Wild Estonian and Russian sea trout (Salmo trutta) in Finnish coastal sea trout catches : results of genetic mixed-stock analysis

Wild Estonian and Russian sea trout ( Salmo trutta ) in Finnish coastal sea trout catches: results of genetic mixed-stock analysis MARJA-LIISA KOLJONEN 1 , RIHO GROSS 2 and JARMO KOSKINIEMI 3 1 Finnish Game and Fisheries Research Institute, Helsinki, Finland 2 Estonian University of Life Sciences, Institute of Veterinary Medicine and Animal Sciences, Department of Aquaculture, Tartu, Estonia 3 University of Helsinki, Department of Agricultural Sciences, Finland Koljonen, M.-L., Gross, R. and Koskiniemi, J. 2014 . Wild Estonian and Russian sea trout ( Salmo trutta ) in Finnish coastal sea trout catches: results of genetic mixed-stock analysis. – Hereditas 151 : 177 – 195. Lund, Sweden. eISSN 1601-5223. Received 1 September 2014. Accepted 4 November 2014. For responsible fi sheries management of threatened species, it is essential to know the composition of catches and the extent to which fi sheries exploit weak wild populations. The threatened Estonian, Finnish and Russian sea trout populations in the Gulf of Finland are targets of mixed-stock fi sheries. The fi sh may originate from rivers with varying production capacities, from different countries, and they may also have either a wild or hatchery origin. In order to resolve the composition of Finnish coastal sea trout catches, we created a standardized baseline dataset of 15 DNA microsatellite loci for 59 sea trout populations around the Gulf of Finland and tested its resolution for mixed-stock analysis of 1372 captured fi sh. The baseline dataset provided suffi cient resolution for reliable mixture analysis at regional group level, and also for most of the individual rivers stocks. The majority (76 – 80%) of the total catch originated from Finnish sea trout populations, 6 – 9% came from Russian and 12 – 15% from Estonian populations. Nearly all Finnish trout in the catch were of hatchery origin, while the Russian and Estonian trout were mostly of wild origin. The proportion of fi sh in the Finnish catches that originated from rivers with natural production was at least one fi fth (22%, 19 – 23%). Two different spotting patterns were observed among the captured trout, with a small and sparsely spotted form being markedly more common among individuals of Russian (28%) and Estonian origin (22%) than among fi sh assigned to a Finnish origin (0.7%). Marja-Liisa Koljonen, Finnish Game and Fisheries Research Institute, P.O. Box 2 FI-00791 Helsinki, Finland . E-mail: marja-liisa.koljo- [email protected] , marja-liisa.k[email protected] after 1 January 2015 Hereditas 151: 177–195 (2014) © 2015 The Authors. This is an Open Access article. DOI: 10.1111/hrd2.00070 There are currently about 101 rivers or brooks draining into the Gulf of Finland from Finland, Russia or Estonia in which there is an anadromous trout population ( Salmo trutta L . ). Of these populations, 85 can be regarded as native wild stocks (ICES 2013). The remaining populations have been supported by hatchery releases. For about one-third of the anadromous trout populations, the conservation status is very poor, as for 29 populations the current smolt production level is less than 5% of the potential smolt production level of the river. In addition, the conservation status is weak and uncertain for another 30 populations. In the Finnish Red Data Book, the anadromous trout is listed as Critically Endangered, because natural reproduction is unstable in most Finnish Baltic Sea populations due to intensive fi shing that also targets immature fi sh, migration obstructions and highly alternating water fl ow levels in rivers ( KAUKORANTA et al. 2000, KALLIO-NYBERG et al. 2001, HEINIMAA et al. 2007, URHO et al. 2010). In order to increase the escapement of sea trout from the fi shery, the legal minimum catch size of sea trout was increased at the beginning of 2014 from 50 cm to 60 cm. In addition, in 2013, the legal minimum catch size of sea trout was increased in the Finnish governmentally controlled offshore sea area of the Gulf of Finland from 50 cm to 65 cm for adipose fi n-clipped trout, all others are completely protected from fi shery and should be released if caught. In Russia, a declaration of trout preservation is in force so that no legal trout fi shing should occur. However, despite many of the rivers being situated in the border zone, poaching does occur. In the Estonian Red List of Threatened Species, the sea trout is listed as Near Threatened ( RED DATA BOOK OF ESTONIA 2008). In Finland, dam construction has been especially active and several sea trout stocks have been destroyed ( KALLIONYBERG et al. 2001). In order to compensate for the decreased population abundance and production levels, artifi cial reproduction in hatcheries and the release of reared fi sh or eggs into rivers are commonly practiced with the aim of re-establishing extinct or enhancing weak populations. In addition, hatchery releases to improve sea trout catches have also been widely used along the coastal area. In all, 293 000 sea trout smolts were released into the Gulf of Finland in 2012. The majority of these (74%, i.e. 216 000 smolts, mainly of Isojoki and Ingarskilanjoki origin) were from Finnish releases, 22% (64 000 smolts, only sea trout from the River Luga) came from Russia and the 178 M.-L. Koljonen et al. Hereditas 151 (2014) remaining 4% (13 000 smolts, mainly of Pudisoo origin) were from Estonian releases (ICES 2013). This also refl ects the approximate levels of recent years. Estonia has announced its aim to end sea trout releases. The mixed-stock fi sheries in the Gulf of Finland not only target fi sh of rivers with varying production capacities or separate national origins, but also a mixture of wild and hatchery fi sh. This in many respects resembles the situation with Atlantic salmon in the Baltic Sea ( KOLJONEN 2006, ICES 2013), although the trout fi shery is more restricted to coastal areas. Some mixing of anadromous trout from different countries is known to occur, as tagging experiments have generally shown about 5 – 10% of the trout tagged in Finland to be returned from the Estonian coast and some also from Russia. Correspondingly, sea trout tagged in Estonia have to some extent been recaptured in Finnish coastal waters. The coastal sea trout catch in 2012 was 13 300 kg in Estonia and 15 900 kg in Finland. In addition, Finland announced a total catch of 3800 kg from rivers (ICES 2013). In Russia, wild sea trout populations are found in at least 40 rivers or streams. The majority are situated along the northeastern coast of the Gulf of Finland, but the rivers with the most abundant smolt production are in the southern River Luga area (Fig. 1). The total smolt production of Russian rivers has been estimated to be at least 10 000 – 15 000 smolts. Smolt trap experiments indicate that between 2000 and 8000 sea trout smolts of natural origin annually migrate into the sea