© 2005 International Council for the Exploration of the Sea
Spring-spawning herring (Clupea harengus L.) in the southwestern Baltic Sea: do they form genetically distinct spawning waves?
a Danish Institute for Fisheries Research, Department of Inland Fisheries Vejlsøvej 39, DK-8600 Silkeborg, Denmark
b Department of Genetics and Ecology, Biological Institute, University of Aarhus Building 540, Ny Munkegade, DK-8000 Aarhus C, Denmark
*Correspondence to H. B. H. Jørgensen: tel: +45 89 21 31 00; fax: +45 89 21 31. e-mail: hannebhj{at}yahoo.com.
Temporal sampling within the spring-spawning season has revealed differentiation in length-at-age in herring at Rügen and differentiation in, e.g., Anisakis infestation rate, otolith microstructure, and gillraker counts in Gda
sk Bay, leading to the expectation that spawning waves consist of distinct herring populations. We tested this expectation by analysing genetic variation at nine microsatellite loci in samples collected at different times during the March to May spawning season in 2 consecutive years, 2002 and 2003. Length-at-age, mean length, and age distributions were compared among samples within locations but did not show consistent temporal patterns. Pairwise genetic differentiation among temporal samples within season was low and non-significant in the Gdask Bay (0 < FST < 0.0025) but higher among Rügen samples (0.0008 < FST < 0.0113). Samples from Rügen collected in 2002 differed significantly from each other, and individual assignment tests showed increased divergence with time. Differentiation was not confounded by effects of age class or sex. We conclude that spawning waves are not genetically differentiated among Gdask samples based on factors analysed in this study, whereas genetically distinct but sympatric spawning populations may be found at Rügen.
Keywords: Clupea harengus L., growth rate, individual assignment, microsatellite DNA, spatio-temporal variation, spawning waves
Received 4 August 2004; accepted 21 April 2005.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Freshwater and anadromous fish often exhibit considerable morphological and ecological variability, even at very local scales. Studies applying molecular markers have revealed that different morphs may reflect distinct sympatric populations or even species (e.g. Gíslason et al., 1999; Taylor, 1999; Peichel et al., 2001). In contrast, gene flow among populations of marine fish is generally thought to be high and effective population sizes are assumed to be large, resulting in limited genetic drift and thereby low levels of genetic differentiation between spatially separated populations in contrast to freshwater fish (Ward et al., 1994; DeWoody and Avise, 2000). Typical estimates of FST reported among populations of marine fish at a regional scale are in the range 0.002–0.005 (e.g. Ruzzante et al., 2001; Knutsen et al., 2003; McPherson et al., 2004), and nearly all higher estimates of genetic differentiation involve either populations separated by thousands of kilometres or populations separated by steep environmental gradients (Mork et al., 1985; Nielsen et al., 2003). The expected low levels of differentiation make it particularly important to separate noise due to sampling error from a weak but "biologically meaningful" genetic signal when studying population structure in marine species (Waples, 1998; Hedrick, 1999). To enable meaningful conclusions to be drawn from studies on genetic population structure, one must therefore employ a sampling design that takes into consideration the overall biology of the species of interest, making sure to include only one suspected population in each sample. This is not only important from a strictly scientific perspective; understanding the distribution of intraspecific biodiversity (i.e. the genetic structuring below the species level) is essential for biologically meaningful conservation and fisheries management (Carvalho and Hauser, 1994; Ryman et al., 1995). Thus, in exploited fish species it is important to know if individuals harvested at any one location actually represent one or more populations, some of which could be overexploited.
The Atlantic herring exhibits a complex population structure with more or less divergent populations referred to as subspecies, races, tribes, stocks, groups etc. (Ryman et al., 1984; McQuinn, 1997). The identification and designation of populations, however, remain controversial. In the Baltic Sea, analyses of morphometric and meristic traits (Nævdal, 1972; Aro, 1989; Rechlin, 2000), spawning time and location (Ojaveer, 1990; Rechlin, 1991; Parmanne et al., 1994), growth rate, and otolith microstructure (Parmanne et al., 1994) have shown clear differences between local spawning groups of herring. Conversely, allozyme studies have shown very low genetic differentiation among spawning groups of herring in Scandinavian waters (e.g. Andersson et al., 1981; Ryman et al., 1984), and Ryman et al. (1984) found no indication of a correlation between genetic and morphological variation among putative spawning groups.
The situation is further complicated by the fact that differentiation may occur on both spatial and temporal scales. Atlantic herring are known to consist of both spring and autumn spawning populations, though currently the latter constitute only a minute proportion of the overall herring spawning biomass in the Baltic Sea (Aro, 1989). Even below that level, i.e. within the same spawning season, more than one spawning group of herring may use the same spawning ground, arriving at the locations in so-called "spawning waves". This has been observed in both Pacific herring (Clupea harengus pallasi) and Atlantic herring, spawning waves, arriving at the spawning locations separated by anything from a few days to more than a month (Hay, 1985; Lambert, 1987). Lambert (1987) suggested that the spawning waves are produced by the reproductive contribution of dominant year classes, and that age structure thereby determines the number of spawning waves per season in a population. Larger, older herring (repeat spawners) arrive earlier in the season than smaller, younger herring (recruit spawners). This is supported by a study by Rajasilta et al. (1993) in the Archipelago Sea of the Baltic.
