821
0.877
0.931
16
13.8
SEB31
17
16.4
86
0.894
0.872
0.301
21
16.7
188
0.793
0.798
0.619
18
16.2
93
0.905
0.860
0.096
17
15.6
93
0.898
0.849
0.079
17
13.995
94
0.816
0.809
0.461
15
12.6
84
0.758
0.738
0.344
14
13.4
57
0.864
0.912
0.920
15
13.1
SEB33
32
31.6
67
0.958
0.940
0.300
33
27.4
186
0.950
0.930
0.141
32
27.2
93
0.957
0.989
0.985
36
29.6
93
0.958
0.968
0.765
34
26.832
93
0.948
0.935
0.351
32
27.1
83
0.954
0.940
0.325
27
25.8
57
0.957
0.965
0.714
29
25.8
SEB45
21
21.0
63
0.929
0.952
0.842
28
22.3
187
0.900
0.941
0.987
21
17.6
93
0.911
0.892
0.313
24
19.3
93
0.927
0.925
0.524
24
19.483
93
0.886
0.925
0.929
24
20.1
83
0.893
0.855
0.164
23
20.7
56
0.893
0.804
0.023
21
19.3
Sal1
16
15.7
80
0.915
0.913
0.533
15
13.4
187
0.873
0.807
0.005
19
16.2
93
0.915
0.892
0.267
20
16.1
93
0.905
0.914
0.679
17
15.255
94
0.888
0.862
0.242
18
15.5
83
0.867
0.880
0.702
17
16.1
57
0.923
0.982
0.990
17
15.6
Sal3
6
5.9
90
0.513
0.500
0.401
6
5.6
188
0.689
0.665
0.249
7
6.2
91
0.607
0.604
0.520
9
7.4
93
0.575
0.570
0.491
7
6.377
93
0.717
0.634
0.032
7
6.0
83
0.665
0.687
0.741
8
7.4
57
0.567
0.526
0.242
6
5.9
Sal4
3
3.0
91
0.265
0.275
0.755
4
3.9
188
0.273
0.277
0.638
3
2.9
90
0.156
0.167
1.000
3
3.0
93
0.234
0.183
0.016
6
4.652
92
0.307
0.348
1.000
4
3.5
82
0.261
0.232
0.153
4
3.8
57
0.166
0.175
1.000
4
3.9
Smen5
12
11.8
69
0.857
0.855
0.538
14
13.1
177
0.874
0.898
0.877
13
11.4
91
0.853
0.802
0.104
12
11.0
92
0.864
0.804
0.063
13
11.541
91
0.814
0.791
0.312
12
10.8
81
0.839
0.864
0.791
11
10.7
50
0.852
0.800
0.191
11
10.9
Smen10
9
9.0
90
0.850
0.800
0.113
12
8.8
188
0.595
0.574
0.227
10
9.5
93
0.839
0.806
0.225
11
9.7
92
0.832
0.837
0.601
10
8.319
94
0.774
0.745
0.278
8
7.2
83
0.751
0.735
0.406
9
8.5
57
0.818
0.789
0.335
9
8.6
Spi4
14
12.9
86
0.838
0.849
0.664
12
9.9
185
0.789
0.789
0.534
16
13.0
93
0.852
0.849
0.522
13
11.1
93
0.819
0.860
0.911
10
9.966
93
0.878
0.860
0.343
13
11.7
83
0.881
0.855
0.271
11
10.5
57
0.837
0.842
0.613
12
11.0
Spi10
31
28.6
87
0.952
0.885
0.008
12
10.5
179
0.819
0.771
0.047
31
26.6
84
0.953
0.952
0.572
34
28.0
86
0.952
0.953
0.605
12
10.507
93
0.749
0.763
0.705
14
11.1
81
0.669
0.630
0.171
26
24.8
50
0.929
0.840
0.017
12
11.4
Overall
16.25
15.64
91
0.802
0.790
0.098
15.75
12.98
188
0.759
0.751
0.127
17.08
14.69
93
0.802
0.786
0.031
17.83
14.97
93
0.801
0.780
0.009
14.83
12.49
94
0.780
0.772
0.220
14.50
12.27
84
0.752
0.741
0.136
15.00
14.06
57
0.783
0.772
0.172
13.67
12.51
M. Ö. Stefánsson et al.
7
Variable
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
10
11
12
13
Total
64
0.734
0.734
0.567
14
10.6
134
0.730
0.709
0.283
15
12.5
91
0.729
0.736
0.641
13
11.4
69
0.772
0.768
0.521
15
12.2
90
0.699
0.700
0.575
12
10.2
88
0.714
0.659
0.086