from the Luga, the largest Russian trout river ( TITOV and SENDEK 2008). In Russia, sea trout releases are only carried out into the River Luga and comprise fi sh of its own river system. In 2014, a total of 48 500 one-summer-old and 43 500 twosummer-old juveniles were released. In Estonia, sea trout populations are found in 45 rivers and brooks in the Gulf of Finland region, of which 38 have wild populations (ICES 2013). Rivers with higher 2 1 3 5 4 10 11 12 16 14 13 6 17 18 19 23 35 22 20 24 38-42 32 3334 47 48 49 50 51 52 53 54 56 57 58 59 55 43 44 45 46 Helsinki Tallin St. Petersburg Gulf of Finland 0 15 30 60 90 120 km FINLAND ESTONIA RUSSIA Bay of Vyborg Karelian Isthmus Vyborg 36 37 7 89 25 26 27 28 29 30 31 21 Archipelago Sea Catch 1 Catch 2 Catch 3 Fig. 1 . The sampled brown trout rivers in Finland, Russia and Estonia. The colour of the river indicates its quality as a spawning site and potential environment for brown trout. Red: river is closed; blue: irregular reproduction occurs; and green: open river with regular natural production of brown trout populations. The names and numbering of the rivers are the following: 1) Aurajoki, 2) Paimionjoki, 3) Purilanjoki, 4) Uskelanjoki, 5) Kiskonjoki, 6) Fiskarsinjoki, 7) Ingarskilanjoki, 8) Siuntionjoki, 9) Mankinjoki, 10) Espoonjoki, 11) Vantaanjoki, 12) Sipoonjoki, 13) Koskenkyl ä njoki, 14) Kymijoki, 15) Isojoki (hatchery stock, not in the map), 16) Summanjoki, 17) Virojoki, 18) Urpalanjoki, 19) Santajoki, 20) Vilajoki, 21) Tervajoki, 22) Rakkolanjoki, 23) Mustajoki, 24) Kilpeenjoki, 25) R ö mp ö tinpuro, 26) Myllyoja, 27) Koivistonpuro, 28) Penttil ä noja, 29) Kello-oja, 30) Lohijoki, 31) Papinoja, 32) Toivolanpuro, 33) Notkopuro, 34) Jukkolanpuro, 35) Inojoki, 36) Pikkuvammeljoki, 37) Vammeljoki, 38) Tyrisev ä noja, 39) Hurrinoja, 40) Terijoki, 41) Huumosenoja, 42) Kuokkalanpuro, 43) Rajajoki, 44) Voronka, 45) Sista, 46) Havlonka, 47) Luga, 48) P ü haj õ gi, 49) Kunda, 50) Toolse, 51) Selja, 52) Loobu, 53) Valgej õ gi, 54) Pudisoo, 55) Mustoja, 56) Pirita, 57) V ä ä na, 58) Keila, 59) Vasalemma. The borders of Finnish administrative ELY Centres on the coast are indicated by black lines. Approximate sea trout catch areas are shown as dark blue shaded areas along the Finnish coast. The blue arrows indicate the prevalent water currents in the Gulf of Finland. Hereditas 151 (2014) Genetic mixed stock analysis of sea trout catches 179 dataset for analysing the origin of Finnish sea trout catches in the coastal waters of the eastern Gulf of Finland. The rivers are numbered from west to east clockwise round the gulf so as the river numbers are showing the geographical proximity of the rivers (Fig. 1, Appendix 1 Table A1). Seventeen of the river systems were entirely on the Finnish side of the coast. Seven of the rivers cross the Russian border, so that the upper reaches of the rivers are located in Finland and the lower parts drain into the sea in Russia. In addition, 23 sea trout populations from relatively native rivers on the Russian coast and also from 12 rivers from the Estonian coast were included in the baseline data (Fig. 1). The Finnish sea trout population samples were obtained from rivers discharging into either the nearby western area, the Archipelago Sea or into the Gulf of Finland. In addition to the border rivers (the northern coast of the Bay of Vyborg), the Russian samples covered the southern coast of the Bay of Vyborg, the Karelian Isthmus, the northern coast of the Bay of St. Petersburg and part of the southern coast of the Gulf of Finland. Estonian rivers were located on the southwestern coast of the Gulf. The type and size of the rivers markedly vary among the studied countries and areas. In Finland, the watersheds are typically large and the rivers are often outlets of large lake and river systems. In Russia and Estonia, the rivers are often small and shallow, with the exception of the River Luga system. The status of the river environment in all countries is classifi ed into three categories, which are shown in Fig. 1. Open rivers are indicated in green, partly open in blue and river stretches above migration obstacles are shown in red. Preliminary information has also been used to classify the populations according to their level of originality as original (native), mixed through stocking or introduced, depending on their stocking history. The most interesting and valuable populations from the management point of view are those that are anadromous, viable, original and still genetically diverse. A substructure analysis within each Finnish river system, in cases where several samples were taken, was performed in K OLJONEN et al. (2013), and only anadromous populations were included in the present study. All 12 Estonian baseline populations and most, 16 out of 23, Russian samples were new to this study. The Finnish coast is divided into three governmental administrative sectors according to the Centres for Economic Development, Transport and the Environment (ELY Centres). These Centres are responsible for the regional implementation and development tasks of the government. The rivers are listed separately for each ELY Centre (Appendix 1 Table A1; borders are also shown in Fig. 1). smolt production are situated in the central part of the north Estonian coast (Fig. 1). Watershed-based analyses of the genetic structure of Finnish brown trout populations have also previously been conducted for national purposes with both allozymes and DNA microsatellites ( KOLJONEN 1989, M ARTTINEN and KOLJONEN 1989, KOLJONEN et al. 1992, K OLJONEN and SAURA 1992, NUOTIO and KOSKINIEMI 1995, SAURA 2005, AALTONEN 2009, 2011). In addition, research teams from other Baltic Sea countries have investigated some restricted areas or river systems of the Baltic Sea drainage basin by using microsatellite markers ( WAS and WENNE 2002, HANSEN et al. 2006, L EHTONEN et al. 2009, SAMUILOVIENE et al. 2009). However, this study is the fi rst to cover the whole Gulf of Finland coast. For responsible fi sheries management of threatened species, it is essential to know the composition of catches and the extent to which fi sheries exploit weak wild production. Mixed-stock analysis based on genetic data has been used for decades in analysing the composition of Pacifi c salmon catches in North America ( SHAKLEE et al. 1999, BEACHAM et al. 2001, 2008) and the Atlantic salmon fi shery in West Greenland ( REDDIN and FRIEDLAND 1999). It has also been used to reveal the composition of Atlantic salmon catches in the Baltic Sea since 2000 ( KOLJONEN 2006, ICES 2013). A genetic baseline database for mixedstock analysis has additionally been created for Atlantic salmon for the whole European range ( GRIFFITHS et al. 2010). Furthermore, catch composition analysis has been utilized to analyse stock proportions in brown trout catches in the large, northern, Finnish Lake Inari ( SWATDIPONG et al. 2013). The aim of this study was to test the possibilities of using microsatellite data for mixed-stock analysis of sea trout catches in the Gulf of Finland. The goals were to: 1) collect international baseline data from all sea trout stocks around the Gulf of Finland; 2) test the resolution of this baseline data set with simulations to analyse stock and stock group proportions, and also the usefulness for individual assignments of captured fi sh; 3) analyse the national contributions of trout production in the studied Finnish catches; 4) analyse the proportions of wild and hatcheryproduced fi sh in the sea trout catch; and 5) compare the results obtained with Bayesian and maximum likelihood estimation methods MATERIAL AND METHODS Fish sampling from rivers to create the baseline dataset Altogether, 4224 sea trout individuals from 59 watersheds were sampled in the current study to create the baseline 180 M.