Not much is known about the genetic differentiation between spawning waves (Hay, 1985), but McPherson et al. (2003) tested for genetically distinct spawning waves at four locations in the Scotian-Fundy region of the Northwest Atlantic. Two samples were collected at each location at 6–15-day intervals, and analysed using microsatellite markers. Temporal genetic differentiation within spawning season was very low at three of the locations but statistically significant at the fourth location, even if the spawning groups were spawning only 6 days apart.
In the Baltic Sea, spring-spawning herring often spawn in archipelagos and bays where the water is slightly warmer and food more abundant than offshore, and when offshore larvae are able to disperse actively they tend to shift inshore (Urho, 1999). Many spawning groups undertake yearly spawning and feeding migrations (Aro, 1989). In the present study we analysed genetic differentiation among temporally separated samples of spring-spawning herring from two locations in the southwestern Baltic Sea. One location is the Greifswalder Bodden, a semi-enclosed area south of the German island of Rügen. This is the spawning ground of the Rügen herring which undertake feeding migrations into Danish waters, with a small proportion of the fish migrating to feeding areas around Bornholm and Hanö Bay (Biester, 1979). The second location is off the coast in Gdask Bay. This is believed to be the spawning ground of three putative populations of herring: (i) Anisakis infected coastal herring undertaking feeding migrations into Danish waters, (ii) local coastal herring not undertaking long feeding migrations, and (iii) slower growing, offshore herring undertaking feeding migrations into the central Baltic Sea (M. Wyszynski, Sea Fisheries Institute, Gdynia, Poland, pers. comm.).
At these two locations, previous studies on variation in several characters have suggested the presence of divergent groups of herring spawning within the same location but at different times during the spring-spawning season (e.g. Kompowski, 1971; Biester and Hering, 1977; Biester et al., 1979; Rechlin, 2000; Podolska and Horbowy, 2002). From each location we analysed samples of herring collected at different times during the spring-spawning season. Using a combination of analyses of length measurements, age, and microsatellite DNA markers, we tested three hypotheses: (i) there is no temporal differentiation in length and age composition and length-at-age in samples taken over the spawning season, (ii) there is no temporal genetic differentiation among samples taken over the spawning season, and (iii) there is no agreement in the temporal patterns of length and age variation on the one hand and genetic variation on the other hand. In addition, we tested for interannual differentiation in the temporal samples and defined new subsamples within the samples based on age class and sex to test for confounding genetic differentiation between these subsamples that might affect the overall pattern of genetic differentiation.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Samples
Samples were taken at the Greifswalder Bodden, near Rügen in the western Baltic Sea and at Gda
sk Bay in the southern Baltic Sea (Figure 1). The samples from Rügen were taken to represent both faster growing, early-spawning herring and slower growing, late-spawning herring (Rechlin, 2000). The samples from Gdask Bay were taken to represent the three putative populations described in the Introduction. During the spring-spawning seasons of 2002 and 2003, gillnet samples were taken at each location in early, mid, and late season (at Rügen, the spring-spawning season lasts from March to May while it lasts from April to May, in Gdask Bay). The use of gillnets may cause some size selectivity in the samples. This was, however, not considered a problem in this study since the sampling procedure was identical for all samples, and the relative length-at-age, rather than the absolute length-at-age, was of interest here. The use of samples from consecutive spawning seasons increases the likelihood that results showing low genetic differentiation reflect a genetic signal rather than noise (Waples, 1998).
|
All specimens were in spawning condition, i.e. in maturity stages V and VI (Anon, 1962), to ensure that samples included only specimens originating from the spawning groups of interest. Specimens caught that could not be sexed were discarded (they made up only ca. 5% of the total number of specimens handled). Table 1 describes sampling times and locations. For all 100 specimens in each sample, sex, age (based on otolith readings), and length were recorded, and gill tissue was stored in 96% ethanol for the extraction of DNA. When age could not be determined, the specimen was excluded from the length-at-age analyses (Table 2). Age readings based on otoliths were performed by experienced age-readers at the Danish Institute for Fisheries Research (Rügen herring) and at the Polish Sea Fisheries Institute (herring from Gda
sk Bay).
|
|
Statistical analyses of morphological data
Growth rates are typically represented as regressions of length on age. To test for the influence of sampling time on length-at-age we set up an analysis of covariance, ANCOVA (Sokal and Rohlf, 1995), for each sampling location, with length as the dependent variable and age as the covariate. By means of the ANCOVAs we tested for differences in regression slopes and mean length-at-age among sampling times.
To test if larger, older herring arrive early at the spawning grounds while smaller, younger herring arrive later in the season (e.g. Lambert, 1987), rather than separate populations with different length-at-age, we compared length and age distributions of the samples. Length data were normally distributed and were compared by one-way ANOVAs (Sokal and Rohlf, 1995) for each sampling location and year. Because data on age were not normally distributed, age distributions were compared by the non-parametric Kruskal–Wallis Rank Sum Test (Zar, 1996) for each sampling location and year. Sampling identical populations in 2002 and 2003, one would expect no significant differentiation between samples taken at the same time during the two consecutive spawning seasons (e.g. RU02-1 and RU03-1). Such pairwise interannual comparisons of length data were done by pairwise t-tests, whereas Mann–Whitney U tests (Zar, 1996) were used to conduct pairwise comparisons of age. Statistical analyses of length, age, and length-at-age were carried out in S-PLUS ver. 6.1 (Insightful Corporation, 2002) and JMP ver. 3.2.2 (SAS Institute Inc, 1997).
