Genetic effects of supportive stockings on native pikeperch populations in boreal lakes - three cases, three different outcomes

Gene ic e ec s o suppo i e s ockings on na i e pikepe ch popula ions in bo eal lakes – h ee cases, h ee di e en ou comes MATTI SALMINEN 1 , MARJA-LIISA KOLJONEN 1 , MARJATTA S Ä IS Ä 2 and JUKKA RUUHIJ Ä RVI 3 1 Finnish Game and Fishe ies Resea ch Ins i u e, Helsinki, Finland 2 Depa men o Ag icul u al Sciences, Uni e si y o Helsinki, Helsinki, Finland 3 Finnish Game and Fishe ies Resea ch Ins i u e, E o, Finland Salminen, M., Koljonen, M.-L., S ä is ä , M. and Ruuhij ä i, J. 2011 . Gene ic e ec s o suppo i e s ockings on na i e pikepe ch popula ions in bo eal lakes – h ee cases, h ee di e en ou comes. – He edi as 149 : 1–15. Lund, Sweden. eISSN 1601-5223. Recei ed 27 May 2011. Accep ed 20 Sep embe 2011. The gene ic consequences and gene l ow o pikepe ch ( Sande luciope ca ) s ocking we e assessed in h ee bo eal lakes based on admix u e model analysis and compa ison o he p e- and pos - elease pa e ns o gene ic a iabili y a 9 DNA mic osa elli e loci in he ecipien popula ions. In wo ou o he h ee cases, he eleases o i sh om o eign popula ions caused signi i can changes in he gene ic s uc u e o he ecipien popula ion. The la ges changes we e obse ed in Lake Ouluj ä i, whe e he pos - elease sam- ple was almos iden ical o he eleased Lake Vanajanselk ä popula ion, and abou 90% o he ca ch was composed o he eleased popula ion. The gene ic composi ion o Lake Lohjanj ä i pikepe ch also shi ed ma kedly owa ds ha o he eleased Lake Vana- janselk ä popula ion, and abou hal o he la e ca ch was o eleased Vanajanselk ä o igin. In Lake Vanajanselk ä , in con as , eleases o pikepe ch om lakes Painio and A e ia had only a small impac on he gene ic s uc u e o he pikepe ch popula ion. These esul s indica e ha he cu en s ocking p ac ices c ea e an e ec i e a i i cial gene l ow ha may s ongly shape and educe he gene ic di e en ia ion among he emaining na i e pikepe ch popula ions. A common ea u e o all h ee cases was he lack o p io app aisal o he po en ial gene ic and ecological isks in ela ion o he expec ed bene i s o he elease p og ammes. Ma i Salminen, Finnish Game and Fishe ies Resea ch Ins i u e, PO Box 2, FIN-00791 Helsinki, Finland. E-mail: ma i.salminen@ k l. i The na i e dis ibu ion o pikepe ch ( Sande luciope ca ) in no he n Eu ope is assumed o be ela ed o he Lake Ancylus ( eshwa e ) s age o he p esen Bal ic Sea (L Ö NNBERG 1899; LEHTONEN e al. 1996), which ca 9200 – 9000 BP (B J Ö RCK 1995) p o ided he species a dis ibu- ion pa h o a eas co e ed by he o me lake, up o 100 – 150 m abo e he p esen wa e le el o he Bal ic Sea. Since he ini ial coloniza ion, adap i e selec ion, gene l ow and d i ha e shaped he pa e ns o gene ic di e si y wi hin he species, esul ing in a ela i ely high le el o gene ic a iabili y (B J Ö RKLUND e al. 2007; S Ä IS Ä e al. 2010) compa able o ha epo ed in Scandina ian whi e- i sh ( Co egonus la a e us, S Ä IS Ä e al. 2008), g ayling ( Thymallus hymallus, K OSKINEN e al. 2000), pe ch ( Pe ca l u ia ilis , NESB Ø e al. 1999) and A lan ic salmon popula- ions ( Salmo sala , S Ä IS Ä e al. 2005). Since he la e 1800s, he dis ibu ion and gene ic di e - si y o pikepe ch has also been shaped by human-in l u- enced gene l ow h ough in oduc ions and ha che y eleases. In Denma k, he i s in oduc ion ook place in 1898, and since hen pikepe ch ha e been es ablished in o e 70 wa e bodies ( DAHL 1982). In Sweden, No way and Finland he pikepe ch is a na i e species, bu has also been he mos commonly in oduced non-salmonid i sh ( TAMMI e al. 2003). In Finland, he pikepe ch is na i e in ca 650 lakes, bu due o in oduc ions, he p esen dis ibu ion includes ca 2300 lakes ( LAPPALAINEN and TAMMI 1999). The expansion o pikepe ch has been exclusi ely ega ded as a posi i e phenomenon, unlike in Tu key, he UK and Denma k, o ins ance, whe e he expansion has in some cases in ol ed undesi able ecological e ec s ( CRIVELLI 1995; COWX e al. 1997; JEPSEN e al. 2000). Un il he mid-1900s, in oduc ions in Finland we e mos ly ca ied ou using epea ed ans e s o adul i sh o e ilized eggs ( HALME 1961, 1962), a me hod ha seems o ha e been qui e e ec i e in ans e ing