23
22.4
1232
0.754
0.718
0.000
64
0.851
0.859
0.639
20
14.6
133
0.837
0.827
0.405
19
14.5
89
0.838
0.775
0.064
17
15.1
66
0.852
0.848
0.527
16
13.5
88
0.842
0.864
0.771
15
12.4
88
0.835
0.784
0.120
25
24.8
1219
0.851
0.838
0.088
64
0.755
0.703
0.160
19
14.0
134
0.779
0.746
0.160
12
11.5
91
0.797
0.846
0.943
14
12.5
67
0.792
0.761
0.279
18
14.9
90
0.823
0.844
0.788
14
11.9
88
0.747
0.750
0.582
27
26.9
1229
0.828
0.805
0.007
64
0.952
0.938
0.367
34
25.8
134
0.953
0.970
0.886
29
25.6
86
0.955
0.965
0.754
27
25.8
50
0.947
0.920
0.270
29
26.4
76
0.955
0.947
0.452
32
27.6
82
0.951
0.939
0.368
42
41.8
1164
0.956
0.950
0.192
63
0.891
0.921
0.847
24
18.5
134
0.889
0.918
0.900
26
20.5
88
0.879
0.864
0.377
21
18.2
60
0.864
0.850
0.426
21
17.4
86
0.880
0.919
0.914
24
19.3
85
0.873
0.906
0.879
37
36.6
1184
0.901
0.905
0.676
64
0.901
0.922
0.782
15
13.3
131
0.869
0.885
0.760
15
13.1
90
0.865
0.878
0.703
16
14.7
68
0.897
0.868
0.256
18
15.8
89
0.885
0.854
0.204
16
14.1
86
0.892
0.930
0.918
24
23.8
1215
0.896
0.882
0.056
64
0.640
0.594
0.208
6
5.6
134
0.657
0.657
0.535
6
5.9
89
0.694
0.719
0.768
6
5.9
69
0.706
0.754
0.877
7
6.5
85
0.607
0.600
0.475
6
5.9
85
0.739
0.718
0.355
10
9.9
1221
0.663
0.637
0.012
63
0.376
0.349
0.271
3.0
3
132
0.275
0.242
0.064
4
3.5
83
0.293
0.313
0.847
3
3.0
60
0.281
0.250
0.195
4
3.5
84
0.248
0.250
0.628
6
4.8
84
0.312
0.321
0.704
6
6.0
1199
0.267
0.261
0.169
63
0.870
0.825
0.184
12
11.3
130
0.845
0.792
0.054
12
11.5
76
0.854
0.789
0.068
11
11.0
43
0.804
0.814
0.645
12
11.0
62
0.826
0.806
0.384
12
11.2
69
0.817
0.826
0.638
15
15.0
1094
0.863
0.826
0.000
64
0.792
0.750
0.237
9
7.6
133
0.765
0.835
0.985
8
7.2
90
0.783
0.833
0.913
9
8.2
66
0.773
0.788
0.667
11
8.8
89
0.770
0.753
0.382
9
7.8
85
0.762
0.753
0.467
12
12.0
1224
0.796
0.756
0.000
64
0.845
0.875
0.821
13
11.8
134
0.852
0.858
0.630
13
11.3
89
0.861
0.809
0.095
11
10.6
64
0.838
0.813
0.332
13
11.8
83
0.872
0.880
0.641
13
11.8
82
0.866
0.841
0.294
18
17.9
1206
0.856
0.841
0.069
61
0.754
0.787
0.846
11
10.4
132
0.728
0.705
0.250
11
9.9
84
0.669
0.655
0.398
10
9.6
57
0.639
0.632
0.506
11
10.1
78
0.648
0.641
0.478
10
9.9
80
0.675
0.613
0.057
39
38.7
1152
0.821
0.757
0.000
64
0.780
0.771
0.237
15.00
12.21
134
0.765
0.762
0.357
14.17
12.25
91
0.768
0.765
0.389
13.17
12.16
69
0.764
0.755
0.243
14.58
12.66
90
0.755
0.755
0.505
14.08
12.23
88
0.765
0.753
0.125
23.17
22.97
1232
0.788
0.765
0.000
Depth and genetic structure of Sebastes mentella across the North Atlantic
9
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
A
r
n
HE
HO
HW
p-values in bold are significant at a ¼ 0.05, p , 0.00032.
a
Allelic richness per locus and population based on a minimum sample size of 43 diploid individuals.