Molecular genetic analyses
DNA was extracted from gill tissue using either the Chelex-proteinase K protocol of Estoup et al. (1996) or the Hot Sodium Hydroxide and Tris (HotSHOT) protocol of Truett et al. (2000). Nine microsatellite DNA loci were analysed: Cha1017, Cha1020, Cha1027, Cha1202 (McPherson et al., 2001), Cpa101, Cpa107, Cpa111, Cpa112, Cpa114 (Olsen et al., 2002). Microsatellite loci were amplified by polymerase chain reaction (PCR) using standard reagents and with annealing temperatures ranging between 54°C and 60°C for the different loci (exact protocols are available from the authors on request). The amplified microsatellite loci were analysed on a BaseStation 51TM DNA fragment analyser (MJ Research) and gels were typed using the software Cartographer 1.2.6 (MJ Geneworks, Inc.).
Statistical analyses of molecular data
Exact tests for conformity to Hardy–Weinberg expectations were conducted according to Guo and Thompson (1992) using the software GENEPOP 3.3 (Raymond and Rousset, 1995). Pairwise genetic differentiation between samples was analysed by Weir and Cockerham's (1984) 
, an unbiased estimator of FST, and by exact pairwise tests (Goudet et al., 1996) using FSTAT 2.9.3 (Goudet, 1995). The P values were compared to Dunn–
idák adjusted significance levels (Sokal and Rohlf, 1995) (k = 15, the number of sample comparisons at each location). Furthermore, a multidimensional scaling analysis, based on pairwise values between all samples, was used to visualize the genetic relationship among samples. This analysis was conducted using the program Vista 5.6.3 (Young, 1996).
We also conducted three hierarchical analyses of genetic differentiation (AMOVA, Analysis of Molecular Variance) using Arlequin 1.1 (Schneider et al., 2000). This approach is in principle a nested ANOVA based on the variance in allele frequencies at different levels: individuals within populations, populations within groups, and groups within the total genetic variation. In the first AMOVA, the total genetic variance (T) was divided into sampling years (representing groups, G) that were further divided into samples (representing populations, P) which, in turn, were divided into individuals (I). Following this nomenclature, FGT is genetic variance among groups, FPG is genetic variance among populations within groups, and FIP is genetic variance among individuals within populations. Two additional AMOVAs were set up to test for differences in genetic variation among age classes and between sexes: subsamples with individuals from the original samples separated by sex, and age-class subsamples formed by cohort (e.g. individuals of age 4 in 2002 and age 5 in 2003), including age classes 4–8 for Rügen and 4–9 for Gdask Bay. Only subsamples consisting of 
10 individuals were included in these analyses. The significance of variance components was tested by 16 000 permutations (i.e. by permuting alleles among individuals, individuals among samples, and samples among groups).
Finally, to further analyse the genetic affinities of individuals we used assignment tests, or more specifically we tested whether individuals were "accepted" or "rejected" in baseline samples (Paetkau et al., 2004). The principle of this test is to randomly generate "gametes" from each of the individuals of a sample and then generate a number of diploid individuals (in this case 10 000) based on the simulated gametes. Next, a frequency distribution of likelihood values is generated. The likelihood of each individual can then be compared with this distribution, and if the value is below a certain threshold, the individual is "rejected" from the sample. These analyses were performed using the software GENECLASS2 (Piry et al., 2004). The assignments were carried out using the Bayesian method of Rannala and Mountain (1997). We used the earliest temporal sample taken at each location as a "baseline sample" (i.e. RU02-1 from Rügen and GD02-1 from Gdask Bay) and then tested whether individuals from all samples were accepted or rejected from the baseline samples.
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Morphological diversity among samples
Length-at-age
The ANCOVA results showed a non-significant difference in length-at-age among the six RU samples (p = 0.804). The variance around the regression lines was substantial for both RU and GD samples (RU: 0.198
r2 0.374; GD: 6 x 10–6 R2 0.06), giving rise to a spurious, significant difference in length-at-age among GD samples (p = 0.018), as well as apparently negative length-at-age regressions (Table 2).
Length
Mean length (Table 2) differed significantly between samples within seasons at Rügen (0.000 < p 0.002) with decreasing mean length during the season, while the significant difference in mean length among the Gdask Bay samples (0.014 p 0.015) showed no temporal pattern.
Age
Among the Rügen samples, there were significant differences in age in 2003 (p < 0.001) but not in 2002 (p = 0.655). Age increased significantly in the course of both sampling years in the Gdask Bay (0.000 < p < 0.001).
Test for interannual differences in length and age showed significant differences in mean length in two of three comparisons (early, mid, and late season) at each location (RU: 0.004 p 0.770; GD: 0.000 < p 0.375) and differences in age distributions in all three tests for GD (0.0002 p 0.012) but not for RU (0.097 p 0.949).