gene ic ma e ial ( S Ä IS Ä e al. 2010). Some o he mos i al na u ally ep o- ducing pikepe ch popula ions in Finland s em om hese ea ly in oduc ions, including wo majo sou ce s ocks o cu en ha che y p oduc ion. Apa om he s ocking p es- su e (no. o s ocking e en s and indi iduals eleased; KOLAR and LODGE 2001) and he ai s o he species, he success o he ea ly in oduc ions may ela e o he a ailabili y o a ou able habi a s a al i udes abo e he Lake Ancylus wa e le el, and o he ac ha he sou ce s ocks we e usu- ally loca ed close o he ecipien lakes. In he Rh ô ne del a o F ance, he in asi e success o pikepe ch has simila ly been explained by he ai s o he species and he s ocking p essu e, main aining a high le el o gene ic a iabili y in he in oduced popula ions ( POULET e al. 2008). A new e a o pikepe ch s ocking was ca alyzed by he collapse o se e al impo an pikepe ch s ocks du ing he He edi as 149: 1–15 (2012) © 2012 The Au ho s. This is an Open Access a icle. DOI: 10.1111/j.1601-5223.2011.002230.x 2 M. Salminen e al. He edi as 149 (2012) 1960s and ea ly 1970s ( COLBY and LEHTONEN 1994). The managemen s a egy de ised in his si ua ion was based on he s ocking o young-o - he-yea (YOY) pikepe ch. Mo e e i cien ea ing me hods we e de eloped ( RUUHI- J Ä RVI and HYV Ä RINEN 1996), inc easing he p oduc ion o YOY pikepe ch o ca 10 million yea ⫺1 ( ANON . 2004). The p oduc ion was mainly di ec ed o new in oduc ions and e-s ocking p ojec s aiming a e-es ablishing los pike- pe ch popula ions. In ecen yea s, he i sh ha e mainly been used in a ious enhancemen p ojec s aimed a he mi iga ion o ep oduc ion ailu es o he e ec s o o e - i shing, o simply a imp o ing i shing possibili ies. The popula i y o enhancemen p og ammes has aised ques ions abou hei sus ainabili y. Despi e he high cos s o he p og ammes, hei economic bene i s ha e o en no been p ope ly e alua ed. S ocking also has been a emp ing solu- ion in many cases, as i educes he need o s ic i shing egula ion. The main conce n, howe e , is he po en ial ad e se e ec o eleases o o eign gene ic ma e ial on he gene ic di e si y and adap abili y o he species as a whole. Due o insu i cien con ol and planning, p ac ically all s ock- ing p og ammes in Finland ha e elied on h ee o ou sou ce s ocks, all o hem wi h a sou he n o igin. The possibly small e ec i e popula ion size o he b oods ocks used has also aised conce ns, as i causes loss o di e si y ( RYMAN and L AIKRE 1991). Du ing he las wo decades, p ac ically all emaining indigenous eshwa e pikepe ch s ocks a e likely o ha e been subjec ed o he l ow o o eign sou he n genes h ough ha che ies. The gene ic e ec s o eleases such as compe i ion be ween na i e and in oduced species o s ocks, he eplacemen o na i e s ocks, mixing o s ocks and ex inc- ion o na i e s ocks ha e been widely discussed ( COWX e al. 1997). In his s udy, we assessed he gene ic consequences o pikepe ch s ockings o indigenous na i e popula ions in h ee bo eal lakes in Finland: Lake Lohjanj ä i, Lake Vanajanselk ä and Lake Ouluj ä i. Gene ic in o ma ion was a ailable om he h ee ecipien pikepe ch popula- ions be o e eleases, om he h ee eleased ha che y popula ions and also he admixed popula ions in each lake a e he eleases, allowing us o assess he p opo ion o new gene ic ma e ial in he na i e popula ions a e he eleases. Finally, o add ess he o e all sus ainabili y o cu en s ocking p ac ices, we con as ed he obse ed gene ic e ec s o he h ee s ocking p og ammes wi h hei expec ed socioeconomic bene i s. MATERIAL AND METHODS S udy lakes Lake Lohjanj ä i The indigenous pikepe ch is ec ea ionally and econom- ically one o he mos impo an i sh species in Lake Lohjanj ä i (a ea 89 km 2 , mean dep h 13 m), sou he n Finland (Fig. 1). In mail su eys, epo ed pikepe ch ca ches om 1981 – 2002 a ied om 4 – 21 ons yea ⫺1 ( SALMINEN and RUUHIJ Ä RVI 2004). Abou 90% o he ca ches ha e been aken by gillne s, while olling accoun s o 10% o he o al yield. The ca ches a e mainly used o household needs, bu some i she men also sell i sh. L EHTONEN and MIINA (1988) epo ed high i shing mo - ali y and a low age a ec ui men (4 – 6 yea s) o Lake Lohjanj ä i pikepe ch. G ow h o e - i shing was ega ded as e iden and ec ui men o e - i shing possible. A