685
686
M. Ö. Stefánsson et al.
Table 3. The partition of 13 S. mentella samples from the North
Atlantic using the program BAPS (Corander et al., 2003, 2004).
Cluster
D
I
S
Habitat
Irminger Sea deep-zone: archive; current;
Faroe Islands west
Icelandic shelf west
Irminger Sea shallow-zone; Faroe Islands
east; Norwegian shelf; Barents Sea;
Norwegian international waters
Sample
number
1a; 3; 4; 7
2
5; 6; 8; 9; 10;
11; 12; 13
Discussion
Maximum posterior probability [p (Sjdata) ¼ 1.000].
a
Sample number 1 and d1 represent the archive individuals processed with
current samples.
Table 4. Admixture analysis for three clusters carried out in the
program BAPS.
No. Habitat
1
Irminger Sea deep
archivea
2
Icelandic shelf west
3
Irminger Sea deep
4
northeast
5
Irminger Sea
6
shallow
southwest
7
Faroe Islands west
8
Faroe Islands east
9
10 Norwegian shelf
11 Barents Sea
12 Norwegian
13
international
waters
Cluster D
Cluster I
0.840 (0.227) 0.052 (0.086)
Cluster S
0.108 (0.227)
0.049 (0.107)
0.857 (0.193)
0.830 (0.263)
0.067 (0.130)
0.041 (0.087)
0.839 (0.225)
0.063 (0.117)
0.078 (0.142)
0.130 (0.189)
0.121 (0.180)
0.112 (0.200)
0.080 (0.156)
0.092 (0.210)
0.803 (0.248)
0.838 (0.195)
0.708 (0.305)
0.101 (0.183)
0.048 (0.103)
0.045 (0.091)
0.054 (0.101)
0.057 (0.093)
0.045 (0.101)
0.162
0.154
0.107
0.098
0.122
0.104
0.102
0.130 (0.209)
0.745 (0.291)
0.845 (0.197)
0.857 (0.184)
0.824 (0.226)
0.839 (0.213)
0.853 (0.196)
(0.263)
(0.257)
(0.163)
(0.150)
(0.197)
(0.179)
(0.174)
samples grouped in Cluster D in all analyses. The Mantel tests
indicated significant correlations between genetic distance
measures and depth (Z ¼ 2.4965, r 2 ¼ 0.3394, p 0.0023), but
no such relationship was found among genetic distance and geographic distance (Z ¼ 3.3455, r 2 ¼ 0.0298, p 0.0855). dbRDA
confirmed these findings, and the results show that depth differences between redfish populations were not influenced by geographic distance (Table 7).
Mean admixture proportions (q) and standard deviation (s.d.) are given for
13 samples of S. mentella from the Irminger Sea. The program BAPS
(Corander and Marttinen, 2006) was used to estimate individual values of q,
i.e. the estimated proportion of an individual’s genotype originating from
one of three populations of origin. Values in bold represent the highest q
values (s.d.) for each sample.
a
Sample number 1 and d1 represent the archive individuals processed with
current samples.
The main objective of our study was to use genetic markers to
assess whether commercial trawling for S. mentella is directed at
one or more stocks inhabiting the North Atlantic.
Contemporary samples were used to assess the spatial structuring,
and comparison with archive samples was used to assess the temporal stability of the genetic pattern detected. All analyses revealed
that S. mentella is genetically structured within the North Atlantic
and that the genetic pattern was stable over a period of 10 years.
Three main groups were distinguishable: Cluster D (Irminger Sea
deep-sea plus Faroe Islands west), Cluster S (Irminger Sea shallow,
Faroe Islands east, Norwegian shelf, Barents Sea, and Norwegian
international waters), and Cluster I (Icelandic shelf west). A
detailed discussion of the genetic variability between these clusters,
as well as the level of genetic differentiation observed and the
potential consequences for fisheries management, is provided
below.
Genetic variability of the current samples
The genetic variability of S. mentella across the study area is similar
to that observed for other commercial marine species (Ruzzante
et al., 1997; Pampoulie et al., 2006; Hyde et al., 2008; Was et al.,
2008), and is comparable with findings reported previously for
the species (Roques et al., 2002; Stefánsson et al., 2004;
Pampoulie and Danı́elsdóttir, 2008). All samples were in HWE
across loci within populations, and a deficit of heterozygotes was
detected across samples, as would be expected under the
Wahlund effect, indicating phenotypic mixing. Variation measures
(A, HO, HE, and r) were higher for samples assigned to Cluster D,
but more similar for samples in Clusters I and S, indicating greater
genetic variability in deep-sea S. mentella.