Genetic diversity among samples
Complete genotypes for the nine loci were scored in 87–100% of the specimens in each sample. Tests for conformity to Hardy–Weinberg equilibrium revealed six cases of significant departures in the 108 single-locus tests after Dunn–idák adjusting the significance level to k = 9 (the number of loci per sample) (Table 3). This is no more than one would expect due to type 1 errors when 
= 0.05.
|
Values of
among samples within years ranged from 0 to 0.0025 in Gdask Bay and 0.0008 to 0.0113 at Rügen. The genetic differentiation was non-significant for all Gdask Bay samples (0.028 p 0.922) and the 2003 samples from Rügen (0.030 p 0.325) (Table 4). However, all three samples from Rügen in 2002 differed significantly from each other (0.0049 p 0.0113). Multidimensional scaling axes 1 and 2 explained 56% and 19% of the variance, respectively (Figure 2). RU02-1 had an isolated placing in the plot, and GD03-2 separated out along dimension 2.
|
|
The first hierarchical AMOVA estimated the percentage of the overall genetic variation found among sampling years, among samples within sampling year, and among individuals within samples, respectively. Results showed highly significant genetic differentiation among samples within sampling year at Rügen and significant genetic differentiation among temporal samples from Gda
sk Bay (Table 5). The genetic variation among individuals within samples also contributed significantly to the overall genetic variance both at Rügen and in Gdask Bay. In the following two AMOVAs, genetic variation due to sex and age classes was found to be non-significant at both spawning locations (Table 5). Even when the RU herring were divided into age classes, the genetic variance among samples (early, mid, and late in both 2002 and 2003) contributed significantly to the overall genetic variance found at Rügen.
|
The assignment analysis of Rügen samples resulted in more individuals sampled in 2002 being rejected from the baseline sample, RU02-1, with larger separation in time (12% for RU02-2 and 30% for RU02-3) (Table 6). All 2003 samples had rejection levels similar to the mid and late season samples from 2002. The rejection levels from the GD02-1 baseline sample ranged from 2% to 20%, but rejection levels were high in the mid sample in 2002 and the early and late samples in 2003 (Table 6).
|
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Morphological diversity among samples
The existence of distinct spawning groups of herring has previously been found in studies of length-at-age at Rügen and in several characters in Gda
sk Bay (e.g. Kompowski, 1971; Biester and Hering, 1977; Biester et al., 1979; Rechlin, 2000; Podolska and Horbowy, 2002). Other studies have shown that early spawners were larger and older (Biester and Hering, 1977; Biester et al., 1979) than late spawners, which is a pattern often found in herring (Lambert, 1987). The temporal samples used in the present study were tested for differences in mean length, age (significant difference would indicate that temporal samples comprised different portions of fish from the same population), and length-at-age (significant difference would indicate that temporal samples comprised individuals from separate populations). Unfortunately, large variations around length-at-age regressions render clear conclusions on this character impossible at both locations. The temporal patterns in mean length and age distributions varied between spawning seasons and locations, and interannual tests revealed significant differences between some of the samples taken at similar times during the spawning season. In summary, these characters did not lend strong support to either scenario (one or more populations spawning at each location), and results from genetic analyses were therefore considered without extensive comparisons with the length, age, and length-at-age results.
Genetic diversity among samples
Rügen
Pairwise comparisons for 2002 Rügen samples showed that all samples diverged significantly from each other. RU02-1 diverged significantly from all other samples, and between RU02-1 and RU02-3 was the largest between any two samples in this study (Table 4). A hierarchical AMOVA showed highly significant differentiation among samples (Table 5; FPG = 0.0041*** and FPG = 0.0038***). Among the 2003 samples, however, pairwise genetic differentiation between all samples was non-significant. The assignment tests also showed that rejection levels increased with increasing temporal separation in 2002, but not in 2003 (Table 6). In general, most samples were grouped together regardless of sampling location and time in the MDS plot (Figure 2). However, 25 of the 30 pairwise values ranged between 0.001 and 0.01, which is comparable to samples collected at a regional scale (e.g. Knutsen et al., 2003; McPherson et al., 2004).
Gdask Bay
The pairwise genetic differentiation was non-significant among all Gdask Bay samples despite the isolated placing of GD02-3 along dimensions 1 and 2 in the MDS plot (Figure 2). The assignment tests did not reveal any temporal pattern of genetic differentiation (Table 6). Still, there was small but statistically significant differentiation among the samples according to the hierarchical AMOVA (Table 5; FPG = 0.0011*).
There was significant genetic variation among individuals within the samples in Gdask Bay and at Rügen according to the hierarchical AMOVA, but this variation was not explained by genetic differentiation among age classes or between sexes (Table 5; age classes: FGT = –0.0006 and 3.0 x 10–5; sexes: FGT = –0.0006 and 0.0012).
Are samples adequately representing populations?
