© 2006 International Council for the Exploration of the Sea
Annual changes in the proportions of wild and hatchery Atlantic salmon (Salmo salar) caught in the Baltic Sea
Finnish Game and Fisheries Research Institute PO Box 2, FIN-00791 Helsinki, Finland
*Correspondence to M-L. Koljonen: tel: +358 205 751 315; fax: +358 205 751 201. e-mail: marja-liisa.koljonen{at}rktl.fi.
DNA-level information from an eight-loci microsatellite baseline database of 32 Atlantic salmon (Salmo salar) stocks was used with a Bayesian estimation method to assess the stock and stock group proportions of Finnish salmon catches in the Baltic Sea area. The proportions of seven stock groups, important to fisheries management, were assessed in catch samples taken between 2000 and 2005. In the Gulf of Bothnia area, the proportion of wild fish in catches showed an increasing trend in all areas until 2003, mainly because of the decrease in total catches caused by the relatively greater mortality of hatchery-reared fish compared with wild fish. In 2004, the total number of wild fish caught had also increased, indicating an increase in the abundance of wild stocks. In catches from the Åland Sea, the proportion of wild fish increased from 44% in 2000 to 70% in 2004, while the catch during the same period increased from 4628 to 7329 fish. In the Gulf of Finland, the local Neva salmon stock, which is released by Estonia, Finland, and Russia, made the largest contribution. In the western part of the Gulf of Finland, fish originating in the Baltic Main Basin also made a substantial contribution to catches. The threatened eastern Estonian and Russian wild stocks were recorded only in the western part of the Gulf of Finland, where the proportion of wild fish increased from 9% in 2003 to 19% in 2004.
Keywords: Atlantic salmon, DNA, microsatellites, stock mixture analysis
Received 30 September 2005; accepted 7 April 2006.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Most Atlantic salmon (Salmo salar) fisheries in the Baltic Sea catch mixtures of fish from different river stocks and of both wild and hatchery-reared origin. In the 1960s and 1970s, extensive hatchery rearing and release programmes were established to compensate for lost smolt production in rivers dammed for the generation of hydroelectricity. In 1997–2004, the number of smolts released annually was relatively stable, ranging from 5.0 to 5.8 million within the whole Baltic Sea (ICES, 2005). As the management goals for wild and hatchery-reared salmon differ, the fish must be harvested at different intensities or with different strategies. Therefore, it is important to know in which fisheries and in what quantities the wild salmon stocks are exploited. Genetic differences among fish stocks can be used to estimate stock proportions within catches. Genetic tags have the following advantages over external tags: there are no costs associated with applying tags; no tags are lost; there is no need to consider the effects of the tags on the viability and catchability of the fish. Moreover, all fish are tagged for life, so studies can be conducted on fish that cannot be tagged by other methods, e.g. wild fish in remote areas. Wild stocks, in particular, can be studied on the same terms as hatchery stocks. This is a definite advantage because usually only hatchery salmon are tagged. In assessment modelling, the parameters derived from hatchery salmon are often applied to wild stocks, even though the population dynamics of the two stock types frequently differ.
With genetic stock identification, the time and place of sampling can be chosen more freely and precisely than with external tagging because it is not dependent on tagging and release programmes. Furthermore, there is no need to assess the reporting rate of the tags returned by fishers. In addition, genetic data can be combined with non-genetic data (e.g. scale characteristics and smolt age). Restrictions on genetic mixed stock analysis are imposed only by the limited genetic differentiation among baseline stocks, which can be critical to the quality of the subsequent analysis. The method's accuracy relies on the baseline samples being representative of the genetic characteristics of the stocks that could potentially contribute to the fishery. For a more detailed evaluation of the method, see Koljonen et al. (2005). Further, genetic stock structure information can be used to define management units based on genetic similarities between stocks (Koljonen et al., 1999; Koljonen, 2001).