Table 5. Estimates of pairwise genetic differentiation (FST) among 13 redfish samples (below diagonal) and their p-values (above diagonal).
Irminger
Faroe Islands
Norwegian shelf
Irminger Sea
Sea
Icelandic Irminger Sea deep
shallow
archived
shelf
West
East
East
Mid
North
No.
2
3
4
5
6
7
8
9
10
11
1a
1
0.00001
0.84874 0.41072 0.00001
0.00001 0.10779
0.00001 0.00001
0.00001 0.00001
2
0.03980
0.00001 0.00001 0.00001
0.00001 0.00001
0.00001 0.00001
0.00001 0.00001
3
20.00050 0.03570
0.84337 0.00001
0.00001 0.20659
0.00001 0.00001
0.00001 0.00001
4
20.00080 0.03800 20.00100
0.00001
0.00001 0.04944
0.00001 0.00001
0.00001 0.00001
5
0.03240 0.02260
0.03150 0.03230
0.91258 0.00001
0.03168 0.00405
0.06462 0.30660
6
0.03760 0.02390
0.03770 0.03810 0.00040
0.00001
0.07699 0.11924
0.27532 0.52373
7
0.00290 0.02760
0.00110 0.00240 0.02530
0.02960
0.00001 0.00001
0.00001 0.00001
8
0.02420 0.02110
0.02620 0.02510 0.00180
0.00130 0.01940
0.66460
0.12737 0.09068
9
0.03040 0.02560
0.03210 0.03120 0.00200
0.00070 0.02610 20.00120
0.09305 0.26228
10
0.03390 0.02390
0.03380 0.03440 0.00020 20.00140 0.02750 20.00090 0.00020
0.04304
11
0.03490 0.02510
0.03410 0.03550 0.00180
0.00230 0.02940
0.00210 0.00080
0.00050
12
0.03030 0.02820
0.03080 0.03150 0.00300
0.00110 0.02330
0.00120 0.00130
0.00080 0.00010
13
0.03810 0.02200
0.03700 0.03800 0.00160
0.00040 0.02990
0.00100 0.00190 20.00140 0.00000
International
waters
12
0.00001
0.00001
0.00001
0.00001
0.13121
0.34292
0.00001
0.34027
0.18203
0.08164
0.71560
13
0.00001
0.00001
0.00001
0.00001
0.13428
0.49849
0.00001
0.58447
0.13410
0.82637
0.21921
0.11519
0.00200
See Table 1 for sample numbers. p-values in bold are significant at: a ¼ 0.05, p , 6.6 1024; a ¼ 0.01, p , 1.3 1024; a ¼ 0.001, p , 1.3 1025.
a
Sample number 1 and d1 represent the archive individuals processed with current samples.
687
Depth and genetic structure of Sebastes mentella across the North Atlantic
Table 6. A hierarchical AMOVA.
Source of variation
Spatial dimension
Among clusters
Among samples
within clusters
Within samples
Temporal dimension
Cluster D
Among years
Among samples
within years
Within samples
Cluster I
Among years
Among samples
within years
Within samples
Cluster S
Among years
Among samples
within years
Within samples
Percentage of
variation
F-statistics
p-valuea
3.12
20.40
0.03117
0.00415
,0.00001
,0.00001
97.28
0.02715
,0.00001
20.43
0.02
20.00434
0.00016
.0.8 n.s.
.0.5 n.s.
100.42
20.00471
.0.9 n.s.
20.13
20.01
20.00128
20.00011
.0.8 n.s.
.0.5 n.s.
100.14
20.00140
.0.8 n.s.
0.05
20.32
0.00047
20.00324
.0.4 n.s.
.0.9 n.s.
100.28
20.00277
.0.9 n.s.
Spatial dimension: comparison among Clusters D, I, and S from
Bayesian-based cluster analysis in Table 4 for 13 samples of S. mentella from
the North Atlantic. Temporal dimension: comparison among contemporary
and archive samples from the same habitat.
a
Probability of having more extreme variance component and F-statistic
than the observed values by chance alone.
n.s., not significant.