AMOVAs showed significant genetic variance among samples at both spawning locations. In addition, the 2002 Rügen samples showed significant pairwise genetic differentiation. Genetic differentiation among temporal samples from the same spawning location can be due to the presence of reproductively isolated populations, or it can be due to so-called "fluctuating genetic patchiness". The latter term comprises selection on early life stages (Hellberg et al., 2002), "sweepstakes reproduction" where few adults reproduce at the time and place of optimal oceanographic conditions, and low effective populations size, Ne (a measure of the number of adults that successfully contribute to reproduction), that can be caused by strongly biased reproductive success, size-dependent fecundity, and large variations in year-class strength (Hedgecock, 1994; Hauser et al., 2002). These factors may all lead to genetic differentiation among cohorts by selection or random drift. Fluctuating genetic patchiness is not expected when age frequency distributions do not differ among samples, as was the case for RU samples. Year-class strength showed large variation in the GD samples with a very strong 1994 age class, but non-significant genetic variance among age classes makes it less likely that genetic patchiness caused the significant genetic differentiation among GD samples. Furthermore, microsatellite DNA is non-coding and therefore presumably selectively neutral (Jarne and Lagoda, 1996). Also, it is unlikely that the sampled herring populations have small effective populations sizes, Ne, because the genetic variation at most loci is high, typically ranging between 0.4 and 0.9 and with the number of alleles ranging from two to four at Cpa107, up to 17–22 at Cha1027 (Table 3). Together with the consistent variance components attributable to sampling, this indicates that the identified genetic differentiation among temporal samples within the spawning season at Rügen and in Gdask Bay may be attributed to the sampling of populations showing more or less reproductive isolation.
Conversely, the significant interannual pairwise comparisons of all samples, except the early GD samples, and the non-significant pairwise genetic differentiation within seasons at GD and among the 2003 RU samples, together with the lack of divergence in length, age, and length-at-age suggest that the pattern of genetically divergent, temporal samples is less clear (Table 4). One factor that might cause an apparent fluctuating genetic patchiness at the two locations would be the sampling of different populations in spawning groups of herring do not show exactly home to natal spawning grounds. Numerous spawning grounds are located along the north coast of Germany and Poland (Tomas Gröhsler, German Institute for Baltic Sea Fisheries, Rostock, and Miroslaw Wyszynski, pers. comm.), thereby making it possible that spawners from other natal populations may visit the Rügen and Gdask Bay spawning grounds.
Sampling error might be another possible confounding factor. The non-significant genetic differentiation combined with the high rejection levels for all 2003 samples when compared to RU02-1 could reflect insufficient sampling in 2003 at Rügen. The majority of the early spawners may have disappeared from the spawning ground before the first sample was taken. The significant difference between the Gdask Bay samples in the hierarchical AMOVA and in growth rates indicate some genetic differentiation among samples, but the pairwise values and the rejection levels suggest no temporal pattern in differentiation among samples. The age frequency distributions for Gdask Bay had very high peaks of 1994 from individuals in both mid and late spawners, indicating that these samples may have been taken from the same populations. However, the mid and late samples were not less genetically differentiated from each other than each of them from the early spawners. It is possible that data on otolith microstructure and gillrakers, traits previously used to assist in the discrimination between spawning groups of herring in Gdask Bay (Strzy
ewska, 1969; Kompowski, 1971), could have helped to shed some light on whether the samples actually represented the expected three populations.
Finally, one may ask whether the use of microsatellite markers for the identification of temporal herring populations is appropriate. Polymorphism at microsatellite loci is typically very high in cold-water fish (Carvalho and Hauser, 1998). This permits "high resolution" analyses but it also leads to a deflation of the maximum FST obtainable between samples (Hedrick, 1999). Hence, an apparently low value may indicate important differentiation, and Wright (1978) pointed out that differentiation as low as 0.05 or even less is by no means negligible. However, even in the case of statistically significant differentiation it is still necessary to evaluate the biological significance of the differentiation (Waples, 1998; Hedrick, 1999). A well-prepared sampling design including sampling on the spawning grounds and at consistent times during the spawning season may considerably improve the quality of the data (Carvalho and Hauser, 1998).
The average FST value at Rügen was 0.0048, and in Gdask Bay it was 0.0007. These values were similar to those found by McPherson et al. (2003), who studied genetic differentiation in spawning waves of Atlantic herring at four locations off the east coast of Canada. In the latter study, each location was sampled twice during the spring-spawning season and the FST values ranged from 0 to 0.0058. Differentiation was significant at one of the four locations. Genetic differentiation has also been studied in herring at different geographical scales, with FST values ranging between 0 and 0.0142 at local scales (O'Connell et al., 1998; McPherson et al., 2004), and up to 0.0035–0.0681 at the pan-Atlantic scale (McPherson et al., 2004). To our knowledge, no other studies have been published which estimated genetic differentiation of Baltic spawning groups of herring using microsatellites, but a study of cod samples collected along the southern coast of Norway showed significant genetic differentiation among samples, with an average FST of 0.0023 (Knutsen et al., 2003). Studies of cod populations at regional scales typically yield FST estimates below 0.005 (e.g. Ruzzante et al., 1999, 2000). Thus, the genetic differentiation among samples within the Rügen spawning location is of the same order of magnitude as that for herring and cod at separate but geographically closely located spawning areas.