Genetic mixed stock analysis (MSA) has traditionally been based on allozyme data and maximum likelihood estimation (MLE) (Fournier et al., 1984; Pella and Milner, 1987). The method determines the relative contributions of baseline stocks with the highest likelihood of providing the observed multilocus genotype frequencies in the catch sample. Allozyme data have also been used to analyse the stock composition of Baltic salmon catches (Koljonen and Pella, 1997). Today, allozyme data can be replaced by DNA microsatellite data and MLE by Bayesian methods. Both of these new approaches increase the resolution power of the stock composition estimation.
DNA microsatellite variation has greatly increased the amount of genetic information available on Baltic salmon stocks (Koljonen et al., 2005; Säisä et al., 2005). Pella and Masuda (2001) have developed a Bayesian estimation method that is particularly useful for microsatellite data with a large number of alleles. This study demonstrates the power and use of microsatellite data and the Bayesian method in assessing the annual variation in stock and stock group proportions in Atlantic salmon catches in the Baltic Sea fishery.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Sampling of fish
The baseline data of potentially contributing salmon stocks were gathered by taking tissue samples from 2337 Atlantic salmon belonging to 32 stocks in rivers draining into the Baltic Sea (Figure 1, Table 1). The catch samples, totalling 5362 fish, were collected during five years (2000–2005) from six Atlantic salmon fisheries in the Baltic Sea (Figure 1, Table 2). In 2000, three sampling sites along the Finnish coast of the Gulf of Bothnia were included (Åland Sea: catch sampling site 1; Bothnian Sea: catch sampling site 2; and Bothnian Bay: catch sampling site 3; see Figure 1). At these sites, the samples were taken along the route of the northward spawning migration, which extends from the Baltic Main Basin across the Åland Sea to the northern rivers in Finland and Sweden. Sampling was conducted on catches from the coastal trapnet and driftnet fishery (see Table 2). From 2002, catch samples were also collected from two sites in the Gulf of Finland (eastern Gulf of Finland: catch sampling site 4; and western Gulf of Finland: catch sampling site 5), and from the international fishery in the Baltic Main Basin (catch sampling site 6). In the eastern part of the Gulf of Finland, salmon were caught with trapnets during their spawning migration, and in the western part of the Gulf and in the Main Basin with longlines and driftnets during their feeding migration in late autumn. The samples from the Gulf of Bothnia and the Gulf of Finland were representative either of Finnish catches in each area and the fishing period or, in some cases, of the whole spawning run at the site.
|
|
|
In the Gulf of Bothnia, the migration patterns and run timings of wild and hatchery Atlantic salmon populations differ slightly. Wild fish tend to return to their natal rivers earlier than hatchery fish, resulting in a larger proportion of wild fish among the first arrivals than among fish arriving later in the migration season. To increase the escapement of wild fish into the spawning rivers, the fishery is regulated by staggering the opening days in four management sectors sequentially from south to north along the coast of Finland during the summer. The opening dates for the Bothnian Sea sector, the southern sector of Bothnian Bay, the northern sector of Bothnian Bay, and the sector close to the Tornionjoki River mouth were 16 June, 21 June, 26 June, and 1 July, respectively (the opening dates and the Finnish fishing sectors are shown in Figure 1). The Åland Sea area is an exception, as fishing is allowed throughout the spawning run. Thus, the Åland Sea samples from the strait between Finland and Sweden can be expected to represent all fish returning to the entire Gulf of Bothnia area, and not only catches for a particular fishery. In addition, the whole of the spawning run was sampled in the Bothnian Sea and Bothnian Bay in 2002 and in the Bothnian Sea in 2003 (Table 2).
The samples were taken separately from three of the four coastal regulation sectors. The fish captured at each sampling site were measured, and scale samples were taken daily from a subsample stratified by size. The ages of the fish were determined by scale reading, and the DNA subsample was taken from the fish stratified by sea age (1–5 sea years). The fish caught during the spawning run were relatively large. For example, in 2005 the mean weight of the individuals in the Åland Sea catches was 7 kg (3.1–15.9 kg), the mean total length was 89.5 cm (72–113 cm), and the mean sea age was 2.2 years. From 2002 to 2004, samples were also collected on random days during the fishing season in the Baltic Main Basin fishery from catches landed on Bornholm Island, Denmark. These samples were not representative of the whole Main Basin fishery.