Figure 2. MDS graphs based on pairwise FST values. (a) Analysis
based on 13 samples within the North Atlantic and 12 microsatellite
loci. (b) An additional ten temporal samples at nine loci included
(see Table 1 for sample codes). Letters D, I, and S denote clusters
from Bayesian-based cluster analysis on samples from (a).
Temporal stability of the genetic structure
All genetic analyses revealed the significant genetic structure of
S. mentella, supporting earlier findings (Johansen et al., 2000;
Roques et al., 2002; Joensen and Grahl-Nielsen, 2004; Stefánsson
et al., 2004, 2006, in press; Stefánsson and Pampoulie, 2006;
Danı́elsdóttir et al., 2008). The genetic differentiation observed
was low (mean FST ¼ 0.019, p , 0.00001, range 0.000 –0.0398),
but similar to that observed in other exploited fish species in the
North Atlantic (O’Leary et al., 2007; Was et al., 2008), and to
the earlier study of Roques et al. (2002). Despite the low level of
differentiation, more than half of the pairwise FST comparisons
were significant and revealed the presence of different geographic
groups, also reflected by the MDS analysis. Cluster S grouped
samples from Norwegian international waters with samples from
the Irminger Sea shallow southwest area, Faroe Islands east,
Norwegian shelf, and Barents Sea. Cluster I consisted of samples
only from the Icelandic shelf, whereas Cluster D consisted of
samples from the Irminger Sea deep northeast area and Faroe
Islands west. The clustering method (BAPS) clearly confirmed
the presence of these three clusters (q . 0.708).
In our study, we also investigated temporal stability using two
different but non-exclusive methods. First, an archive sample collected at 740 m in the Irminger Sea in 1995 was used in all analyses
performed on the contemporary samples. Each of these analyses
indicated that this sample was not genetically different from
samples in Cluster D (,500 m). Second, we compared the contemporary and archive data using a MDS and AMOVA, to test
temporal stability within the clusters we detected. The MDS analysis clustered the contemporary and archive samples in three distinct genetic groups: Irminger deep-sea (Cluster D, samples 3, 4,
7, d1, d2, d3, and d4), Irminger shallow (Cluster S, samples 5, 6,
8, 9, 10, 11, s1, s2, s3, and s4), and Icelandic shelf west (Cluster
I, samples 2, i1, and i2). In addition, the AMOVA showed that
no portion of the variation was attributable to the among-years
variance component, indicating temporal stability within all clusters. Although the period investigated may not be sufficient to
detect any substantial genetic changes (S. mentella is long-living),
both approaches suggest that the structure was persistent over an
extensive period, so supporting inferences that the population
structure described here is real (Waples, 1998).
Depth: the third dimension to fisheries management
Although we detected clear genetic variation, the geographical distribution of clusters appeared to overlap, indicating the sympatric
existence of three S. mentella gene pools. However, when individual admixture proportions (q) were plotted against depth, samples
from Clusters D and S showed clear depth segregation. This trend
688
M. Ö. Stefánsson et al.
Figure 3. Locations of clusters D, I, and S as interpreted from Bayesian-based cluster analysis (Table 4). The bathymetry at 1000 m is indicated
as dotted lines.
Figure 4. Contour plots representing the distribution of admixture proportion (q) in contemporary samples with depth (m) in the North
Atlantic. Only values .0.5 were plotted for each cluster. All graphs represent two-dimensional space, where the position of the fish is
determined by q on the x-axis and ocean depth on the y-axis. Contours were drawn according to the number of fish behind them in this
two-dimensional space. The scale bar shows the number of fish.
Table 7. Effect of depth and geographic distance on genetic
differentiation (using FST/(12FST) of S. mentella using dbRDA) with
depth and geographic distance tested separately, and with the
effect of depth tested with geographic distance as a covariate.
Predictors
Pseudo-F
Permutation p
Individual sets of predictor variables
Depth
19.4
0.0001
Distance
1.6
0.2085
Geographical distance as a covariate
Depth
18.3
0.0027
% variance
23.3
2.5
22.0
Pseudo-F statistics and permutation p-values are shown. % variance is the
percentage of the variation explained by each predictor variable.
was noticeable particularly within the Irminger Sea, as previously
described in analyses of archived data (Stefánsson et al., 2004, in
press; Stefánsson and Pampoulie, 2006). The current results corroborate previous conclusions that geographical emphasis in fisheries management cannot fully account for the distribution of
pelagic fish, and that depth could also be a useful measure for
management. Indeed, the emphasis put on the depth dimension
in the current study may have facilitated the capture of genetic
variation across the panoceanic zone described by Roques et al.