la ge gill-ne mesh size (50 – 55 mm, ba leng h) and a la ge minimum landing size (MLS; 40 – 42 cm) we e sugges ed and in 1992 also implemen ed o inc ease he age a ec ui men . Releases o YOY pikepe ch we e also expanded o inc ease and s abilize ca ches, and in he 1990s hey eached he le el o 100 000 indi iduals yea ⫺1 (ca 11 ind. ha ⫺1 , Fig. 2). N 0100 km ESRI 1999 © Oulujä i Lohjanjä i Vanajanselkä A e ia Painio Ki ijä i Kemijä i FINLAND Fig. 1. The loca ion o he h ee s udy lakes (blue do s; Lakes Lohjanj ä i, Vanajanselk ä and Ouluj ä i) and he h ee pike- pe ch popula ions addi ionally used in he eleases (g een do s; Lakes A e ia, Painio and Ki ij ä i). The pikepe ch popula ion o Lake Kemij ä i (yellow do ) was used as an ou g oup in he gene ic dis ance analyses. He edi as 149 (2012) Gene ic e ec s o suppo i e s ockings on pikepe ch 3 Lake Lohjanjä i 0 20 000 40 000 60 000 80 000 100 000 120 000 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 B2: 1998-2002B1: 1991-1995 Lake Vanajanselkä 0 10 000 20 000 30 000 40 000 50 000 60 000 70 000 80 000 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 A: 1984-1986 B: 1998-2002 Lake Oulujä i 0 100 000 200 000 300 000 400 000 500 000 600 000 700 000 800 000 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 A e ia Vanajanselkä Painio A: 1936-1943, 1961-1968 B2: 2001B1: 1990-1992 A: 1979-1981 Fig. 2. Releases o YOY pikepe ch acco ding o he yea -class (1979 – 2003) and sou ce s ock in Lake Lohjanj ä i, Lake Vana- janselk ä and Lake Ouluj ä i. A ows indica e he yea -classes o p e- (A) and pos -s ocking (B) samples (Table 1). Lake Vanajanselk ä Lake Vanajanselk ä is he la ges lake (103 km 2 , mean dep h 8 m) o he Vanaja esi wa e cou se in he Ri e Kokem ä enjoki basin. The ela i ely s ong and s able indigenous pikepe ch popula ion ( TOIVONEN e al. 1981) suppo s impo an pikepe ch i she ies and has also se ed as a sou ce o ea ly ans e s o adul i sh and eggs ( HALME 1961, 1962; RUUHIJ Ä RVI and SALMINEN 1992), and since he 1980s also o la ge-scale YOY p oduc ion ( RUUHIJ Ä RVI and HYV Ä RINEN 1996). Pikepe ch ca ches o ec ea ional i she men ha e ecen ly been ca 20 ons yea ⫺1 . Se en p o essional i she men (in 2006, Pekka Ko honen pe s. comm.) also i sh pikepe ch in he lake, hei ca ch being 5 – 10 ons yea ⫺1 . Gill-ne ing wi h mainly 45 mm ne s accoun s o ca 90% o he pikepe ch ca ch, he es being caugh mainly by olle s ( KIVINEN 2009). Despi e he good g ow h a e o Lake Vanajanselk ä pikepe ch, he ecommended ises in he minimum landing size (p esen 37 cm) and gill-ne mesh size ha e no been implemen ed. Ins ead, enhance- men eleases ha e been ca ied ou (Fig. 2), despi e he good ec ui men o he na i e popula ion. Lake Ouluj ä i Un il he la e 1950s, Lake Ouluj ä i (a ea 928 km 2 , mean dep h 7.6 m) pikepe ch sus ained a l ou ishing i she y, wi h ca ches amoun ing o 100 – 150 ons yea ⫺1 ( SUTELA and HYV Ä RINEN 2002). The ea e , ca ches began o decline and he bo om, less han 100 kg yea ⫺1 , was eached in he ea ly 1980s. In he 1990s, ca ches began o eco e and ha e ecen ly been a ound 100 ons yea ⫺1 . Pikepe ch a e mainly aken by ec ea ional gill-ne e s and angle s, bu also p o ide an impo an a ge o he 32 p o essional i she men ope a ing in he lake (in 2006, Pekka Ko honen pe s. comm.). Po en ial causes o he collapse o he pikepe ch s ock and ca ches in Lake Ouluj ä i, as in nume ous o he bo eal lakes, we e he declining empe a u es in 1940 – 1970 com- bined wi h inc easing exploi a ion caused by he shi om co on o mo e e i cien nylon mono i lamen gill- ne s ( COLBY and LEHTONEN 1994). Enhancemen s ocking was he main managemen ac ion ecommended in his si ua ion ( SALOJ Ä RVI e al. 1981), and based on his ecom- menda ion eleases we e s a ed in 1985, eaching he le el o 500 000 indi iduals yea ⫺1 in he 2000s (Fig. 2, ca 5 ind. ha ⫺1 ). The p oduc ion and eleases o ha che y pikepe ch The p oduc ion o YOY pikepe ch elies on he annual cap u e o wild spawne s om he spawning a eas using ap-ne s ( SALMINEN and RUUHIJ Ä RVI 1991; RUUHIJ Ä RVI and H YV Ä RINEN 1996). A e ha ching he la ae a e s ocked a 20 000 – 40 000 ind. ha ⫺1 in o 1 – 2 ha dams p oducing na u- al ood. The a e age ha es in la e Augus – Sep embe , i.e. a e 3 – 4 