Microsatellite DNA analyses
The multilocus genotype frequencies of baseline stocks and catch samples were used to estimate the stock compositions of the catches. Total genomic DNA was extracted from muscle tissue, adipose fins, or scale samples. Variation was determined at eight microsatellite loci: Ssa85, Ssa289 (McConnell et al., 1995), Ssa171, Ssa197, Ssa202 (O'Reilly et al., 1996), SSOSL85, SSOSL417 (Slettan et al., 1995), and SSOSL438 (Slettan et al., 1996).
The polymerase chain reaction (PCR) and DNA labelling were conducted as described in Säisä et al. (2005). Microsatellite genotypes were detected with a Li-Cor automated DNA sequencer (Li-Cor, Inc., Lincoln, NE, USA) and analysed with Gene ImagIRTM fragment software (version 3.52 Scananalytics).
Analyses of genetic differentiation
The genetic differentiation between stocks for eight loci was quantified with the DA distance of Nei et al. (1983). A dendrogram was constructed using the neighbour-joining (NJ) method (Saitou and Nei, 1987). The bootstrap test of loci was performed for the NJ tree by recalculating the distance for all loci 1000 times. Distance analysis was conducted and mean heterozygosities (Nei, 1973) were computed with the DISPAN package (Ota, 1993).
Estimation method
For genetic mixed stock analysis (MSA), mixture modelling using the Bayesian estimation method was implemented using the BAYES program (Pella and Masuda, 2001). The program calculates posterior probability distributions for unknown stock proportions in the mixed-catch samples. The posterior distribution for the stock proportions combines prior baseline information with information on the mixed-catch sample to estimate the stock composition of the mixture sample while updating the multi-locus genotype distributions of the baseline stocks for each draw. Samples of the posterior distributions were drawn by Markov chain Monte Carlo (MCMC) methods. The outcome of the analysis, i.e., the result of the stock and stock group proportion estimation, is expressed in terms of probability distributions.
The number of chains run per application (for each catch sample) for the posterior probability distributions for the stock proportion estimates ranged from 20 (for the Gulf of Bothnia samples) to 32 (for the Gulf of Finland and Main Basin samples). The convergence of the chains to the posterior probability distribution was tested and approved for each catch sample (Gelman and Rubin, 1992). The last 1000 MCMC iterations of each 5000-iteration chain were combined and used to describe the posterior probability distributions of the proportions of each baseline stock and of the seven stock groups within the catch samples. Since medians of the posterior distributions were used as proportion estimates, they do not always sum exactly to 1.0. For a more detailed explanation of the method, see Koljonen et al. (2005).
Proportions were always estimated first for the 32 individual baseline stocks. Then these estimates were pooled for the seven stock groups useful for management. When these individual stock proportion estimates are pooled, the genetic similarity among them is reflected in the resulting probability interval for the group estimate.
The proportion estimates were also used to assess the total number of fish in each stock group in the Åland Sea fishery, which was assumed to represent the relationships of the stock proportions in the entire Gulf of Bothnia area, partly because of the central location of the sampling site, and partly because fishing covered the whole spawning run and the number of fishers was relatively small and stable (about 15), resulting in little annual variation in fishing effort, which was measured as driftnet days.