-L. Koljonen et al. Hereditas 151 (2014) volume with 3 μ l of extracted DNA, 5 μ l of kit master mix and primers with concentrations and dyes as presented in the Appendix 1 Table A2. PCR reactions were carried out in PTC200 Thermal Cyclers (MJ Research), and the temperature profi le of the PCR program was suggested in the Type-it Microsatellite kit manual. The annealing temperature was 56 ° C. The amplifi cation products were separated by capillary electrophoresis on AB3130 (in Finland) and AB3500 (in Estonia) Genetic Analyzers (Applied Biosystems, Foster City, CA), and the sizes of the microsatellite alleles were determined using Genemapper ver. 4.0 and ver. 4.1 software, respectively (Applied Biosystems, Foster City, CA), and manually checked. Cross-laboratory calibration and standardisation of allele sizes for all loci was conducted between laboratories from the University of Helsinki, Department of Agricultural Sciences and the Estonian University of Life Sciences, Institute of Veterinary Medicine and Animal Sciences, Department of Aquaculture. This allowed the pooling of genotype data to establish a joint baseline dataset, and enables the future analysis of all catch samples in both laboratories. Statistical analysis Pairwise F st values were calculated with FSTAT ver. 2.9.3.2. (February 2002) ( GOUDET 1995, 2001) (  www2. unil.ch/popgen/softwares/fstat.htm  ). Analysis of the differences between populations was based on genotype frequencies and was also tested with FSTAT, which includes Bonferroni correction for multiple tests. Genetic distances between sea trout populations were calculated using Nei ’ s D A distances ( NEI et al. 1983). Phylogenetic trees were constructed using a neighbour-joining (NJ) algorithm ( SAITOU and NEI 1987, TAKEZAKI 1998) with Populations 1.2.32 software ( LANGELLA 1999) (  http:// bioinformatics.org/ ∼ tryphon/populations/  ). Bootstrapping with 1000 replicates was used to test the statistical strength of the branches. The genetic distance tree was drawn with TreeView ver. 1.6.6. ( PAGE 1996) (  http:// taxonomy.zoology.gla.ac.uk/rod/treeview.html  ). The catch composition was analysed with the Bayesian estimation method for mixed-stock analysis using the BAYES program ( PELLA and MASUDA 2001). Genetic information on 59 sea trout populations around the Gulf of Finland and Archipelago Sea from Finland, Russia and Estonia was available as baseline population data. Ten iteration chains were calculated for each catch estimate. The prior of each potential baseline population had a maximum proportion of 71.0% in turn, and the others evenly 0.005%. The convergence of the chains to the posterior probability distribution was tested for each catch sample with convergence statistics ( GELMAN and RUBIN 1992), included in the BAYES software package. The last 1000 Sea trout sampling from fi shery catches For the catch analysis, a total of 1373 trout were caught by fi shermen, mostly using bottom gill nets (70%), but also by rod (17%) or fyke net (13%). Samples were collected from both professional and recreational fi shery. The catch sites were divided into three sectors according to the nearest coastal municipality (Table 1, Fig. 1). These catch sectors are not the same as the ELY Centre borders. All catch samples mostly came from the eastern part of the Gulf of Finland, between the city of Helsinki and the Russian border, and relatively close to the Finnish coast. No catch samples were available from the westernmost part of the coast from the area of the ELY Centre for southwest Finland. Fishery sampling was more effi cient in the easternmost area (Table 1). The samples were collected over the years from 1996 to 2012 and they do not thus represent any particular time period, but rather provide an overview of the situation in recent years, when hatchery population releases have been quite intensive in Finland. Most samples were collected between years 2006 – 2012, and only samples less than 50 individuals in previous years. (Table 1). Fishermen also visually inspected fi sh in their catch and reported adipose fi n-clipped fi sh, and also fi sh that somehow deviated in their appearance. Some fi sh were reported to look more like salmon, being slender and having fewer and smaller spots than fi shermen were used to have seen. These fi sh were identifi ed from the individual assignment database and also analysed separately. DNA methods Total genomic DNA was extracted from scale or tissue samples in 95% alcohol using the DNeasy Blood & Tissue Kit (Qiagen) or according to the simplifi ed method of L AIRD et al. (1991). Variation was determined at 16 microsatellite loci (Appendix 1 Table A2). However, locus SSa289 was omitted from the statistical analysis, as it was not included in the Estonian baseline data. For each sample, two multiplex PCR reactions were performed using the Qiagen Type-it Microsatellite kit in a 10μ l reaction Table 1. Numbers, sampling sites and main sampling years (  50 fi sh year 1 ) of the analysed Finnish sea trout catch samples. Area Municipality Location n Main years 1 Virolahti, Hamina, Kotka East 656 2006 – 2012 2 Pyht ä ä , Ruotsinpyht ä ä , Loviisa, Pernaja Middle 492 2009 – 2012 3 Porvoo, Sipoo, Helsinki, Espoo, Kirkkonummi West 224 2010 – 2011 Total 1372 Hereditas 151 (2014) Genetic mixed stock analysis of sea trout catches 181 10, with the corresponding mixture sample sizes of 295, 413 and 590. RESULTS Genetic differentiation and distances between baseline populations The level of differentiation (as estimated by F st ) between pairs of baseline populations varied from 0.002 to 0.437. Most of the baseline populations were signifi cantly differentiated from each other, and statistically non-signifi - cant differences in genotype frequencies (after Bonferroni correction for multiple tests) were only observed for two Finnish population pairs and one Estonian population pair (F st values for these pairs are presented in red in Table 2): Aurajoki – Fiskarsinjoki (F st 0.009), Ingarskilanjoki – Koskenkyl ä njoki (F st 0.013) and Pudisoo – P ü haj õ gi (F st 0.002). The genotype frequencies of these population pairs were thus identical to each