mon hs o ea ing, is abou 10 000 YOY pike- pe ch ha ⫺1 , anging in o al leng h om 60 – 80 mm. As he backg ound o YOY pikepe ch has no been an issue among he manage s, he p oduc ion o la ae has concen a ed on a limi ed numbe o dense sou ce popula- ions, all wi h a sou he n o igin. Ano he ac o po en ially in l uencing he gene ic consequences o eleases is amily size, i.e. he numbe o ju eniles p oduced pe emale. Gi en he high ecundi y in a i i cial ea ing ( SALMINEN e al. 1992) and high su i al du ing ea ing, one la ge emale may p oduce up o 200 000 ju eniles, which may 4 M. Salminen e al. He edi as 149 (2012) 493 (0.5 – 3.0 kg) adul i sh om Lake Vanajanselk ä (Fig. 3, Table 1) ( SUTELA e al. 1995). Pikepe ch samples The analysis o gene ic impac s was based on admix u e modelling and he compa ison o gene ic a iabili y wi hin he h ee ecipien popula ions in Lakes Lohjanj ä i, Vana- janselk ä and Ouluj ä i, be o e and a e s ocking. F om all h ee lakes, one p e-s ocking sample and one o wo (in Lake Ouluj ä i) pos - elease admix u e samples we e analyzed (Table 1). The h ee pikepe ch popula ions addi ionally used in he eleases ( om lakes A e ia, Painio and Ki ij ä i) we e also sampled o he DNA analyses. The pikepe ch popula- ion o Lake Kemij ä i was used as an ou g oup in he gene ic dis ance analyses ( S Ä IS Ä e al. 2010). In he case o Lohjanj ä i i was possible o o ganize an addi ional es , independen o he gene ic da a, as he ju eniles ( om Lake A e ia and Lake Vanajanselk ä ) eleased he e in 1991 – 1995 we e ma ked using ho - b anding ( SAURA 1996; SALMINEN and RUUHIJ Ä RVI 2004), allowing he iden i i ca ion o hei s ock o o igin (Table 1, Fig. 2). Thei s ock-speci i c p opo ions we e eco ded among adul pikepe ch sampled in 1994 – 2000. hen cons i u e a la ge pa o o e en he whole yea -class s ocked in one o wo la ge o se e al smalle lakes. To p e- en his, ha che ies ha e been ad ised o mix he o sp ing o emales ha spawn a app oxima ely he same ime, bu i is no known o wha ex en his ac ually happens. Ou h ee s udy lakes o e ypical examples o he complex ans e s o pikepe ch om one wa e -body o ano he : 1) In Lake Lohjanj ä i, h ee o eign popula ions we e used in he eleases (Fig. 3, Table 1). Two o hem (Lake A e ia and Lake Painio) a e gene ically close o Lake Lohjanj ä i pikepe ch ( S Ä IS Ä e al. 2010), as hey ha e hei oo s in ea ly (1930s) ans e s om he same Lake Lohjanj ä i, while he hi d sou ce popula ion (Lake Vanajanselk ä ) is gene ically mo e dis an ( S Ä IS Ä e al. 2010). 2) In Lake Vanajanselk ä , bo h i s own indigenous popula- ion, and he popula ions o Lake A e ia and Lake Painio ha e been used in he eleases (Fig. 3, Table 1). 3) In Lake Ouluj ä i, pikepe ch om Lake Vanajanselk ä ha e been s ocked, mos ly ia a special b ood- i sh lake (Lake Ki ij ä i) housing a pikepe ch popu- la ion es ablished in he la e 1980s by ans e ing Lake Lohjanjä i Lake Vanajanselkä Lake Oulujä i Lake Ki ijä i Lake A e ia Lake Painio 1985–2003 1981–2003 1980s 1993–2003 1999–2001 1930s 2000 1930s 1993–1996 A e ia (Lohjanjä i) - S1/o iginal (1984) - S2/ om Lohjanjä i (1994–2000) Painio (Lohjanjä i) - S3/o iginal (2006) Case Lohjanjä i - S4/o iginal (1982–1986) - S5/wild (1994–2000) - S6/mixed (2002–2003) Case Vanajanselkä - S7/o iginal (1988–1991) - S8/ om Lohjanjä i (1994–2000) - S9/mixed (2001–2006) Case Oulujä i - S11/o iginal (1946/1974) - S12/mixed - S13/mixed (2002) Ki ijä i (Vanajanselkä) - S10/o iginal (2008) Fig. 3. Gene ic backg ound and popula ion ans e his o y o he sampled Finnish pikepe ch popula ions. Dashed lines indica e ans- e s o adul i sh, solid lines ans e s o YOY. Sample numbe s (S) e e o Table 1. He edi as 149 (2012) Gene ic e ec s o suppo i e s ockings on pikepe ch 5 be ween popula ions we e assessed by he - es o pai ed obse a ions ( NEI 1987). Analysis o he di e ences be ween samples was based on allele equency di e - ences, using pai wise F s alues ( WEIR and COCKERHAM 1984), which we e es ima ed wi h FSTAT e . 2.9.3 ( GOUDET 2001). S anda d de ia ions and con i dence in e als we e es ima ed h ough boo s apping. Gene ic dis ances be ween samples we e calcula ed using Nei ’ s D A dis ances ( NEI e al . 