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Genetic differentiation between baseline stocks
The genetic differentiation between the baseline stocks varied substantially, forming a pattern (Figure 2) similar to that observed previously in allozyme data (Koljonen et al., 1999). The northern Baltic Sea stocks from Finland and Sweden in the Gulf of Bothnia constituted one compact branch of the dendrogram of genetically similar stocks, which are of Atlantic origin (Koljonen et al., 1999; Säisä et al., 2005). The eastern and southern Baltic Sea stocks, from Russia, Estonia, and Latvia, formed another branch, which originated after glaciation in eastern refugial glacial lakes (Koljonen et al., 1999; Säisä et al., 2005). Two stocks from rivers in southern Sweden (Emån and Mörrumsån), which drain into the southern Baltic Main Basin, formed a distinct intermediate branch, the origin of which cannot be explained by either western or eastern connections. It has been suggested that these stocks originated in southern refugial areas; for a more detailed analysis see Säisä et al. (2005). The GST over all baseline populations and eight loci was 0.12, and the mean heterozygosity was 0.69, ranging from 0.58 for the Kunda River to 0.76 for the Ljungan River.
|
Stock group proportions
Gulf of Bothnia
In the Gulf of Bothnia area, sampling sites 1–3 (Åland Sea, Bothnian Sea, and Bothnian Bay; Figure 1), Atlantic salmon catches consisted mainly of three stock groups: wild fish originating in rivers draining into the Gulf of Bothnia (group 1), and hatchery fish originating in Finnish (group 2) and Swedish (group 3) releases into this area (Table 3). The combined proportion of other stock groups accounted for <3% in all years; these proportions are not shown in the figures for Gulf of Bothnia catches (Figure 3a–c). The major river components in the wild stock group were from the Tornionjoki and Kalixälven Rivers.
|
|
The proportion of wild fish in catches in the Gulf of Bothnia area has been increasing at all three sampling sites since 2000. In the southern part, the Åland Sea, the proportion of wild fish rose from 44% (see probability intervals in Table 3) in 2000 to 84% in 2003, while coincidentally the proportion of Swedish hatchery fish decreased from 40% to 1.6%. The proportion of Finnish hatchery stocks remained relatively constant (13%–16%) during the same period (Table 3, Figure 3a). In 2004, the proportion of Swedish hatchery fish increased to 14%, but decreased again in 2005 to 6.6%.
In the more northerly coastal Bothnian Sea (sampling site 2), the major contribution (45%–57%) to catches in this area between 2000 and 2002 was from Finnish hatchery fish. However, in 2003 and 2004, the proportion of these fish fell to 22% and 26%, respectively, of the total catch and to about one-third in 2005. The proportion of Swedish hatchery fish fell to almost zero in 2002 (Figure 3b), but increased to 11% in 2005. At the same time, the proportion of wild stocks in the catches increased from 38% in 2000 and 40% in 2002 to a peak of 77% in 2003 (Table 3). The locally released Neva salmon made a marked contribution to the Finnish hatchery group in all years, 7%–27% of the catches.
In the most northerly area, near the spawning rivers of Bothnian Bay, the main contributors to catches were wild stocks and Finnish hatchery stocks, whereas Swedish hatchery fish accounted for no more than 12% (in 2000) of the total. A clear trend in wild stock proportions could also be seen, increasing from 40% in 2000 to 64% in 2003 and to 67% in 2005 (Figure 3c).
Gulf of Finland
The composition of catches was more diverse in the Gulf of Finland than in the Gulf of Bothnia, and all seven potential stock groups made some contribution to catches in either the eastern or the western part of the Gulf (Table 3, Figure 3d and e). In both areas, however, the largest group in the catches (41%–88%) came from hatchery releases into the Gulf of Finland (group 5, Neva and Narva stocks). Since 2003, northern stocks, originating in the Gulf of Bothnia, have also made a significant contribution to the eastern Gulf of Finland catch during the period of the spawning migration. In June, these northern wild stocks accounted for, at most, about one-third of the total catch (24%–34%) (Figure 3d), and hatchery-released fish for no more than 12% when the catch levels in general were low. Eastern wild stocks did not occur in this trapnet fishery, which targets at spawning-run fish, because the home rivers of the stocks are on the south coast of the Gulf of Finland.