other, and fi sh transfers or releases are known to have occurred in all of them. The difference was also only signifi cant at the 5% nominal level for two very similar population pairs of rivers, Paimionjoki (2) – Siuntionjoki (8) in Finland and Hurrinoja (40) – Kuokkalanpuro (42) (Fig. 1) on the Karelian Isthmus in Russia. The latter pair of rivers should have contained native populations, but are geographically very close to each other (Fig. 1). All rivers numbered from 38 to 42 almost discharge into the same location. In addition to the already mentioned very similar population pairs, F st values smaller than 0.02 were observed between thirteen Estonian population pairs, and also between occasional population pairs in some other geographically close rivers from other areas (all shown shaded in Table 2). All F st values less than 0.02 were, however, observed between population pairs of the same geographical area. Separating these very similar populations on the basis of genetic data could be assumed to be diffi cult, and these populations are unlikely to be distinguished as individual populations in mixed-stock analysis, and were thus of special interest in the simulation analysis. In most of the cases with a very high similarity between populations, genetic effects of hatchery releases of known sources could be assumed according to the history of hatchery fi sh releases. Among the Finnish populations, Aurajoki trout have been released into the River Fiskarsinjoki and Koskenkyl ä njoki has been restocked using Ingarskilanjoki trout, which explains their identity, while Isojoki trout have also regularly been released into the Rivers Kymijoki and Summanjoki, which can be seen in the high similarity between them. In Estonia, Pudisoo and P ü haj õ gi trout have at least also been mixed, in addition to some others, but the geographical proximity there is also high. In Russia, mixing of populations is not known to MCMC iterations of each 6000-iteration chain were combined and used to describe the posterior probability distributions of the proportions of each baseline population, and the eight population groups (formed based on the genetic distances among baseline populations), as well as for the individual assignments. The Bayesian estimation was chosen for the estimation of the fi nal results, as it has been shown in previous studies to out-perform the maximum likelihood-based methods ONCOR and SPAM, e.g. in tests where mixed samples of known origin were available ( GRIFFITHS et al. 2010, MORAN et al. 2014). For comparison, the maximum likelihood estimates of the same catches were assessed with ONCOR software (  www.montana.edu/kalinowski/Software/ONCOR. htm  ) ( KALINOWSKI et al. 2007, ANDERSON et al. 2008). The analysed sea trout baseline dataset for the Gulf of Finland was very large, including several similar sets of populations. It consisted of nearly 60 sea trout populations, and they originated from a geographically relatively small area for some locations, which sets challenges for the estimation method. For example, when analysing a baseline data set of 57 Atlantic salmon river populations in the southern part of the European range, clear diffi culties were met regarding the estimation accuracy because of the genetic similarity of the populations, but possibly also because of the small baseline sample sizes ( GRIFFITHS et al. 2010). Thus, the capacity of the estimation method in general to separate so many individual trout populations and groups was fi rst assessed by the simulation options included in the ONCOR software. The population proportion estimation of individual river populations was evaluated with so-called 100% simulation, in which each baseline population in turn contributed 100% of the analysed mixture. The baseline samples were the same as in the empirical baseline. The mixture sample size was set to 200 and 1300, which were approximately the smallest and largest catch sizes used here (224, 492, 656 and 1372, Table 1), and simulations were repeated 100 times for both individual populations and the eight population groups used here as reporting groups. The accuracy of individual assignment was tested with the leave-one-out procedure and results were obtained similarly for individual populations and reporting groups. In addition, the accuracy of the estimation was assessed by a simulation option called ‘ Realistic fi shery simulation ’ , with simulated mixtures of 200 individuals so that each of the baseline populations in a set of 20 populations contributed 5% to the simulated catch. The results of these simulations are combined into one table. In addition, the option ‘ Three way error decomposition: Fishery, Loci, Baseline ’ , was used to analyse the source of observed uncertainty for each baseline population by setting the number of individuals in each population to be 5, 7 and 182 M.-L. Koljonen et al. Hereditas 151 (2014) Table 2. Genetic differentiation of the most similar baseline populations when measured as pairwise F st values. Values less than 0.01 were statistically nonsignifi cant and are highlighted in red, and all observed values less than 0.02 are highlighted in orange. Only columns and rows with values less than 0.02 are shown . Pop 1 Aurajoki 7 Ingarskilanjoki 14 Kymijoki 31 Papinoja 33 Notkopuro 35 Inojoki 39 Hurrinoja 44 Voronka 48 Pühaj õ gi 49 Kunda 51 Selja 52 Loobu 56 Pirita 57 V ä ä na 58 Keila 6 Fiskarsinjoki 0.009 FIN 13 Koskenkyl ä njoki 0.150 0.013 RUS RUS 15 ISOJOKI 0.046 0.116 0.010 32 Toivolanpuro 0.107 0.171 0.138 0.020 35 Inojoki 0.068 0.152 0.101 0.039 0.015 37 Vammeljoki 0.086 0.153 0.099 0.061 0.032 0.017 39 Hurrinoja 0.070 0.147 0.105 0.029 0.019 0.014 41 Huumosenoja 0.070 0.160 0.112 0.063 0.032 0.020 0.016 42 Kuokkalanpuro 0.094 0.165 0.119 0.029 0.027 0.024 0.012 45 Sistajoki 0.065 0.125 0.090 0.074 0.041 0.042 0.046 0.017 46 Havlonka 0.057 0.145 0.093 0.077 0.046 0.036 0.045 0.020 51 Selja 0.047 0.125 0.086 0.075 0.053 0.045 0.045 0.049 0.025 0.017 52 Loobu 0.071 0.142 0.107 0.082 0.057 0.058 0.064 0.067 0.034 0.011 0.020 EST 54 Pudisoo 0.056 0.153 0.097 0.085 0.067 0.058 0.061 0.071 0.002 0.032 0.028 0.033 56 Pirita 0.064 0.128 0.096 0.086 0.055 0.052 0.051 0.054 0.041 0.032 0.019 0.036 57 V ä ä na 0.051 0.125 0.096 0.073 0.052 0.045 0.042 0.052 0.025 0.024 0.015 0.027 0.031 58 Keila 0.057 0.133 0.098 0.064 0.041 0.035 0.042 0.047 0.024 0.014 0.012 0.017 0.028 0.015 59 Vasalemma 0.045 0.124 0.081 0.058 0.043 0.034 0.036 0.040 0.023 0.013 0.011 0.017 0.018 0.014 0.010 Hereditas 151 (2014) Genetic mixed stock analysis of sea trout catches 183 the relatively native Mankinjoki, Espoonjoki and Sipoonjoki populations from the nearby area could at least partly be assumed to be a result of their common evolutionary history. The four westernmost river populations formed the Uskelanjoki group, and these rivers mostly drain into the Archipelago Sea. The Russian populations grouped very precisely according to their geographical distances. The Bay of