1983). A phylogene ic ee was cons uc ed using a neighbou joining (NJ) algo i hm ( SAITOU and NEI 1987) wi h DISPAN so wa e ( OTA 1993). Boo s apping wi h 1000 eplica es was used o es he s a is ical s eng h o he b anches. To analyse he popula ion mix u es in he ecipien lakes, Bayesian clus e ing algo i hms we e used in he p og am STRUCTURE ( e . 2.2.3) ( PRITCHARD e al. 2000; F ALUSH e al. 2003; PRITCHARD and WEN 2004). The p o- g am assigns indi idual i sh o one o mo e g oups, wi h hei ela i e equency o p edic ed membe ship o alling 1.00. This allowed analyses o admixed popula ions wi h p io in o ma ion om he sou ce popula ions. In all cases, he op ions o using popula ion in o ma ion and applying he admix u e model o he unknown ca ch sample we e used. Mo eo e , alpha , desc ibing he amoun o popula- ion mixing, was allowed o a y in all cases and was di - e en o each popula ion. The o al leng h o he uns was 150 000 i e a ions, wi h a bu n-in o 100 000 i e a ions and he las 50 000 i e a ions being used o he es ima es. DNA analysis and calcula ions DNA ex ac ion and mic osa elli e labo a o y analysis we e conduc ed acco ding o he desc ip ion o S Ä IS Ä e al. (2010). Va ia ion in he ollowing nine mic osa elli e loci was de e mined: P l aL3, P l aL8 ( LECLERC e al. 2000), S i4, S i6, S i18, S i33 ( BORER e al. 1999), S iL7, S iL8 and S iL11 ( WIRTH e al. 1999). The numbe o alleles in samples was compa ed using a a e ac ion-based allelic ichness measu e ( EL MOUSADIK and PETIT 1996; PETIT e al. 1998), which was calcula ed wi h FSTAT so wa e e . 2.9.3 ( GOUDET 2001). The p o- g am calcula es allelic ichness o he smalles numbe o indi iduals yped o any locus. Each locus was calcula ed sepa a ely wi h he same numbe o e all popula ions, and he mean was calcula ed o e loci. Popula ion di e en ia- ion was analyzed wi h he GENEPOP ( e . 4.0) so wa e package ( RAYMOND and ROUSSET 1995; ROUSSET 2008) wi h Ma ko chain pa ame e s, 300 ba ches and 3000 i e a ions. The Bon e oni co ec ion ( RICE 1989) was applied o co ec o he numbe o es s in he H-W equi- lib ium. The po en ial occu ence o null alleles was also checked by GENEPOP. Indica ion o a null allele in locus S i33 could be seen, bu i was s ill included in he analysis, as indica ion o null alleles may esul om Ha dy-Weinbe g de ia ions as well. The expec ed he e ozygosi y le el in each sample was cal- cula ed using Popgene e . 1.32 ( YEH and BOYLE 1997). Di - e ences in he mean he e ozygosi ies and allele ichness ’ s Table 1. Analysed pikepe ch samples. Sample numbe , gene ic backg ound o he popula ion, sampling si e (Fig. 1), sampling pe iod (A ⫽ p e-s ocking, B ⫽ pos -s ocking), sample size (n), ca ch yea , yea class and s a us o s udied pikepe ch samples. * Indi iduals o he sample we e ecognized om ho b anding ma ks. Sample Gene ic backg ound Sampling lake Sampling pe iod n Ca ch yea Yea class S a us S1 Lohjanj ä i (b oods ock) A e ia – 63 1984 1979 – 1982 o iginal/sou ce S2 Lohjanj ä i Lohjanj ä i B1 59 1994–2000 1991 – 1993 Lake A e ia o igin * S3 Lohjanj ä i (b oods ock) Painio – 74 2006 1985 – 2005 o iginal/sou ce S4 Lohjanj ä i Lohjanj ä i A 60 1982–1986 1979 – 1981 O iginal S5 Lohjanj ä i Lohjanj ä i B1 60 1994–2000 1991 – 1995 wild * S6 Lohjanj ä i and Vanajanselk ä Lohjanj ä i B2 200 2002–2003 1998 – 2002 cu en , mixed S7 Vanajanselk ä Vanajanselk ä A 60 1988–1991 1984 – 1986 o iginal S8 Vanajanselk ä Lohjanj ä i B1 60 1994–2000 1993 – 1995 Vanajanselk ä o igin * S9 Vanajanselk ä and Lohjanj ä i Vanajanselk ä B 208 2001–2006 1998 – 2002 cu en , mixed S10 Vanajanselk ä (b oods ock) Ki ij ä i – 31 2008 1992 – 1998 o iginal/sou ce S11 Ouluj ä i Ouluj ä i A 8, 15 1946, 1974 1936 – 1943, 1961 – 1968 o iginal S12 Ouluj ä i and Vanajanselk ä Ouluj ä i B1 100 1990s 1990 – 1992 mixed S13 Ouluj ä i and Vanajanselk ä Ouluj ä i B2 100 2002 2001 cu en , mixed 6 M. Salminen e al. He edi as 149 (2012) Fig. 4a – d. Es ima ed popula ion s uc u e o pikepe ch om Bayesian STRUCTURE analysis o ou ca ch mix u es. Each indi idual is ep esen ed by a hin e ical line, which is pa i ioned in o K colou ed segmen s ha ep esen i s es ima ed popula ion g oup mem- be ship ac ions. Black lines sepa a e indi iduals om di e en numbe ed samples. Sample names and numbe s a e gi en sepa a ely o each