In the Gulf of Finland, local wild fish from Estonia and Russia (group 4: Luga, Kunda, and Keila) were observed in longline catches from the western part of the Gulf, comprising 10% of these feeding migration fish in 2002 and 19% in 2003 (Figure 3e). In addition to fish released into the Gulf of Finland from hatcheries (41%–57%), salmon stocks from the eastern Main Basin (including the Daugava and Gauja Rivers from Latvia) made a clear contribution to the Finnish autumn fishery in the western part of the Gulf of Finland (27% in 2002 and 17% in 2003). These stocks tend to migrate northwards from their home rivers for feeding.
Although the samples from the Baltic Main Basin were not representative of the total fishery, they still indicate the stock group composition of catches in this international fishery (Table 3, Figure 3e). The main component, and about half of the catches (49%–71%), came from the northern, Gulf of Bothnia, wild stocks, mainly from the Tornionjoki and Kalixälven Rivers, but fish from other wild stocks, e.g. the Swedish Byskeälven and Vindelälven Rivers, were also caught in the Main Basin. No eastern, Gulf of Finland, stocks were observed in these catches. Fish from rivers draining into the eastern and western Main Basin, the Mörrumsån and Emån Rivers, provided a maximum of 8% combined, and thus made a small contribution. In contrast to the stock composition of Finnish catches in the Gulf of Bothnia area, Swedish hatchery stocks appear to have made a greater contribution (14–27%) to these Main Basin catches than Finnish hatchery stocks (0–3%), even in 2003 and 2004, when their proportion was small on the Finnish coast.
The proportion of wild stocks in Gulf of Bothnia catches showed an increasing trend from 2000 to 2003. However, the numbers of wild fish in the catches did not increase at the same rate because the total catch fell sharply in 2002 and 2003, only to increase again in 2004 (Table 4, Figure 4). When stock group proportions were used to estimate the number of fish in each stock group based on the total number of fish taken in Åland Sea catches, a distinct decreasing trend was seen in the number of Swedish hatchery fish caught during the first three years, the number declining from about 4200 individuals in 2000 to approximately 80 in 2003 (see probability intervals in Table 4). However, in 2004, the number of Swedish hatchery fish increased to about 1500, when the total catch from the Åland Sea was approximately the same as in 2000, about 10 500 fish. In 2005, the total catch decreased again, and the number of fish caught was about the same as in 2002.
|
|
The number of wild salmon caught in 2003 (4207 fish) was smaller than in 2000 (4628 fish), but the proportion of wild fish was larger, 83% in 2003 compared with 44% in 2000. In 2004, however, the total number of wild fish caught increased for the first time to more than 7000 in the Åland Sea fishery. Assuming relatively stable fishing effort and only a minor effect of catchability variation, this increase can indicate a greater abundance of wild fish, as the total catch, hatchery plus wild fish, had increased.
If the wild stock proportion estimates are used to evaluate the effect of regulation of the fishery, i.e. the variation in the time the fishery is opened in different fishing sectors, the results suggest that the effect varied annually, depending on the annual timing of the spawning run (Table 5). During the three years when sampling followed the fishing regulation (2000, 2004, and 2005), the proportion of wild fish was 9.2% smaller, on average, in the Bothnian Sea fishery and 18.4% smaller in the Bothnian Bay fishery than in the Åland Sea catches.
|
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Genetic mixed stock analysis was valuable in revealing the variation in catch composition that would otherwise have gone undetected. It also permitted direct observation of the proportions of salmon from different conservation or management units. Here, the stock composition of the catches differed considerably in different parts of the Baltic Sea, most notably between the Gulf of Bothnia and the Gulf of Finland. Thus, the fish caught in those areas originate from different source populations. Significantly, the same salmon stocks tended to occur in the same sea areas in all years, in either the Gulf of Finland or the Gulf of Bothnia, indicating relatively constant annual distributions. However, the proportions of individual stocks within each area and fishery varied considerably from one year to the next, the wild and hatchery components in particular revealing dissimilar patterns. As the catches in different areas come from different source populations, each local fishery should be managed according to the production level of the contributing wild stocks in those areas: in the Gulf of Bothnia according to the production levels of Finnish and Swedish wild stocks and in the Gulf of Finland according to Russian and Estonian wild stocks. When the proportion of wild stocks in the total catches is known, this information can also be used in assessment modelling to improve the estimated production levels of wild stocks (Michielsens et al., 2004).