Vyborg populations formed a tight group, and within it even the subgroups from the northern and southern coast of the bay could be separated, as two populations, Myllyoja (number 25) and R ö mp ö tinpuro (26) from the southern coast, formed a small subgroup of its own (Fig. 2). All the Russian populations from the Karelian Isthmus grouped together, but with some distinction from populahave occurred, although some River Luga trout have been released back into the Luga itself. When a phylogenetic tree was constructed based on the genetic distances between all populations, a clear and logical grouping could be seen in the baseline data, which followed surprisingly well the geographical distances between the populations and the form of the coastline (Fig. 2), despite the long history of releasing hatchery populations into the area. The exceptions to the geographical order along the coastline came from the already known hatchery releases of Isojoki and Aurajoki hatchery populations. In addition, the release history explains the similarity of the Ingarskilanjoki trout with the Koskenkyl ä njoki and Vantaanjoki populations, into which it has been released. The similarity of the Ingarskilanjoki and 27Koivistonpuro 29Kello-oja 28Penttilänjoki 31Papinoja 32Toivolanpuro 30Lohijoki 33Notkopuro 34Jukkolanpuro 38Tyrisevänoja 40Terijoki 39Hurrinoja 41Huumosenoja 42Kuokkalanpuro 35Inojoki 36Pikkuvammeljoki 37Vammeljoki 19Santajoki 22Rakkolanjoki 21Tervajoki 23Mustajoki 24Kilpeenjoki 20Vilajoki 17Virojoki (FIN) 25Römpötinpuro 26Myllyoja 18Urpalanjoki 43Rajajoki 44Voronka 45Sista 46Havlonka 47Luga 49Kunda 52Loobu 50Toolse 51Selja 53Valgejõgi 57Vääna 58Keila 59Vasalemma 48Pühajõgi 54Pudisoo 55Mustoja 56Pirita 7Ingarskilanjoki 13Koskenkylänjoki 9Mankinjokinjoki 11Vantaanjoki 10Espoonjoki 12Sipoonjoki 14Kymijoki 15Isojoki 16Summanjoki 1Aurajoki 6Fiskarsinjoki 5KiskoPerniönjoki 2Paimionjoki 8Siuntionjoki 3Purilanjoki 4Uskelanjoki Estonian stocks Finnish stocks Finnish-Russian border rivers Russia Karelian Isthmus Russia South G3 G8 G7 G5 G4 G2 G1 G6 18 33 46 99 10 100 55 50 15 5 94 709 48 24 10 99 6 40 10 63 57 27 7 85 31 20 24 72 19 11 9 11 22 16 30 18 19 42 1 67 52 18 36 13 59 42 75 30 23 13 9 13 43 5 44 3 Fig. 2 . Genetic distances (Nei ’ s D A distance, neighbour-joining unrooted tree) among Finnish, Russian and Estonian anadromous brown trout stocks in the Gulf of Finland. Watersheds are numbered from west to east clockwise around the gulf and watershed numbers are presented together with the river names. Bootstrap numbers for the branches are shown as a percentage from 1000 repeats. 184 M.-L. Koljonen et al. Hereditas 151 (2014) Estonian coast, a slight subgrouping could additionally be seen, and populations from the Rivers Keila and V ä ä na were also very similar. In all, fi ve major groups could be formed based on their genetic distances: Finnish populations (groups 1 – 4), border river populations (group 5), eastern Russian populations (group 6), southern Russian populations (group 7) and Estonian populations (group 8) (Fig. 2, 3). Deviating from this, the Finnish River Virojoki, located next to the border rivers, grouped with them according to its genetic similarity. Consequently, the Bay of Vyborg group actually includes border rivers and the River Virojoki from Finland near the border. The Finnish populations, in turn, could be further divided into four groups: Archipelago Sea area populations (1), Aurajoki hatchery type populations (2), Ingarskilanjoki type populations (3) and Isojoki hatchery type populations (4). As most of the captured fi sh originated from Finnish populations, fi ner-scale grouping was necessary in order to gain more detailed information on the catch composition. These eight population groups were used as reporting groups in the catch analysis (Fig. 2), and this served as the fi rst output level of the origin information (Table 3). The locations of the rivers in the reporting groups are indicated in Fig. 3. Four of the groups (G1 to G4) are tions north and south of the Inojoki (35). The Inojoki population belonged to the southern group, which mainly includes rivers draining into the Bay of St. Petersburg. The Russian Rivers Papinoja (31) – Toivolanpuro (32) and Notkopuro (33) – Jukkolanpuro (34) formed geographically close pairs. In addition, a very tight group of fi ve small rivers (38 – 42) was also found on the northern coast of the Bay of St. Petersburg, all of which were diffi cult to distinguish from each other. The Rivers Hurrinoja (39) – Huumosenoja (41) and also Tyrisev ä noja (38) – Terijoki (40) formed close pairs, with Kuokkalanpuro (42) belonging to the same group with all of them (Fig. 1, 2). The River Rajajoki (River Siestarjoki) formed an intermediary type to the southern coastal Russian group, in which the River Luga has the largest smolt production. The four rivers on the southern coast drain into two small gulfs, two into each, and their genetic similarity also refl ected this pattern. The Rivers Voronka (44) and Sista (45) formed a pair, as well as the Rivers Havlonka (47) and Luga (48). Interestingly, the Estonian trout populations clearly differed from this Luga type of trout, and were very similar to each other. Some effects of hatchery fi sh releases could be seen, as Pudisoo and P ü haj õ gi were genetically very similar, and Pudisoo trout is known to have been released into the P ü haj õ gi river. Along the 2 1 3 5 4 10 11 12 16 14 13 6 17 18 19 23 35 22 20 24 38-42 32 33 34 47 48 49 50 51 52 53 54 56 57 58 59 55 43 44 45 46 Helsinki Tallin St. Petersburg Gulf of Finland 0 15 30 60 90 120km FINLAND ESTONIA RUSSIA Bay of Vyborg Karelian Isthmus Vyborg 36 37 7 8 9 25 26 27 28 29 30 31 21 Archipelago Sea G1 G2 G3 G4 G5 G6 G7 G8 Fig. 3 . The location of the eight genetic catch analysis reporting groups of the Finnish, Russian and Estonian sea trout stocks in the Gulf of Finland. The river numbering is the same as in Fig 1. Administrative ELY Centre borders on the Finnish coast are indicated by black lines. Hereditas 151 (2014) Genetic mixed stock analysis of sea trout catches 185 population group proportion estimates. When the accuracy of individual assignment was assessed with the leaveone-out procedure, the likelihood of correct assignment was over 90% for only 7 out of the 59 baseline populations. For 20 river populations it was over 80%, while for 26 populations the likelihood of correct assignment was still over 75%. For four Estonian populations (Keila, Kunda, V ä ä na and Pudisoo) it was less than 50%, again indicating a very high degree of similarity among some of them. In particular, the Pudisoo population had a very low likelihood of being correctly assigned, and it was more often assigned to P ü haj õ gi than to its own river, which indicates that it possibly does not even deserve the status of a separate population. However, when the likelihood of correct assignment of the individuals to their own reporting group was estimated, on average up to 93% of the