Spawning waves of herring may be genetically divergent, as indicated by the Devastation Shoal samples in McPherson et al. (2003) and the 2002 samples from Rügen. This may be due to sympatric divergence between spawning waves. One class of models of sympatric divergence comprises the ecological models where divergence is driven by competition for resources (Turelli et al., 2001). The decrease in competition for a suitable substratum for eggs and competition for food among larvae in large herring populations may lead to such sympatric divergence. Alternatively, Ware and Tanasichuck (1989) studied wave spawning in Pacific herring (Clupea harengus pallasi) and found that larger herring mature faster and that wave spawning therefore is determined by the difference in size-at-age combined with faster maturation rates of large individuals. They suggested that this might be a "bet-hedging" strategy to counter environmental uncertainties during the early life stages. Another driving force in models of sympatric divergence is disruptive selection (Turelli et al., 2001). Eggs and larvae of early and late spawners at Rügen and in Gdask Bay experience quite different environments during their development, and there may be a selective advantage for individuals adapted to mean temperature and salinity at the time of spawning and larval growth. For example, fast growth and early appearance on the spawning ground in herring, as opposed to slower growth and later arrival, have been suggested to make up two evolutionary strategies that can gradually become more pronounced and lead to actual reproductive isolation between individuals following the two strategies (sympatric population segregation) (Nævdal, 1972). In any case it follows that even separate spawning waves may hold part of the herring's evolutionary potential, and genetic differentiation between them would indeed be "biologically meaningful".
The proper use of the results from studies on population structure is important to ensure the success of fishery management programmes (Shaklee and Bentzen, 1998), including the ability of the exploited fish species to adapt to future changes in the environment (Hedrick and Miller, 1992; Lande and Shannon, 1996). The results in this study suggest that care should be taken to collect samples from different years consistently in the same part of the spawning season to be comparable when studying spatial herring population structure, and it emphasizes the importance of studying population structure at both regional and local scales.
|
Acknowledgements |
|---|
We thank Henrik Mosegaard at the Danish Institute for Fisheries Research (DIFRES) for suggesting a temporal genetic analysis of Rügen herring. Samples from Gda
sk Bay and Polish literature were kindly provided by Miroslaw Wyszynski at the Sea Fisheries Institute in Gdynia, and Joachim Dröse at the Institut für Ostseefischerei Rostock, Germany, helped us obtain samples from the Greifswalder Bodden. We also thank Stina Bilstrup for otolith readings of Rügen herring and Dorte Meldrup, Karen-Lise D. Mensberg, and Tina Brandt Christensen at DIFRES for technical help and advice in the DNA laboratory. Finally, we thank Daniel Ruzzante, Dalhousie University, Canada, for advice on statistical analyses and Dorte Bekkevold, DIFRES, for producing the sample map. The project was supported by the Danish Network for Fisheries and Aquacultural Research (www.fishnet.dk) financed by the Danish Ministry for Food, Agriculture and Fisheries and the Danish Agricultural and Veterinary Research Council. |
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Andersson L., Ryman N., Rosenberg R., Ståhl G. (1981) Genetic variability in Atlantic herring (Clupea harengus harengus): description of protein loci and population data. Heredity 95:69–78.[CrossRef]
Anon. (1962) Recommendations adopted by the Herring Committee. Rapports et Procès-Verbaux du Conseil International pour l'Exploration de la Mer (Appendix 1): 71–73.
Aro E. (1989) A review of fish migration patterns in the Baltic. Rapports et Procès-Verbaux du Conseil International pour l'Exploration de la Mer 190:72–96.
Biester E. (1979) Der Frühjahrshering Rügen – seine Rolle in der Fischerei der Ostsee und in den Übergangsgebieten zur Nordsee. Inaugural Dissertation. Rostock. 238 pp.
Biester E. and Hering P. (1977) Changes in the population structure of the herring in the Greifswalder Bodden during spawning time in spring. ICES CM 1977/P: 20. 12 pp.
Biester E., Jönsson N., Hering P., Thieme Th., Brielmann N., Lill D. (1994) Studies on Rügen herring 1979. ICES CM 1979/J: 32. 9 pp.
Carvalho G.R. and Hauser L. (1979) Molecular genetics and the stock concept in fisheries. Reviews in Fish Biology and Fisheries 4:326–350.[CrossRef]
Carvalho G.R. and Hauser L. (1998) Advances in the molecular analysis of fish population structure. Italian Journal of Zoology 65:Suppl., 21–33.[Medline]
DeWoody J.A. and Avise J.C. (2000) Microsatellite variation in marine, freshwater and anadromous fishes compared with other animals. Journal of Fish Biology 56:461–473.[CrossRef][Web of Science]
Estoup A., Largiader C.R., Perrot E., Chowrout D. (1996) Rapid one-tube DNA extraction for reliable PCR detection of fish polymorphic markers and transgenes. Molecular Marine Biology and Biotechnology 5:295–298.[Web of Science]
Gíslason D., Ferguson M.M., Skúlason S., Snorrason S.S. (1999) Rapid and coupled phenotypic and genetic divergence in Icelandic Arctic char (Salvelinus alpinus). Canadian Journal of Fisheries and Aquatic Sciences 56:2229–2234.
Goudet J. (1995) FSTAT version 1.2: a computer program to calculate F statistics. Journal of Heredity 86:485–486.
Goudet J., Raymond M., Demeeus T., Rousset F. (1996) Testing differentiation in diploid populations. Genetics 144:1933–1940.[Abstract]
Guo S.W. and Thompson A. (1992) Performing the exact test of Hardy–Weinberg proportion for multiple alleles. Biometrics 48:361–372.[CrossRef][Web of Science][Medline]
Hauser L., Adcock G.J., Smith P.J., Ramírez J.H.B., Carvalho G.R. (2002) Loss of microsatellite diversity and low effective population size in an overexploited population of New Zealand snapper (Pagrus auratus). Proceedings of the National Academy of Sciences of the USA 99:11742–11747.