(2002). As sampling was restricted to the upper 500-m layer in
the latter study, this could offer an explanation for the discrepancy
between our findings and those of Roques et al. (2002), i.e. any
Depth and genetic structure of Sebastes mentella across the North Atlantic
variation associated with depths .500 m remained beyond the
scope of Roques et al. (2002). Further underscoring the importance of the depth dimension in marine fisheries management,
our results showed that this trend persisted over time, with the
archive deep-sea and shallow samples grouping together with
Clusters D and S, respectively.
Consequences for fisheries management
To conclude, we found significant population structure in S. mentella across its distribution, despite the high potential for gene
flow, its high fecundity, long life expectancy, presumably large
effective population sizes, and extensive migration patterns. The
substructure we saw was temporally stable and did not appear to
be linked to isolation by distance; instead, depth seemed to be
the driving force. Our findings, together with those of previous
studies (Danı́elsdóttir et al., 2008; Pampoulie and Danı́elsdóttir,
2008; Stefánsson et al., in press), suggest that the assumption of
panmixia for this species could have detrimental effects on stock
structure and population persistence.
Acknowledgements
We thank Valérie Chosson, Sif Guðmundsdóttir, Torild Johansen,
Svend Lemvig, and Janicke Skadal for their skilled technical assistance, and two anonymous reviewers for valued input to the manuscript submitted.
References
Anderson, M. J. 2001. A new method for non-parametric multivariate
analysis of variance. Austral Ecology, 26: 32 – 46.
Begg, G. A., Friedland, K. D., and Pearce, J. B. 1999. Stock identification and its role in stock assessment and fisheries management:
an overview. Fisheries Research, 43: 1 – 8.
Bohonak, A. J. 2002. IBD (Isolation by Distance): a program for analyses of isolation by distance. Journal of Heredity, 93: 153 – 154.
Christensen, V., Guénette, S., Heymans, J. J., Walters, C. J., Watson, R.,
Zeller, D., and Pauly, D. 2003. Hundred-year decline of North
Atlantic predatory fishes. Fish and Fisheries, 4: 1 – 24.
Corander, J., and Marttinen, P. 2006. Bayesian identification of admixture events using multilocus molecular markers. Molecular
Ecology, 15: 2833 – 2843.
Corander, J., Waldmann, P., Marttinen, P., and Sillanpaa, M. J. 2004.
BAPS 2: enhanced possibilities for the analysis of genetic population structure. Bioinformatics, 20: 2363 –2369.
Corander, J., Waldmann, P., and Sillanpaa, M. J. 2003. Bayesian analysis of genetic differentiation between populations. Genetics, 163:
367– 374.
Danı́elsdóttir, A. K., Gı́slason, D., Kristinsson, K., Stefánsson, M. Ö.,
Johansen, T., and Pampoulie, C. 2008. Population structure of
deep-sea and oceanic Sebastes mentella in the Irminger Sea and
the continental slope in Icelandic waters based on allozyme
markers: are they cryptic species? Transactions of the American
Fisheries Society, 137: 1723– 1740.
Excoffier, L., Laval, G., and Schneider, S. 2005. Arlequin ver. 3.0: an
integrated software package for population genetics data analysis.
Evolutionary Bioinformatics Online, 1: 47 – 50.
FAO. 2007. The State of the World Fisheries and Aquaculture 2006.
FAO, Rome.
Gharrett, A. J., Matala, A. P., Peterson, E. L., Gray, A. K., and Li, Z.
2007. Distribution and population genetic structure of sibling
rougheye rockfish species. In Biology, Assessment, and
Management of North Pacific Rockfishes, pp. 121 – 183. Ed. by J.
Heifetz, J. Dicosmo, A. J. Gharrett, M. S. Love, V. M. O’Connell,
and R. D. Stanley. Alaska Sea Grant College-AK-SG-07-01.
689
Gomez-Uchida, D., Hoffman, E. A., Ardren, W. R., and Banks, M. A.
2003. Microsatellite markers for the heavily exploited canary
(Sebastes pinniger) and other rockfish species. Molecular Ecology
Notes, 3: 387 – 389.
Goudet, J. 1995. FSTAT (Version 1.2): a computer program to calculate F-statistics. Journal of Heredity, 86: 485– 486.
Guo, S. W., and Thompson, E. A. 1992. Performing the exact test of
Hardy– Weinberg proportion for multiple alleles. Biometrics, 48:
361– 372.