individuals were assigned to the correct population group. For as many as 45 river populations out of 59 (76%), the percentage of correct assignments was over 90%. All individual assignments at the group level were correct, at least at the 77% level. The highest uncertainty occurred between some Finnish hatchery populations. When simulated mixtures with the same proportions of each population were analysed, variation in estimation bias could be compared. Biases with a nominal value of over 1% for a 5% proportion occurred for 16 out of 59 baseline populations (Table 5). The uncertainty was highest among the Estonian populations. In Finnish populations, the proportion of Koskenkyl ä njoki trout was underestimated in favour of the Ingarskilanjoki trout, from which it originates. On the Karelian Isthmus, the sample from the river Inojoki (baseline sample size n  98) population seemed to represent well the trout of the area, as its proportion was overestimated at the cost of some of the completely on the Finnish coast, one group (G5) crosses the Finnish – Russian border, two groups are located in Russia (G6 and G7) and one group in Estonia (G8). Resolution of the baseline dataset When the so-called 100% simulation of each baseline population was conducted, the results showed that a likelihood of correct proportion estimation below 90% occurred in the most similar and hatchery rearing-infl uenced Finnish populations (Aurajoki, Fiskarsinjoki, Koskenkyl ä njoki, Kymijoki), for which the stocking history was known (Table 4). In addition, low values were also obtained for four native Russian populations on the Karelian Isthmus, belonging to the tight population group Vammeljoki (37), Tyrisev ä noja (38), Hurrinoja (39), Huumosenoja (41), Kuokkalanpuro (42), Voronka (44, mixing with Sista, 45) and Havlonka (46, mixing with 47 Luga) on the Russian southern coast, and several Estonian populations (Valgej õ gi, Pudisoo, Mustoja, Pirita, V ä ä na, Keila and Vasalemma). The value for the Pudisoo population was especially poor, being only 0.09. Clear uncertainty is thus evident in the individual population estimates for some of these populations, and at least the most similar populations should be grouped in all potential grouping options to obtain reliable estimates, especially for small catch sample analysis. However, for all individual river populations, the likelihood of being correctly assigned to its reporting group was at least 0.97 (Table 4.). If the mixture sample size was increased to the maximum used for the total catch of about 1300 fi sh, it did not essentially affect the result. There was considerably more uncertainty in the individual assignments of the fi sh than for the population or Table 3. Sea trout river populations included in the reporting groups of the catch analysis. The mean F st value over stock pairs within each group is also presented . Group River populations Country Mean F st 1 Aurajoki Aurajoki, Kiskonjoki, Fiskarsinjoki FIN 0.079 2 Uskelanjoki Paimionjoki, Purilanjoki, Uskelanjoki, Siuntionjoki FIN 0.239 3 Ingarskilanjoki Ingarskilanjoki, Mankinjoki, Espoonjoki, Vantaanjoki, Sipoonjoki, Koskenkyl ä njoki FIN 0.078 4 Isojoki Kymijoki, Isojoki, Summanjoki FIN 0.029 5 Bay of Vyborg Virojoki (FIN), Urpalanjoki, Santajoki, Vilajoki, Tervajoki, Rakkolanjoki, Mustajoki, Kilpeenjoki, R ö mp ö tinpuro, Myllyoja FIN/RUS 0.066 6 Karelian Isthmus Koivistonpuro, Penttil ä njoki, Kello-oja, Lohijoki, Papinoja, Toivolanpuro, Notkopuro, Jukkolanpuro, Inojoki, Pikkuvammeljoki, Vammeljoki, Tyrisev ä noja, Hurrinoja, Terijoki, Hurrinoja, Kuokkalanpuro, Rajajoki RUS 0.055 7 Russia south coast Voronka, Sista, Havlonka, Luga RUS 0.032 8 Estonia P ü haj õ gi, Pudisoo, Mustoja, Pirita, Kunda, Toolse, Selja, Loobu, Valgej õ gi, V ä ä na, Keila, Vasalemma EST 0.028 192 M.-L. Koljonen et al. Hereditas 151 (2014) Aaltonen, J. 2011. Kiskonjoen – Perni ö njoen vesist ö n s ä hk ö koekalastukset vuosina 2007 ja 2009 sek ä taimenkannan DNA-analyysi. – Salon seudun kalastusalue. Anderson, E. C., Waples, R. S. and Kalinowski, S. T. 2008. An improved method for estimating the accuracy of genetic stock identifi cation. – Can. J. Fish. Aquat. Sci. 65:1475 – 1486. Aparicio, E., Garcia-Berthou, E., Araguas, R. M. et al. 2005. Body pigmentation pattern to assess introgression by hatchery stocks in native Salmo trutta from Mediterranean streams. – J. Fish Biol. 67: 931 – 949. Beacham, T. D., Candy, J. R., Supernault, K. J. et al. 2001. Evaluation and application of microsatellite and major histocompatibility complex variation for stock identifi cation. – Trans. Am. Fish. Soc. 130: 1116 – 1149. Beacham, T. D., Spilsted, B., Le, K. D. et al. 2008. Population structure and stock identifi cation of chum salmon Oncorhynchus keta from British Columbia determined with microsatellite DNA variation. – Can. J. Zool. 86: 1002 – 1014. Degerman, E., Leonardsson, K. and Lundqvist, H. 2012. Coastal migrations, temporary use of neighbouring rivers, and growth of Sea trout ( Salmo trutta ) from nine northern Baltic Sea rivers – ICES J. Mar. Sci. 69: 971 – 980. Gelman, A. and Rubin, D. B. 1992. Inference from iterative simulation using multiple sequences. – Stat. Sci. 7 : 457 – 511. Goudet, J. 1995. FSTAT (ver. 1.2): a computer program to calculate F-statistics. – J. Hered. 86: 485 – 486. Goudet, J. 2001. FSTAT, a program to estimate and test gene diversities and fi xation indices (ver. 2.9.3). Griffi ths, A. M., Machado-Schiaffi no, G., Dillane, E. et al. 2010. Genetic stock identifi cation of Atlantic salmon ( Salmo salar ) populations in the southern part of the European range. – BMC Genet. 2010: 11 – 31. Hansen, M. M., Bekkevold, D., Jensen, L. F. et al. 2006. Genetic restoration of a stocked brown trout Salmo trutta population using microsatellite DNA analysis of historical and contemporary samples. – J. Appl. Ecol. 43: 669 – 679. Heinimaa, P., Jutila, E. and Pakarinen, T. (eds.) 2007. It ä meren meritaimenty ö paja, (Baltic Sea Trout Workshop). – Riistaja kalatalouden tutkimuslaitos, Kala ja riistaraportteja nro 410. ICES 2013. Report of the Baltic Salmon and Trout Assessment Working Group 2013 (WGBAST), 3 – 12 April, Tallinn, Estonia. – ICES CM 2013/ACOM:08. Kalinowski, S. T., Manlove, K. R. and Taper, M. L. 2007. A computer program for genetic stock identifi cation. Manual. – Dept of Ecology, 310 Lewis Hall, Montana State Univ.  www.montana.edu/kalinowski/ONCOR/ONCOR_ Manual_21Oct2007.pdf  Kallio-Nyberg, I., Koljonen, M.-L. and Jutila, E. 2001. Taimenatlas. – Kalatutkimuksia 173. Kallio-Nyberg, I., Jutila, E., Koljonen, M.-L. et al. 2010. Can the lost migratory Salmo trutta stocks be compensated with resident trout stocks in the coastal rivers? – Fish. Res. 102: 69 – 79. Kaukoranta, M., Koljonen, M.