case. ( a ) Lake Lohjanj ä i 2000, Popula ion 1 ⫽ A e ia, ed (S1), 2 ⫽ Painio, g een (S3), 3 ⫽ Vanajanselk ä , blue (S7), and 4 ⫽ o iginal Lohjanj ä i, yellow (S4), 5 ⫽ ca ch mix u e o Lake Lohjanj ä i in 2000 (S6). ( b ) Lake Vanajanselk ä 2000, (Popula ion 1 ⫽ A e ia, ed (S1), 2 ⫽ Painio, g een (S3) and 3 ⫽ o iginal Vanajanselk ä , blue (S7), 4 ⫽ ca ch mix u e om Lake Vanajanselk ä in 2000 (S9). ( c ) Lake Ouluj ä i 1990. Popula ion 1 ⫽ Ki ij ä i, ed (S10), 2 ⫽ Ouluj ä i o iginal, g een (S11) and 3 ⫽ ca ch mix u e om Lake Ouluj ä i in he 1990s (S12). ( d ) Lake Ouluj ä i 2000. Popula ion 1 ⫽ Ki ij ä i (S10), 2 ⫽ Ouluj ä i o iginal (S11) and 3 ⫽ ca ch mix u e om Lake Ouluj ä i in 2000 (S13). Numbe s in b acke s e e o Table 1, sample numbe . Con e gence o all uns was checked. In all cases, he numbe o con ibu ing popula ions, K, was known, which simpli i ed he analyses. Se e al uns we e ca ied ou o each admix u e and he consis ency o he uns was assessed. Changes in K we e also es ed wi h K alues o one mo e o less han he known numbe o popula ions, and esul s wi h he g ea es pos e io p obabili ies a e p esen ed. Admix u e analysis was also pe o med wi h maximum likelihood es ima ion by WANG (2003), included in he LEADMIX so wa e. This is based on he p inciple ha allele equencies o he admixed popula ions should be He edi as 149 (2012) Gene ic e ec s o suppo i e s ockings on pikepe ch 7 he e ozygosi y (H e ⫽ 0.54) in he p e-s ocking sample o Lake Ouluj ä i (S11; Table 2). The lowes allelic ichness (3.6 alleles) was obse ed in he Lake Painio popula ion (S3) and in he p e-s ocking sample om Lake Lohjan- j ä i (S4), and he lowes expec ed he e ozygosi y in he o iginal Lake A e ia popula ion (S1; 0.41). The geno ype dis ibu ions o he h ee na i e popula ions in lakes Lohjanj ä i, Ouluj ä i and Vanajanselk ä did no de ia e om Ha dy–Weinbe g equilib ium a e Bon e oni co - ec ion. The gene ic di e si y was in gene al ela i ely high in all cases bo h be o e and a e s ock ans e s (Table 2), and no dec ease in gene ic di e si y could be obse ed as a esul o s ocking. In all h ee cases, allelic ichness was ac ually somewha highe a e eleases han in he o igi- nal p e-s ocking popula ion. In Lakes Lohjanj ä i and Vanajanselk ä , he mean he e ozygosi y was also highe in he con empo a y han in he na i e popula ion. The only s a is ically signi i can di e ence in he mean he e ozygosi y was be ween he Lake Painio popula ion (S3) and he con empo a y admixed popula ion o Lake Lohjanj ä i (S6), which had a highe di e si y. Allelic ich- ness was also s a is ically signi i can ly highe in he con- empo a y Lohjanj ä i sample (S6) han in he samples om Lake A e ia (S1), Lake Painio (S3) and e en in he sample om he na i e Lake Lohjanj ä i popula ion (S4). Samples S2 and S8 om Lake Lohjanj ä i, bo h ep e- sen ing h ee consecu i e yea -classes o adul F 1 i sh iden i i ed o hei s ock o o igin (Lake A e ia and Lake Vanajanselk ä , co espondingly) by ma king, showed gene ic di e si y compa able o ha in hei sou ce popu- la ions (S1 and S7), indica ing ha gene ic di e si y was linea combina ions o hose o he con ibu ing pa en al popula ions a he ime when admix u e occu s. In addi- ion o some p e ious p og ams, i also akes in o accoun gene ic d i ha has po en ially occu ed in pa en al pop- ula ions and in admix u e popula ions since he admix- u e, and e en he po en ial d i o he pa en al popula ions be o e admix u e occu ed. The p og am addi ionally gi es 95% con i dence in e als o he admix u e p opo - ions. Resul s o he STRUCTURE and LEADMIX analy- sis we e compa ed wi h he popula ion g ouping analysis based on gene ic dis ances. Analysis was ca ied ou o ou ca ch mix u e popula- ions: Lohjanj ä i 2000 (Table 1, S6), Vanajanselk ä 2000 (Table 1, S9) Ouluj ä i 1990s (Table 1, S12) and Oulu- j ä i 2000 (Table 1, S13). Fo he Lake Lohjanj ä i case, he po en ial addi ional con ibu o s we e he popula ions om Lakes Painio, A e ia and Vanajanselk ä , and o Lake Vanajanselk ä , he popula ions om Lakes Painio and A e ia, (Fig. 3). Fo he Lake Ouluj ä i case, he only po en ial con ibu ing o eign popula ion was he Ki i- j ä i b ood s ock o Vanajanselk ä o igin. RESULTS Gene ic di e si y in pikepe ch samples The o e