Hilborn, R., Quinn, T. P., Schindler, D. E., and Rogers, D. E. 2003.
Biocomplexity and fisheries sustainability. Proceedings of the
National Academy of Sciences of the United States of America,
27: 6564– 6568.
Hyde, J. R., Kimbrell, C. A., Budrick, J. E., Lynn, E. A., and Vetter,
R. D. 2008. Cryptic speciation in the vermilion rockfish (Sebastes
miniatus) and the role of bathymetry in the speciation process.
Molecular Ecology, 17: 1122 – 1136.
ICES. 2001. Planning Group on Redfish Stocks. ICES Document CM
2001/D: 04 Ref. ACFM.
ICES. 2007. Report of the Arctic Fisheries Working Group (AFWG),
18 – 27 April 2007, Vigo, Spain. ICES Document CM 2007/
ACFM: 16. 641 pp.
ICES. 2008. Report of the Arctic Fisheries Working Group (AFWG),
21 – 29 April 2008, ICES Headquarters, Copenhagen. ICES
Document CM 2008/ACOM: 01. 531 pp.
Joensen, H., and Grahl-Nielsen, O. 2004. Stock structure of Sebastes
mentella in the North Atlantic revealed by chemometry of the
fatty acid profile in heart tissue. ICES Journal of Marine Science,
61: 113– 126.
Johansen, T., Danielsdottir, A. K., Meland, K., and Naevdal, G. 2000.
Studies of the genetic relationship between deep-sea and oceanic
Sebastes mentella in the Irminger Sea. Fisheries Research, 49:
179– 192.
Lage, C., Kuhn, K., and Kornfield, I. 2004. Genetic differentiation
among Atlantic cod (Gadus morhua) from Browns Bank,
Georges Bank, and Nantucket Shoals. Fishery Bulletin US, 102:
289– 297.
Magnússon, J. 1972. Tilraunir til karfaveida med midsjávarvörpu ı́
úthafinu. Aegir, 65: 430 – 432.
Magnússon, J., and Magnússon, J. V. 1995. Oceanic redfish (Sebastes
mentella) in the Irminger Sea and adjacent waters. Scientia
Marina, 59: 241 – 254.
Miller, K. M., Schulze, A. D., and Withler, R. E. 2000. Characterization
of microsatellite loci in Sebastes alutus and their conservation in
congeneric rockfish species. Molecular Ecology, 9: 240– 242.
Morato, T., Watson, R., Pitcher, T. J., and Pauly, D. 2006. Fishing down
the deep. Fish and Fisheries, 7: 24 – 34.
O’Leary, D. B., Coughlan, J., Dillane, E., McCarthy, T. V., and Cross,
T. F. 2007. Microsatellite variation in cod Gadus morhua throughout its geographic range. Journal of Fish Biology, 70: 310– 335.
Pampoulie, C., and Danı́elsdóttir, A. K. 2008. Resolving species identification problems in the genus Sebastes using nuclear genetic
markers. Fisheries Research, 93: 54 – 63.
Pampoulie, C., Ruzzante, D. E., Chosson, V., Jörundsdóttir, T. D.,
Taylor, L., Thorsteinsson, V., Danı́elsdóttir, A. K., et al. 2006.
The genetic structure of Atlantic cod (Gadus morhua) around
Iceland: insight from microsatellites, the Pan I locus, and tagging
experiments. Canadian Journal of Fisheries and Aquatic Sciences,
63: 2660– 2674.
Pampoulie, C., Steingrund, P., Stefánsson, M. Ö., and Danı́elsdóttir,
A. K. 2008. Genetic divergence among East Icelandic and Faroese
populations of Atlantic cod provides evidence for historical
imprints at neutral and non-neutral markers. ICES Journal of
Marine Science, 65: 65 – 71.
Raymond, M., and Rousset, F. 1995. An exact test for population
differentiation. Evolution, 49: 1280– 1283.
690
Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution, 43:
223– 225.
Rico, C., Ibrahim, K. M., Rico, I., and Hewitt, G. M. 1997. Stock composition in North Atlantic populations of whiting using microsatellites markers. Journal of Fish Biology, 51: 462 – 475.
Roques, S., Pallotta, D., Sevigny, J. M., and Bernatchez, L. 1999.
Isolation and characterization of polymorphic microsatellite
markers in the North Atlantic redfish (Teleostei: Scorpaenidae,
genus Sebastes). Molecular Ecology, 8: 685 – 687.