Proportion estimates can be used to assess catch numbers for certain stocks or stock groups when the total catch numbers are known and when the sampling represents a particular fishery. Stock proportion estimates do not directly indicate the numbers of fish caught, as catch numbers vary markedly. Here, the increased proportion of wild fish did not necessarily mean larger numbers of wild fish in catches. The number of released hatchery fish has been very stable for several years. Finland and Sweden have been releasing 3.3–3.6 million smolts annually since 1998, the first year in which such releases would contribute to the catches examined in this study, and the proportion of Swedish fish has consistently accounted for slightly more than half of the total (52–56%; Table 6; ICES, 2005). Owing to the stability of hatchery releases, the variation in post-smolt mortality remains the main factor explaining the variation in the proportion of the hatchery component in the catches. Some of the changes in stock proportions may be the result of variations in migration routes, but this is unlikely to explain the large changes. The poorer post-smolt survival of the hatchery stocks has also been observed in the Carlin tagging data (Michielsens et al., 2006). One potential cause of spawner mortality in the Baltic Sea is the M-74 syndrome, the intensity of which was lower in 2003 and 2004, when the percentage of affected females decreased to an average of <10% (ICES, 2005). The possible increased post-smolt mortality of wild stocks is partly compensated for by the increased wild smolt production.
|
Stock proportion estimates derived by using mixed stock fishery analysis can potentially be used as an index of the abundance of populations under certain conditions because catch samples should represent the populations of interest. However, information about stock proportions in catches does not correlate automatically and directly with stock abundance in the sea or with smolt production levels in rivers. Catch samples should be representative of the populations of interest, and these populations should be well represented in the catches. The temporal and spatial distributions of fish stocks must also be taken into account. As a result of their schooling behaviour, salmon are not usually evenly distributed in the sea, and their feeding and spawning migrations may vary. In addition, the potential variation in fishing effort should be considered. Harvesting always depends on fishing effort, on temporal and spatial fishing regulations, on the fishing gear used, and on other fishing parameters; even ice, wind, and currents may alter the catchability of the fish and the pattern of a fishery from year to year.
Fishery sampling, however, can be organized to represent the current fishery and sometimes the more general situation as well, such as adult fish abundance in a certain area. Here, sampling was planned to describe stock compositions in Finnish Atlantic salmon catches in the Gulf of Bothnia and the Gulf of Finland. Furthermore, the sampling site in the Åland Sea was expected to be representative of all stocks migrating northwards in spring, although some annual variation occurs in fishing effort and fish catchability. When fishing effort is known, it can be used to standardize catch data and make the proportion estimates better reflect changes in abundance. When fishing effort for Åland Sea catches was standardized for an abundance index, the overall results remained similar (Figure 5): the number of wild fish was clearly largest in 2004 and remained at a higher level in 2005, 6000–7000 individuals in the standardized catch, than in the previous years, 3000–5000 fish. This indicates not only a relative increase in the wild proportion but also an increase in the abundance of wild Atlantic salmon in the Baltic in 2004 and 2005.
|
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Fournier D.A., Beacham T.D., Riddell B.E., Busack C.A. (1984) Estimating stock composition in mixed stock fisheries using morphometric, meristic, and electrophoretic characters. Canadian Journal of Fisheries and Aquatic Sciences 41:400–408.
Gelman A. and Rubin D.B. (1992) Inference from iterative simulation using multiple sequences. Statistical Science 7:457–511.
ICES. (2005) Report of the Working Group on Baltic Salmon and Trout (WGBAST)5–14 April 2005Helsinki, Finland ICES CM 2005/ACFM: 18.