all expec ed mean he e ozygosi y (H e ) o all sam- ples was 0.51 and he mean allelic ichness o e all sam- ples was 4.1 alleles loci ⫺1 . The F ST o e all popula ions was as high as 0.08. The highes allele ichness was obse ed in he Lake Ki ij ä i b oods ock popula ion (S10), wi h 4.6 alleles loci ⫺1 , and highes expec ed Table 2. The gene ic di e si y o he pikepe ch samples o he h ee cases, he sampling lake, ca ch yea , s a us o he popula ion, mean sample size o nine loci, mean expec ed he e ozygosi y ( H e ) and i s s anda d e o (SE), and a e age allelic ichness (A ) o nine DNA mic osa elli e loci. Sample Sampling lake Ca ch yea S a us Mean n/locus H e SE A Case Lohjanj ä i S1 B oods ock, A e ia, 1984 o iginal/ sou ce 58.0 0.41 0.08 3.7 S2 Lohjanj ä i 1994 – 2000 A e ia, o igin 56.6 0.46 0.08 3.9 S3 B oods ock, Painio 2006 o iginal/ sou ce 69.9 0.45 0.08 3.6 S4 Lohjanj ä i 1982 – 1986 o iginal 59.1 0.44 0.09 3.6 S5 Lohjanj ä i 1994 – 2000 wild 59.4 0.45 0.08 4.0 S6 Lohjanj ä i 2002 – 2003 cu en , mixed 195.0 0.51 0.08 4.4 Case Vanajanselk ä S7 Vanajanselk ä 1988 – 1991 o iginal 58.7 0.49 0.09 4.3 S8 Lohjanj ä i 1994 – 2000 Vanajanselk ä , o igin 58.4 0.48 0.09 4.1 S9 Vanajanselk ä 2001 – 2006 cu en , mixed 204.2 0.53 0.08 4.5 Case Ouluj ä i S10 B oods ock, Ki ij ä i 2008 o iginal/ sou ce 30.7 0.52 0.09 4.6 S11 Ouluj ä i 1946, 1974 o iginal 14.6 0.54 0.07 3.8 S12 Ouluj ä i 1990s Mixed 95.3 0.46 0.10 4.0 S13 Ouluj ä i 2002 cu en , mixed 98.4 0.47 0.09 4.2 Mean 0.51 0.08 4.1 8 M. Salminen e al. He edi as 149 (2012) Table 3. Mean p opo ions (%) o indi iduals assigned o each o he p e-de i ned popula ions om each o he ou admixed pikepe ch samples by he Bayesian STRUCTURE p og am (Bayes es ima e), and s ock p opo ion es ima es wi h hei 95% con i dence in e als (CI) om he maximum-likelihood es ima ion o he LEADMIX p og am (MLE es ima e) . Bayes es ima e % MLE es ima e % 5% CI 95% CI S1 Lohjanj ä i 2000 A e ia 6.4 10.2 0.0 20.9 S3 Painio 5.4 0.0 0.0 12.3 S7 Vanajanselk ä 1980 47.8 54.6 45.8 72.5 S4 Lohjanj ä i 1980 40.5 35.2 14.8 43.6 Vanajanselk ä 2000 S1 A e ia 6.0 10.8 10.7 10.8 S3 Painio 2.2 3.1 3.1 3.2 S7 Vanajanselk ä 1980 91.8 86.2 86.1 86.2 Ouluj ä i 1990 S10 Ki ij ä i-Vanajanselk ä 89.0 99.9 92.8 100.0 S11 Ouluj ä i, o iginal 2.7 0.0 0.0 8.5 S12 Ouluj ä i 1990 8.3 – – – Ouluj ä i 2000 S10 Ki ij ä i-Vanajanselk ä 89.8 98.3 89.9 100.0 S11 Ouluj ä i, o iginal 6.5 1.7 0.0 11.8 S13 Ouluj ä i 2000 3.7 – – – in his case qui e e i cien ly ans e ed h ough ea ing and s ocking om he sou ce lakes o he ecipien popula- ion. Admix u e analysis o h ee pikepe ch popula ions Lake Lohjanj ä i In he Lake Lohjanj ä i case, a ma ked gene ic con i- bu ion could be obse ed as esul o he eleases. The admixed ca ch sample was mainly composed o bo h Lake Vanajanselk ä and Lake Lohjanj ä i popula ions (Table 3, Fig. 4a). A la ge p opo ion o indi iduals we e admix u es. Vanajanselk ä pikepe ch comp ised a leas abou hal o he ca ch in Lake Lohjanj ä i. The maximum likelihood es ima e o he con ibu ion o Vanajanselk ä was 54.6% (95% con i dence in e al (CI), 45.8 – 72.5%), and ha o indigenous Lake Lohjan- j ä i pikepe ch was co espondingly only 35.2% (CI: 14.8 – 43.6%) (Table 3). The Bayesian es ima e o he con ibu ion o Vanajanselk ä was sligh ly less, being 47.8%, and o Lohjanj ä i somewha mo e, 40.5%. In all cases, he Bayesian es ima es we e wi hin he 95% con i dence in e als o he maximum likeli- hood es ima es. The es ima ed con ibu ion o he Vanajanselk ä popula- ion in he admix u e was somewha less han he obse ed con ibu ion o ma ked Vanajanselk ä pikepe ch (62 – 72%) o Lake Lohjanj ä i ca ch samples om yea -classes 1993 – 1995 ( SALMINEN and RUUHIJ Ä RVI 2004). The mos commonly eleased s ock in Lake Lohjan- j ä i has been Lake A e ia pikepe ch (Fig. 2). The la - es eleases ha e been ca ied ou wi h Lake Painio pikepe ch, bu e y li le gene ic e ec o hese eleases could be obse ed, despi e he ela i ely la ge numbe s o eleased i sh. The la ge con ibu ion om he Lake Painio pikepe ch eleases could be excluded (shown in Fig. 4a as g een), bu eleases om Lake A e ia we e possibly con ibu ing o he ca ch sample, as he maximum likelihood es ima e was 10% wi h a p oba- bili y in e al o 0 – 20% o he