-L., Koskiniemi, J. et al. 2000. Atlas of Finnish fi shes, English summary. Distribution of lamprey, brook lamprey, Atlantic salmon, brown trout, Arctic charr, whitefi sh, vendace, grayling, asp, vimba, spined loach and bullhead – the distribution and status of stocks. – Finnish Game and Fisheries Res. Inst. Kocaba s¸ , M. and Ba s¸ ç inar, N. 2013. The effect of salinity on spotting features of Salmo trutta abanticus, S. trutta fario such as Vantaanjoki, Koskenkyl ä njoki and Ingarskilanjoki itself. The Isojoki trout originates from a river that does not drain into the Gulf of Finland at all, but further to the north, into the Gulf of Bothnia. The use of this population for trout releases in the Gulf of Finland could be reconsidered in future, as it may well cause some gene fl ow to the still native rivers. A hatchery broodstock has already founded from the border-river, Mustajoki, population for potential use in releases in the eastern Gulf of Finland ( PEUHKURI et al. 2014). The Russian populations were slightly more common in the eastern catches. The baseline dataset collected in this study is highly valuable and will enable future catch analyses in all areas of the Gulf of Finland with reasonable costs. It thus offers a useful tool for fi sheries management and conservation of the still native sea trout populations of the Gulf of Finland area. Acknowledgements – This work was mainly fi nanced by the INTERREG IV A Programme project HEALFISH ‘ Healthy fi sh stocks – indicators of successful river basin management ’ (2010 – 2013). A proportion of the Russian and Finnish baseline samples were collected within the INTERREG projects ISKALT and ISKALT II ( SAULAMO et al. 2007, see sampling years in Table 1). About half of the catch sample collection and analysis was fi nanced by the Finnish – Russian project ‘ Rivers and fi sh – our common interest ’ (RIFCI 2011 – 2014) funded by the Southeast Finland – Russia ENPI CBC 2007 – 2013 Programme. The earlier analysed data on Russian populations, collected during the ISKALT 2003 – 2007 projects, was updated for 16 DNA microsatellite loci from the previously analysed 10 loci data sets, and this updating was fi nanced by the RIFCI project. Additional fi nancing was provided by the Estonian Ministry of Education and Research (institutional research funding project IUT8-2). We warmly thank all our colleagues from the Centre for Economic Development, Transport and the Environment of Southwest Finland, Uusimaa and Southeast Finland, GosNIORKh, the Finnish Game and Fisheries Research Institute and the Estonian Marine Institute of the University of Tartu for organizing and realizing both river population and catch sampling. We are also grateful to all the numerous fi shermen who sampled their catch fi sh for the project. Frontier guards made it possible to enter the frontier zone for sampling of the Finnish – Russian border rivers. Ari Saura and Aki Janatuinen provided valuable practical information on the trout rivers and their state in the Gulf of Finland during the project. Pekka V ä h ä n ä kki had a signifi cant role in highlighting the importance of this project for the benefi t of common sea trout populations and fi shermen in all three countries. We also want to express our sincere thanks to all the parties that have provided funding for our project. Dr. Roy Siddall checked the English language. REFERENCES Aaltonen, J. 2009. Electro-fi shing in the Uskelanjoki watercourse in year 2006 and genetic analysis of the brown trout ( Salmo trutta L.) stock. – Turku Univ. of Applied Sciences, Fisheries and Environmental care, in Finnish. Hereditas 151 (2014) Genetic mixed stock analysis of sea trout catches 193 Reddin, D. G. and Friedland, K. D. 1999. A history of identifi cation to continent of origin of Atlantic salmon ( Salmo salar L.) at west Greenland, 1969 – 1997. – Fish. Res. 43: 221 – 235. Saitou, N. and Nei, M. 1987. The neighbour joining method: a new method for reconstructing phylogenetic trees. – Mol. Biol. Evol. 4: 406 – 425. Saulamo, K., V ä h ä n ä kki, P. and Peuhkuri N. 2007. It ä isen Suomenlahden kalaston selvitys ja sen seuranta mahdollisten ö ljyja kemikaalionnettomuuksien varalta. – ISKALT IIhankkeen loppuraportti. Unpublished project report. Samuiloviene, A., Kontautas, A. and Gross, R. 2009. Genetic diversity and differentiation of sea trout ( Salmo trutta ) populations in Lithuanian rivers assessed by microsatellite DNA variation. – Fish Physiol. Biochem. 35: 649 – 659. Saura, A. 2005. Taimen Karjaanjoen vesist ö alueella. teoksessa: Karjaanjoen vesist ö – El ä k ö ö n vesi. – Lohjan ymp ä rist ö lautakunta. julkaisu 7/05, pp. 92 – 99. Shaklee, J. B., Beacham, T. D., Seeb, L. et al. 1999. Managing fi sheries using genetic data: case studies from four species of Pacifi c salmon. – Fish. Res. 43: 45 – 78. Segerstr å le, C. 1937. Studier r ö rande havsforellen ( Salmo trutta L.) i S ö dra Finland, speciellt p å Karelska n ä set och I Nyland. – Acta Soc. Fauna Flora Fenn. 60: 696 – 750. Skaala, Ø . and J ø rstad, K. E. 1988. Inheritance of the fi ne-spotted pigmentation pattern of brown trout. – Pol. Arch. Hydrobiol. 35: 295 – 304. Skaala, Ø . and J ø rstad, K. E. 2011. Fine-spotted brown trout ( Salmo trutta ): its phenotypic description and biochemical genetic variation. – Can. J. Fish. Aquat. Sci. 44: 1775 – 1780. Swatdipong, A., Vasem ä gi, A., Niva, T. et al. 2013. Genetic mixed-stock analysis of lake-run brown trout Salmo trutta fi shery catches in the Inari Basin, northern Finland: implications for conservation and management. – J. Fish Biol. 83: 598 – 617. Takezaki, N. 1998. NJBAFD: Neighbor-joining tree construction from allele frequency data. – Natl Inst. Genetics, Misima, Sizuoka-ken, Japan.  http://homes.bio.psu.edu/people/ Faculty/Nei/Lab/software.htm  . Titov, S. and Sendek, D. 2008. Atlantic salmon in the Russian part of the Baltic Sea basin. – Baltic Fund for Nature / Coalition Clean Baltic. Urho, L., Pennanen, J. T. and Koljonen, M.-L. 2010. Fish. – In: Rassi, P., Hyv ä rinen, E., Jusl é n, A. et al. (eds), The Red List of Finnish Species. Ymp ä rist ö ministeri ö & Suomen ymp ä rist ö keskus, Helsinki, pp. 336 – 343. Was, A. and Wenne, R. 2002. Genetic differentiation in hatchery and wild sea trout (Salmo trutta) in the southern Baltic at microsatellite loci. – Aquaculture 204: 493 – 506. Wood, C. C., McKinnell, S., Mulligan, T. J. et al. 1987. Stock identifi cation with the maximum-likelihood mixture model: sensitivity analysis and application to complex problems. – Can. J. Fish. Aquat. Sci. 44: 866 – 881. and S. trutta labrax of cultured brown trout. – Iranian J. Fish. Sci. Short Comm. 12: 723 – 732. Koljonen, M.-L. 1989. Uudenmaan meritaimenkantojen geneettinen tutkimus. – Suomen Kalastuslehti 3: 128 – 131. Koljonen, M.-L. 2006. Annual changes in the proportions of wild and hatchery Atlantic salmon ( Salmo salar ) caught in the Baltic Sea. – ICES J. Mar. Sci. 63: 1274 – 1285 . Koljonen, M.-L. and Saura, A. 1992. Kymijoen meritaimen ja lis ä ä ntyv ä n kannan alkuper ä . – Suomen. Koljonen, M.-L., Marttinen, M. anda Koskiniemi, J. 1992. Karjaanjoen vesist ö ss ä on perinn ö llisesti arvokkaita purotaimenkantoja. – Suomen Kalastuslehti 3: 4 – 7. Koljonen, M.