Hay D.E. (1985) Reproductive biology of Pacific herring (Clupea harengus pallasi). Canadian Journal of Fisheries and Aquatic Sciences 42:Suppl. 1, 111–126.
Hedgecock D. (1994) Does variance in reproductive success limit effective population sizes of marine organisms? In Beaumont A. (Ed.). Genetics and Evolution of Aquatic Organisms(Chapman and Hall, London) pp. 122–134.
Hedrick P.W. (1999) Perspective: highly variable loci and their interpretation in evolution and conservation. Evolution 53:2313–318.[CrossRef][Web of Science]
Hedrick P.W. and Miller P.S. (1992) Conservation genetics: techniques and fundamentals. Ecological Applications 2:30–46.[CrossRef][Web of Science]
Hellberg M.E., Burton R.S., Neigel J.E., Palumbi S.R. (2002) Genetic assessment of connectivity among marine populations. Bulletin of Marine Science 70:Suppl., 273–290.[Web of Science]
Insightful Corporation. (2002) S-SPLUS ver. 6.1.
Jarne P. and Lagoda P.J.L. (1996) Microsatellites, from molecules to populations and back. Trends in Ecology and Evolution 11:424–429.[CrossRef]
Knutsen H., Jorde P.E., Andre C., Stenseth N.C. (2003) Fine-scaled geographic population structuring in a highly mobile marine species: the Atlantic cod. Molecular Ecology 12:385–394.[CrossRef][Medline]
Kompowski A. (1971) Types of otoliths in herring from the southern Baltic. Prace Morskiego Instytutu Rybackiego, Tom 16, Seria A 107–139.
Lambert T.C. (1987) Duration and intensity of spawning in herring Clupea harengus as related to the age structure of the mature population. Marine Ecology Progress Series 39:209–220.[Web of Science]
Lande R. and Shannon S. (1996) The role of genetic variation in adaptation and population persistence in a changing environment. Evolution 59:434–437.
McPherson A.A., O'Reilly P.T., McParland T.L., Jones M.W., Bentzen P. (2001) Isolation of nine novel tetranucleotide microsatellites in Atlantic herring (Clupea harengus). Molecular Ecology Notes 1:40.
McPherson A.A., O'Reilly P.T., Taggart C.T. (2004) Genetic differentiation, temporal stability, and the absence of isolation by distance among Atlantic herring populations. Transactions of the American Fisheries Society 133:434–446.[CrossRef]
McPherson A.A., Stephenson P.L., Taggart C.T. (2003) Genetically different Atlantic herring Clupea harengus spawning waves. Marine Ecology Progress Series 247:303–309.[Web of Science]
McQuinn I.H. (1997) Metapopulations and the Atlantic herring. Reviews in Fish Biology and Fisheries 7:297–329.[CrossRef][Web of Science]
Mork J., Ryman N., Ståhl G., Utter F., Sundnes G. (1985) Genetic variation in Atlantic cod (Gadus morhua) throughout its range. Canadian Journal of Fisheries and Aquatic Sciences 42:1580–1587.
Nielsen E.E., Hansen M.M., Ruzzante D.E., Meldrup D., Grønkjær P. (2003) Evidence of a hybrid-zone in Atlantic cod (Gadus morhua) in the Baltic and the Danish Belt Sea revealed by individual admixture analysis. Molecular Ecology 12:1497–1508.[CrossRef][Medline]
Nævdal G. (1972) Rasedannelse hos sild og brisling. Fauna 25:1–17.
O'Connell M., Dillon M.C., Wright J.M., Bentzen P., Merkouris S., Seeb J. (1998) Genetic structuring among Alaskan Pacific herring populations identified using microsatellite variation. Journal of Fish Biology 53:150–163.[CrossRef][Web of Science]
Olsen J.B., Lewis C.J., Kretchmer E.J., Wilson S.L., Seeb J.E. (2002) Characterization of 14 tetranucleotide microsatellite loci derived from Pacific herring. Molecular Ecology Notes 2:101–103.[CrossRef][Web of Science]
Ojaveer E. (1990) On the condition and management of the herring stocks in the Baltic. Proceedings of the International Herring Symposium October 1990Anchorage, Alaska pp. 521–531.
Paetkau D., Slade R., Burden M., Estoup A. (2004) Genetic assignment methods for the direct, real-time estimation of migration rate: a simulation-based exploration of accuracy and power. Molecular Ecology 13:55–65.[CrossRef][Medline]
Parmanne R., Rechlin O., Sjöstrand B. (1994) Status and future of herring and sprat stocks in the Baltic Sea. Dana 10:29–59.
Peichel C.L., Nereng K.S., Ohgi K.A., Cole B.L.E., Colosimo P.F., Buerkle C.A., Schluter D., Kingsley D.M. (2001) The genetic architecture of divergence between threespine stickleback species. Nature 414:901–905.[CrossRef][Medline]
Podolska M. and Horbowy J. (2002) Infection of Baltic herring (Clupea harengus membras) with Anisakis simplex larvae, 1992–1999: a statistical analysis using generalized linear models. ICES Journal of Marine Science 60:85–93.[CrossRef][Web of Science]
Piry S., Alapetite A., Cornuet J-M., Paetkau D., Baudoin L., Estoup A. (2004) GeneClass2: a software for genetic assignment and first generation migrants detection. Journal of Heredity 95:536–539.