Koljonen M-L. (2001) Conservation goals and fisheries management units for Atlantic salmon in the Baltic Sea area. Journal of Fish Biology 59:Suppl. A, 269–288.[Web of Science]
Koljonen M-L., Jansson H., Paaver T., Vasin O., Koskiniemi J. (1999) Phylogeographic lineages and differentiation pattern of Atlantic salmon in the Baltic Sea with management implications. Canadian Journal of Fisheries and Aquatic Sciences 56:1766–1780.
Koljonen M-L. and Pella J.J. (1997) The advantage of using smolt age with allozymes for assessing wild stock contributions to Atlantic salmon catches in the Baltic Sea. ICES Journal of Marine Science 54:1015–1030.
Koljonen M-L., Pella J.J., Masuda M. (2005) Classical individual assignments versus mixture modelling to estimate stock proportions in Atlantic salmon (Salmo salar) catches from DNA microsatellite data. Canadian Journal of Fisheries and Aquatic Sciences 62:2143–2158.
McConnell S.K., O'Reilly P., Hamilton L., Wright J.N., Bentzen P. (1995) Polymorphic microsatellite loci from Atlantic salmon (Salmo salar): genetic differentiation of North American and European populations. Canadian Journal of Fisheries and Aquatic Sciences 52:1863–1872.
Michielsens C. G. J., Koljonen M-L., Mäntyniemi S. (2004) The use of genetic stock identification results for the assessment of wild Baltic salmon stocks. ICES CM 2004/EE: 03.
Michielsens C.G.J., McAllister M.K., Kuikka S.M., Pakarinen T., Karlsson L., Romakkaniemi A., Perä I., Mäntyniemi S. (2006) Bayesian state-space mark-recapture model to estimate fishing mortality rates within a mixed stock fishery. Canadian Journal of Fisheries and Aquatic Sciences 63:321–334.
Nei M. (1973) Analysis of gene diversity in sub-divided populations. Proceedings of the National Academy of Sciences of the United States of America 70:3321–3323.
Nei M., Tajima F., Tateno Y. (1983) Accuracy of estimated phylogenetic trees from molecular data. Journal of Molecular Evolution 19:153–170.[CrossRef][Web of Science][Medline]
O'Reilly P.T., Hamilton L.C., McConnell S.K., Wright J.M. (1996) Rapid analysis of genetic variation in Atlantic salmon (Salmo salar) by PCR multiplexing of dinucleotide and tetranucleotide microsatellites. Canadian Journal of Fisheries and Aquatic Sciences 53:2292–2298.
Ota T. (1993) DISPAN: genetic distance and phylogenetic analysis. Institute of Molecular Evolutionary Genetics, Pennsylvania State University, 328 Mueller Laboratory, University Park, PA 16802.
Pella J. and Masuda M. (2001) Bayesian method for analysis of stock mixtures from genetic characters. Fisheries Bulletin US 99:151–167.
Pella J.J. and Milner G.B. (1987) Use of genetic marks in stock composition analysis. In Ryman N. and Utter F. (Eds.). Population Genetics and Fisheries Management(University of Washington Press, Seattle) pp. 247–276.
Säisä M., Koljonen M-L., Gross R., Nilsson J., Tähtinen J., Koskiniemi J., Vasemagi A. (2005) Population genetic structure and postglacial colonization of Atlantic salmon in the Baltic Sea area based on microsatellite DNA variation. Canadian Journal of Fisheries and Aquatic Sciences 62:1887–1904.
Saitou N. and Nei M. (1987) The neighbour joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4:406–425.[Abstract]
Slettan A., Olsaker I., Lie O. (1995) Atlantic salmon, Salmo salar, microsatellites at the SSOSL25, SSOSL85, SSOSL311, SSOSL417 loci. Animal Genetics 26:277–285.[Web of Science][Medline]
Slettan A., Olsaker I., Lie O. (1996) Polymorphic Atlantic salmon (Salmo salar) microsatellites at the SSOSL438, SSOSL439, and SSOSL444 loci. Animal Genetics 27:57–58.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||