-L., Janatuinen, A., Saura, A. et al. 2013. Genetic structure of Finnish and Russian sea trout populations in the Gulf of Finland area. – Working Papers Finnish Game Fish. Inst. 25/2013. Laird, P. W., Zijderveld, A., Linders, K. et al. 1991. Simplifi ed mammalian DNA isolation procedure. – Nucleic Acid Res. 19: 4293 Langella, O. 1999. Populations 1.2.28 (12/5/2002): a populations genetic software. – CNRS UPR9034.  http://bioinformatics. org/ ∼ tryphon/populations/  Lehtonen, P. K., Tonteri, A., Sendek, D. et al. 2009. Spatiotemporal genetic structuring of brown trout ( Salmo trutta L.) populations within the River Luga, northwest Russia. – Conserv. Genet. 10: 281 – 289. Marttinen, M. and Koljonen, M.-L. 1989. Uudenmaan meritaimenkantojen inventointi ja geneettinen tutkimus. – Uudenmaan Kalastuspiirin Kalastustoimiston tiedotus. No. 4. Moran, P., Bromaghin, J. F. and Masuda, M. 2014. Use of genetic data to infer population-specifi c ecological and phenotypic traits from mixed aggregations. – PLoS ONE 9(6): e98470. Nei,. M., Tajima. F. and Tateno. Y. 1983. Accuracy of estimated phylogenetic trees from molecular data. – J. Mol. Evol. 19: 153 – 170. Nuotio, E. and Koskiniemi, J. 1995. Varsinais-Suomen purotaimenselvitys. Maaja mets ä talousministeri ö , Kalaja riistaosasto. – Kalaja riistahallinnon julkaisuja 16: 22 – 61. Page, R. D. M. 1996. TREEVIEW: an application to display phylogenetic trees on personal computers. – Comput, Appl. Biosci. 12: 357 – 358. Pella, J. J. and Masuda, M. 2001. Bayesian methods for analysis of stock mixtures from genetic characters. – Fish. Bull. 99: 151 – 167. Peuhkuri, N., Saura, S., Koljonen, M.-L. et al. 2014. Current state and restoration of sea trout and Atlantic salmon populations in three river systems in the eastern Gulf of Finland. – Working Papers Finnish Game Fish. Inst. 26/2014. Red Data Book of Estonia. 2008. – Commission for Nature Conservation of the Estonian Academy of Sciences.  http://elurikkus.ut.ee/kirjeldus.php?lang  eng & id  172944 & rank  70 & id_puu  172944 & rank_puu  70  accessed 27-08-2014. 194 M.-L. Koljonen et al. Hereditas 151 (2014) Table A1. Analysed sea trout river population samples for the baseline data from Finland, Russia and Estonia. The sampled river, country of origin, tributary or area, sampling year, number of individuals and originality of the sampled population are presented. No River Country Tributary Year n Originality West coast, ELY Centre for Southwest Finland 1 Aurajoki FIN 2006 37 Introduced 2 Paimionjoki FIN V ä h ä joki, Karhunoja 2004, 2008 22 Original 3 Purilanjoki FIN 2011 15 Original 4 Uskelanjoki FIN Pitk ä koski, Kaukolankoski, Haukkalankoski, Hitolanjoki,Terttil ä njoki 2007 57 Original, Introduced 5 Kiskonjoki FIN Perni ö njoki, Juottimenoja-Piilioja, Pakapy ö lin Lohioja 2008 50 Original Middle coast, ELY Centre for Uusimaa 6 Fiskarsinjoki FIN Main stream 2010 50 Introduced 7 Ingarskilanjoki FIN Main stream, P ä rthyvelb ä cken, Kr ä mars 2005 192 Original 8 Siuntionjoki FIN Passilankoski 2010 15 Original 9 Mankinjoki FIN Espoonkartanonkoski, Gumb ö lenjoki 2005, 2008, 2010 133 Original 10 Espoonjoki FIN Main stream, Glomsinjoki, Glimsinjoki 2008, 2010 72 Original 11 Vantaanjoki FIN Vantaankoski, Pitk ä koski, Ruutinkoski, Nukarinkoski, Longinoja 2010 207 Mixed, Introduced 12 Sipoonjoki FIN Ritob ä cken, Byab ä cken 2010 46 Original 13 Koskenkyl ä njoki FIN Hammarfors, Kvarnfors, K ä kikoski, Sahakoski 2010 31 Introduced East coast, ELY Centre for Southeast Finland 14 Kymijoki FIN Main stream 2006, 2010 96 Introduced 15 Isojoki FIN Lapv ä ä rtin Isojoki, Hatchery stock 2006 – 2008 98 Hatchery 16 Summanjoki FIN Main stream 2003, 2008 19 Mixed 17 Virojoki FIN Saarasj ä rvenoja 2003, 2005, 2011 80 Original Border rivers 18 Urpalanjoki FIN/ RUS Urpalanjoenpuro (RUS), Main stream (RUS, FIN) 1999, 2005, 2006, 2010, 2012 69 Original 19 Santajoki (Kaltonjoki) FIN/ RUS Main stream (RUS) 2005, 2006 50 Original 20 Vilajoki FIN/ RUS Main stream (RUS) 2001, 2004, 2005, 2006 33 Original 21 Tervajoki FIN/ RUS Tervajoenpuro (RUS) 2004, 2005 48 Original 22 Rakkolanjoki FIN/ RUS Main stream (RUS), Hanhijoki (RUS) 2004, 2005, 2006 109 Original 23 Mustajoki FIN/ RUS Main stream (FIN, RUS), Kananoja (RUS), Tupakkamyllynoja (FIN), Alhonpuro (FIN), P ö lkkyoja (FIN) 2004, 2005, 2006, 2007, 2008, 2012 438 Original 24 Kilpeenjoki FIN/ RUS Main stream (RUS) 2001, 2006 16 Original Russian rivers 25 R ö mp ö tinpuro RUS 2005 50 Original 26 Myllyoja RUS 2005 44 Original 27 Koivistonpuro RUS 2005 50 Original 28 Penttil ä noja RUS 2004, 2005 50 Original 29 Kello-oja RUS 2004, 2005 54 Original 30 Lohijoki RUS 2004, 2005 56 Original 31 Papinoja RUS 2004, 2005 46 Original 32 Toivolanpuro RUS 2004, 2005 52 Original 33 Notkopuro RUS 2006 51 Original 34 Jukkolanpuro RUS Three close streams 2004, 2005 148 Original Appendix 1 Continued Hereditas 151 (2014) Genetic mixed stock analysis of sea trout catches 195 No River Country Tributary Year n Originality 35 Inojoki RUS Main stream and two tributaries 2004, 2005 98 Original 36 Pikkuvammeljoki RUS 2006 50 Original 37 Vammeljoki RUS 2004, 2005, 2006 50 Original 38 Tyrisev ä noja RUS 2004, 2005 48 Original 39 Hurrinoja RUS 2004, 2005 54 Original 40 Terijoki RUS 2004, 2005 45 Original 41 Huumosenoja RUS 2004, 2005 50 Original 42 Kuokkalanpuro RUS Two close streams 2006 23 Original 43 Rajajoki RUS Main stream, Siesjoki 2006 21 Original 44 Voronka RUS 2005 23 Original 45 Sista RUS 2005 64 Original 46 Havlonka RUS River Tchornaja 2005 29 Original 47 Luga RUS River Solka, R. Lemovzha, Vudoni stream 2005 115 Original Estonian rivers 48 P ü haj õ gi EST 2010, 2011 176 Mixed 49 Kunda EST 2009, 2010 83 Original 50 Toolse EST 2011 49 Original 51 Selja EST 2010, 2011 116 Mixed 52 Loobu EST 2010, 2011 179 Original 53 Valgej õ gi EST 2010 39 Mixed 54 Pudisoo EST 2011 22 Mixed 55 Mustoja EST 2009 36 Mixed 56 Pirita EST Leiva 2010 37 Original 57 V ä ä na EST 2011 68 Original 58 Keila EST 2010, 2011 66 Original 59 Vasalemma EST 2010, 2011 99 Original 4224 Table A2. Microsatellite loci used for sea trout analysis. References, multiplexes, dyes and primer concentrations are also indicated. Locus Reference Multiplex Dye Primer concentration 1 BS131 Estoup et al. 1998 MP 1 VIC 0.03 μ M 2 OneU9 Schribner et al. 1996 MP 2 VIC 0.03 μ M 3 SSa197 O’Reilly et al. 1996 MP 1 NED 0.02 μ M 4 SSa289 McConnell et al. 1995 MP 1 PET 0.30 μ M 5 Ssa407 Cairney et al. 2000 MP 1 NED 0.15 μ M 6 SSa85 McConnell et al. 1995 MP 2 VIC 0.02 μ M 7 Ssosl311 Slettan et al. 1995 MP 2 NED 0.07 μ M 8 SSosl417 Slettan et al. 1995 MP 1 PET 0.04 μ M 9 SSosl438 Slettan et al. 1996 MP 2 VIC 0.07 μ M 10 SSsp1605 Patterson et al. 2004 MP 2 NED 0.04 μ M 11 SSsp2201 Patterson et al. 2004 MP 1 6-FAM 0.03 μ M 12 Str15INRA Estoup et al. 1993 MP 1 6-FAM 0.05 μ M 13 Str60lNRA Estoup et al. 1993 MP 2 PET 0.04 μ M 14 Str73lNRA Estoup et al. 1993 MP 1 VIC 0.04 μ M 15 Str85lNRA Presa and Guyomard 1996 MP 2 6-FAM 0.40 μ M 16 Strutt58 Poteaux 1995 MP 2 6-FAM 0.30 μ M Table A1. Continued.