Rajasilta M., Eklund J., Hänninen J., Kurkilahti M., Kääriä J., Rannikko P., Soikkeli M. (1993) Spawning of herring (Clupea harengus membras L.) in the Archipelago Sea. ICES Journal of Marine Science 50:233–246.
Rannala B. and Mountain J.L. (1997) Detecting immigration by using multilocus genotypes. Proceedings of the National Academy of Science of the USA 94:9197–9201.[CrossRef]
Raymond M. and Rousset F. (1995) Genepop (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity 86:248–249.
Rechlin O. (2000) Difference in length growth of western Baltic Sea spring spawning herring. Archive of Fisheries and Marine Research 48:153–160.
Rechlin O.K. (1991) Tendencies in the herring population development of the Baltic Sea. Internationale Revue der Gesamten Hydrobiologie 76:405–412.[CrossRef]
Ruzzante D.E., Taggart C.T., Cook D. (1999) A review of the evidence for genetic structure of cod (Gadus morhua) populations in the NW Atlantic and population affinities of larval cod off Newfoundland and the Gulf of St. Lawrence. Fisheries Research 43:79–97.[Medline]
Ruzzante D.E., Taggart C.T., Doyle R.W., Cook D. (2001) Stability in the historical pattern of genetic structure of Newfoundland cod (Gadus morhua) despite the catastrophic decline in population size from 1964 to 1994. Conservation Genetics 2:257–269.[CrossRef]
Ruzzante D.E., Wroblewski J.S., Taggart C.T., Smedbol R.K., Cook D., Goddard S.V. (2000) Bay-scale population structure in coastal Atlantic cod in Labrador and Newfoundland, Canada. Journal of Fish Biology 56:431–447.[CrossRef][Web of Science]
Ryman N., Lagercrantz U., Andersson L., Chakraborty R., Rosenberg R. (1984) Lack of correspondence between genetic and morphologic variability patterns in Atlantic herring (Clupea harengus). Heredity 53:687–704.[Web of Science]
Ryman N., Utter F., Laikre L. (1995) Protection of intraspecific biodiversity of exploited fishes. Reviews in Fish Biology and Fisheries 5:417–446.[CrossRef][Web of Science]
SAS Institute Inc. (1997) JMP ver. 3.2.2.
Schneider S., Kueffer J-M., Roessli D., Excoffier L. (2000) Arlequin ver 2.000. A Software for Population Genetics Data Analysis(Genetics and Biometry Lab., Dept. of Anthropology, University of Geneva, Switzerland).
Shaklee J.B. and Bentzen P. (1998) Genetic identification of stocks of marine fish and shellfish. Bulletin of Marine Science 62:589–621.[Web of Science]
Sokal R.R. and Rohlf F.J. (1995) Biometry(W.H. Freeman and Company, New York) 850 pp.
Strzyewska K. (1969) A comparison study on herring populations spawning off Polish Baltic coasts. Prace Morskiego Instytutu Rybackiego, Tom 16, Seria A 211–277.
Taylor E.B. (1999) Species pairs of north temperate freshwater fishes: evolution, taxonomy, and conservation. Reviews in Fish Biology and Fisheries 9:299–324.[CrossRef][Web of Science]
Truett G.E., Heeger P., Mynatt R.L., Truett A.A., Walker J.A., Warman M.L. (2000) Preparation of PCR-quality mouse genomic DNA with hot sodium hydroxide and Tris (HotSHOT). BioTechniques 29:52–54.[Web of Science][Medline]
Turelli M., Berton N.H., Coyne J.A. (2001) Theory and speciation. Trends in Ecology and Evolution 16:330–343.[CrossRef]
Urho L. (1999) Relationship between dispersal in larvae and nursery areas in the Baltic Sea. ICES Journal of Marine Science 56:Suppl., 114–121.[CrossRef][Web of Science]
Waples R.S. (1998) Separating the wheat from the chaff: patterns of genetic differentiation in high gene flow species. Journal of Heredity 89:438–450.
Ward R.D., Woodmark M., Skibinski D.O.F. (1994) A comparison of genetic levels in marine, freshwater, and anadromous fishes. Journal of Fish Biology 44:213–232.[CrossRef][Web of Science]
Ware D.M. and Tanasichuck R.W. (1989) Biological basis of maturation and spawning waves in Pacific herring (Clupea harengus pallasi). Canadian Journal of Fisheries and Aquatic Sciences 46:1776–1784.
Weir B.S. and Cockerham C.C. (1984) Estimating F-statistics for the analysis of population structure. Evolution 38:1358–1370.[CrossRef][Web of Science]
Wright S. (1978) Variability Within and Among Natural Populations. Evolution and the Genetics of Populations(University of Chicago Press, Chicago) vol. 4: 590 pp.
Young F.W. (1996) ViSta: The Visual Statistics System 2nd edn. (L.L. Thursone Psychometric Laboratory, University of North Carolina, Chapell Hill, NC) Research Memorandum 94-1(b).
Zar J.H. (1996) Biostatistical Analysis(Prentice-Hall, New Jersey) 662 pp.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||