Roques, S., Sevigny, J. M., and Bernatchez, L. 2002. Genetic structure
of deep-water redfish, Sebastes mentella, populations across the
North Atlantic. Marine Biology, 140: 297 – 307.
Ross, I., and Gentleman, R. 1996. R: a language for data analysis and
graphics. Journal of Computational and Graphical Statistics, 5:
299– 314.
Ruzzante, D. E., Taggart, C. T., Cook, D., and Goddard, S. V. 1997.
Genetic differentiation between inshore and offshore Atlantic cod
(Gadus morhua) off Newfoundland: a test and evidence of temporal stability. Canadian Journal of Fisheries and Aquatic
Sciences, 54: 2700– 2708.
Sambrook, J., and Russell, D. W. 2001. Molecular Cloning: a
Laboratory Manual. Cold Spring Harbour Laboratory Press,
New York.
Sigurðsson, Th., Kristinsson, K., Rätz, H-J., Nedreaas, K. H., Melnikov,
S. P., and Reinert, J. 2006. The fishery for pelagic redfish (Sebastes
mentella) in the Irminger Sea and adjacent waters. ICES Journal of
Marine Science, 63: 725– 736.
Sokal, R. R., and Rohlf, F. J. 1995. Biometry, 3rd edn. W. H. Freeman
and Co., New York.
Stefánsson, M. Ö., Gislason, D., Thorgilsson, B., Ragnarsdóttir, A.,
Danı́elsdóttir, A. K., and Sigurðsson, Th. 2004. Population structure of S. mentella within and around the Irminger Sea. Working
Document to the Study Group on Stock Identity and
Management Units of Redfishes. ICES Document CM 2003/
ACFM WD: 07.
Stefánsson, M. Ö., and Pampoulie, C. 2006. The influence of depth in
fishery management: evidence from redfish (Sebastes mentella) and
cod (Gadus morhua). Journal of Fish Biology, 69(Suppl. C): 256.
M. Ö. Stefánsson et al.
Stefánsson, M. Ö., Pampoulie, C., Thorgilsson, B., Ragnarsdóttir, A.,
Gı́slason, D., Danı́elsdóttir, A. K., and Sigurðsson, Þ. 2006.
Natural selection and adaptation of the deep-sea redfish Sebastes
mentella. Journal of Fish Biology, 69(Suppl. C): 235.
Stefánsson, M. Ö., Sigurdsson, Th., Pampoulie, C., Danı́elsdóttir,
A. K., Thorgilsson, B., Ragnarsdóttir, A., Gı́slason, D., et al.
Pleistocene genetic legacy suggests incipient species of Sebastes
mentella in the Irminger Sea. Heredity, in press.
Stephenson, R., and Kenchington, E. 2000. Conserving fish stock
structure is a critical aspect of preserving biodiversity. ICES
Document CM 2000/Mini: 07. 7 pp.
Tåning, A. V. 1949. On the breeding places and abundance of the
redfish (Sebastes) in the North Atlantic. Journal du Conseil
Permanent International pour l’Exploration de la Mer, 16: 85– 95.
Vinnichenko, V. I. 2007. Russian investigations and fishery of the
pelagic redfish in the Norwegian Sea. Working Document
9. ICES Arctic Fisheries Working Group, Vigo, 18 – 27 April
2007. 23 pp.
Was, A., Gosling, E., McCrann, K., and Mork, J. 2008. Evidence for
population structure of blue whiting (Micromesistius poutassou)
in the Northeast Atlantic. ICES Journal of Marine Science, 65:
216– 225.
Walsh, P. S., Metzger, D. A., and Higuchi, R. 1991. Chelex-100 as a
medium for simple extraction of DNA for PCR-based typing
from forensic material. Biotechniques, 10: 506 – 513.
Waples, R. S. 1998. Separating the wheat from the chaff: patterns of
genetic differentiation in high gene flow species. Journal of
Heredity, 89: 435 – 450.
Ward, R. D. 2000. Genetic in fisheries management. Hydrobiologia,
420: 191 – 201.
Weir, B. S., and Cockerham, C. C. 1984. Estimating F-statistics for the
analysis of population-structure. Evolution, 38: 1358 – 1370.
Westgaard, J-I., and Fevolden, E. 2007. Atlantic cod (Gadus morhua L.)
in inner and outer coastal zones of northern Norway display divergent genetic signature at non-neutral loci. Fisheries Research, 85:
320– 329.
doi:10.1093/icesjms/fsp059