A e ia s ock con ibu- ion. The mo e success ul Lake Vanajanselk ä i sh we e only eleased in 1993 – 1995, and also in much smalle numbe s. Lake Vanajanselk ä In he Lake Vanajanselk ä case, e y li le gene ic con ibu- ion could be seen as a esul o he eleases (Fig. 4b), and he o iginal Vanajanselk ä pikepe ch s ill accoun ed o he majo i y o he ca ch. Acco ding o he Bayesian es ima e, 91.8% o he admix u e gene pool o igina ed om he indigenous Vanajanselk ä popula ion, and wi h he maxi- mum likelihood es ima ion he p opo ion was somewha less, being 86.2% (Table 3). Abou 10% o igina ed om He edi as 149 (2012) Gene ic e ec s o suppo i e s ockings on pikepe ch 9 (Lohjanj ä i g oup), Lakes Vanajanselk ä , Ouluj ä i mixed samples, Ki ij ä i and Lohjanj ä i samples he second g oup (Vanajanselk ä g oup) and he wo no he n na i e popula ions om Lakes Kemij ä i and Ouluj ä i he hi d g oup (Fig. 5). The gene ic di e en ia ion be ween allele equencies o he popula ions be o e and a e eleases emained s a is ically signi i can wi hin all hese g oups, and in ac all pai wise compa isons be ween popula ions we e s a is ically signi i can . The mean gene ic dis ance wi hin he Lohjanj ä i g oup was 0.06 and wi hin he Vanaja esi g oup 0.04. The o igi- nal Lake Ouluj ä i sample was he mos dis inc o he o he samples and g ouped oge he wi h he sample o indigenous pikepe ch om he no he n Lake Kemij ä i (Fig. 1), which was used as an ou g oup in he analysis. In gene al, he esul s on he gene ic s uc u e con i med he changes obse ed in he ecipien popula ions in admix- u e analysis, and esul s om bo h analyses we e cong u- en o all cases. Lake Lohjanj ä i Fi e o he six samples o igina ing om Lake Lohjanj ä i pikepe ch g ouped in o he same b anch in he dend o- g am (Fig. 5). The A e ia 1984 sample was mos simila o he o iginal Lohjanj ä i sample, Painio had some unique ea u es and he ca ch sample om he 1990s (wild i sh iden i i ed by ma king) had also shi ed somewha om he o iginal popula ion. The mos dis inc i e was, howe e , he las sample, he admix u e om he yea s 2002 and 2003, which g ouped in o he Vanajanselk ä g oup ins ead o he Lohjanj ä i g oup. The F ST be ween he Lohjanj ä i he A e ia popula ion eleases, and ha dly any om he Painio popula ion eleases, al hough his popula ion was he mo e commonly used s ock in eleases, and easily dis- inguishable om o he popula ions. The 95% con i dence in e al o he MLE es ima e was na ow, and he Bayes- ian es ima e did no all wi hin hese limi s. Lake Ouluj ä i In he case o Lake Ouluj ä i pikepe ch, he indigenous popula ion has nea ly disappea ed. The esul s show ha a la ge majo i y o he con empo a y Lake Ouluj ä i pike- pe ch popula ion o igina ed om he eleases o Ki ij ä - i-Vanajanselk ä pikepe ch. In bo h mixed samples, om 1990 and 2000, he p opo ion o i sh o igina ing om he eleases was abou 90% when es ima ed wi h he Bayesian me hod (Table 3, Fig. 4c – d). When he po en ial e ec o gene ic d i was included in he maximum-likelihood me hod, he p opo ion o he o iginal Vanajanselk ä popu- la ion inc eased e y close o 100% (Table 3). Fo bo h Ouluj ä i mixed samples, he pos e io p obabili ies we e highe in he Bayesian me hod when he numbe o con- ibu ing popula ions was se o h ee a he han o wo, indica ing ha he admix u e could no comple ely be explained by he wo popula ions, and he esul s a e he e- o e gi en acco dingly. Gene ic di e en ia ion among s ocks be o e and a e s ocking Acco ding o gene ic dis ances, he popula ions g ouped in o h ee main g oups, whe e Lakes A e ia, Painio and he o iginal Lohjanj ä i samples o med he i s g oup Oulujä i 1990s, mixed, S12 90 KEMIJ Ä RVI 2006, o iginal OULUJÄRVI 1946, 1974, o iginal, S11 91 60 Oulujä i 2002, mixed, S13 42 VANAJANSELKÄ 1988–1991, o iginal, S7 82 Vanajanselkä 2001-2006, cu en mixed, S9 36 Vanajanselkä 1994–2000, in Lake Lohjanjä i, S8 Ki ijä i 2008, Vanajanselkä o igin, S10 77 Lohjanjä i 2002–2003, cu en mixed, S6 A e ia 1994–2000, in Lake Lohjanjä i, S2 Lohjanjä i 1990s, wild, S5 54 58 58 78 Painio 2006, S3 A e ia 1984, S1 LOHJANJÄRVI 1980s, o iginal, S4 Fig. 5 . Gene ic dis ances based on nine mic osa elli e DNA loci. Names o indigenous popula ions in capi al le e s. Sample numbe s (S) e e o Table 1.