Phenotypic plasticity in growth and fecundity induced by strong population fluctuations affects reproductive traits of female fish

Pheno ypic plas ici y in g ow h and ecundi y induced by s ong popula ion luc ua ions a ec s ep oduc i e ai s o emale ish Juha Ka jalainen 1 , Olli U panen 1,2 , Tapio Keskinen 1,3 , Hannu Huuskonen 4 , Jouko Sa ala 5 , Pen i Valkeaj€ a i 3 & Timo J. Ma jom€ aki 1 1 Depa men o Biological and En i onmen al Science, Uni e si y o Jy € askyl€ a, P.O. Box 35, Jy € askyl€ a FI-40014, Finland 2 Me s€ ahalli us, Jy € askyl€ a, P.O. BOX 36, Jy € askyl€ a FI-40100, Finland 3 Na u al Resou ces Ins i u e Finland, Su on ie 9A, Jy € askyl€ a FI-40500, Finland 4 Depa men o Biology, Uni e si y o Eas e n Finland, P.O. Box 111, Joensuu FI-80101, Finland 5 Depa men o Biology, Sec ion o Ecology, Uni e si y o Tu ku, Tu ku FI-20014, Finland Keywo ds Age o ma u a ion, co egonids, emb yonic de elopmen , ishe ies, la al de elopmen , ma e nal e ec , s ock luc ua ions. Co espondence Juha Ka jalainen, Depa men o Biological and En i onmen al Science, Uni e si y o Jy € askyl€ a, P.O. Box 35, Jy € askyl€ a FI-40014, Finland. Tel: +358405134865; Fax: +35814617239; E-mail: [email p o ec ed] Funding In o ma ion This wo k was inancially suppo ed by he Minis y o Ag icul u e and Fo es y and Finnish Cul u al Founda ion/Cen al Finland. Recei ed: 18 June 2015; Re ised: 7 Decembe 2015; Accep ed: 9 Decembe 2015 Ecology and E olu ion 2016; 6(3): 779–790 doi: 10.1002/ece3.1936 Abs ac Fish a e known o hei high pheno ypic plas ici y in li e-his o y ai s in ela- ion o en i onmen al a iabili y, and his is pa icula ly p onounced among salmonids in he No he n Hemisphe e. Resou ce limi a ion leads o ade-o s in pheno ypic plas ici y be ween li e-his o y ai s ela ed o he ep oduc ion, g ow h, and su i al o indi idual ish, which ha e consequences o he age and size dis ibu ions o popula ions, as well as hei dynamics and p oduc i - i y. We s udied he e ec o plas ici y in g ow h and ecundi y o endace emales on hei ep oduc i e ai s using a se ies o long- e m incuba ion expe imen s. The wild pa en al ish o igina ed om ou sepa a e popula ions wi h ma kedly di e en densi ies, and hence na u ally induced di e ences in hei g ow h and ecundi y. The ene gy alloca ion o soma ic issues and eggs p io o spawning se ed as a p oxy o o al esou ce a ailabili y o indi idual emales, and i s e ec s on o sp ing su i al and g ow h we e analyzed. Vendace emales alloca ed a a he cons an p opo ion o a ailable ene gy o eggs (pe body mass) despi e di e en g ow h pa e ns depending on he o al esou ces in he di e en lakes; in es men in o eggs hus dic a ed he sha e emaining o g ow h. The ene gy alloca ion o eggs pe mass was highe in young han in old spawne s and he egg size and he ela i e ecundi y di e ed be ween hem: Young emales p oduced mo e and smalle eggs and la ae han old spawne s. In con as o ea lie obse a ions o salmonids, a sho age o ma e nal ood esou ces did no inc ease o sp ing size and su i al. Vendace emales in spa se popula ions wi h ample esou ces and high g ow h p oduced la ge eggs and la ae. Vendace accommoda e s ong popula ion luc ua ions by hei high plas- ici y in g ow h and ecundi y, which a ec hei o sp ing size and conse- quen ly hei ec ui men and p oduc i i y, and accoun o hei pe sis ence and esilience in he ace o high ishing mo ali y. In oduc ion Pheno ypic plas ici y occu s when he pheno ype exp essed by a gi en geno ype changes as en i onmen al condi ions a y (Nussey e al. 2007). Fish a e known o hei high pheno ypic plas ici y in many li e-his o y ai s in ela ion o en i onmen al a iabili y, and his is pa ic- ula ly p onounced among salmonids (Salmonidae) in he No he n Hemisphe e (Skulason and Smi h 1995). T ade- o s in pheno ypic plas ici y be ween li e-his o y ai s ela ed o he ep oduc ion, g ow h, and su i al o indi- idual ish may ha e consequences o he age and size dis ibu ions o popula ions, as well as hei dynamics and p oduc i i y (S ea ns and Koella 1986; Heino e al. 2002). Age and size a ma u a ion o ishes wi h mode - a e- o-high ecundi y, high age a ma u a ion, and low ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. 779 p ema u a ion su i al ha e been widely s udied using he eac ion no m app oach (Heino e al. 2002; E ans e al. 2010; Jonsson e al. 2012; Kokkonen e al. 2015). Those species ha e high a ia ion in age and size a ma u- a ion ela i e o ce ain salmonid species, such as en- dace (Co egonus albula (L.), Fig. 1), wi h ea ly ma u a ion, high ep oduc i e e o , small size, sho li e span, high demog aphic esilience, and low ju enile su - i o ship (Ka jalainen and Viljanen 1994). Fu he mo e, g ow h o endace is s ongly dependen on he popula- ion densi y due o in aspeci ic ood compe i ion (Vilja- nen 1986; Helminen e al. 1992; Ma jom€ aki and Ki jasniemi 1995; Au inen e al. 2000), so ha he we mass o imma u e ish a e he i s g ow h season can a y by as much as ou old be ween successi e yea s (Helminen e al. 1992). In endace popula ions, in e coho compe i ion has been p oposed o induce a quan i a i e ma e nal e ec on o sp ing quali y du ing he g owing season p io o spawning (Ham in and Pe sson 1986). This in e ac ion can be illus a ed by he ollowing chain o associa ion: High young-o - he-yea abundance causes se e e ood compe i ion be ween young-o - he-yea and ma u e ish which leads o slow g ow h and weakened condi ion o ma u e ish wi h de e io a ion in some essen ial quali y o quan i y componen s in he eggs. Rep oduc i e p oduc s o poo quali y will lead o a coho e ec o low su i al o emb yos o la ae, and hus less success ul ec ui men o he nex yea class. Fu he , a high mo ali y o en- dace due o in ensi e ha es ing will simpli y he age s uc u e o he spawning popula ion om a mul icoho i e opa ous s a e o an almos o ally single-coho semel- pa ous s a e and main ains he induced 2-yea gene a ion cyclici y in he popula ion dynamics (Ham in and Pe s- son 1986) as well as he g ow h o indi idual ish (Helmi- nen and Sa ala 1994). Thus, ha es ing may signi ican ly adjus he ep oduc i e ai s o indi iduals and popula- ions in endace, ou model species. Vendace is a special case among i e opa ous ishes, because he age a he i s ep oduc ion o emales in many endace s ocks has e y li le a ia ion: Almos all emales spawn o he i s ime in hei second au umn (young, no ice spawne s a age o 1+) (Ham in and Pe s- son 1986; Huusko and Hy € a inen 2005), and only a small p opo ion o hem su i e o epea spawning in he ol- lowing yea s. Howe e , some endace popula ions wi h highe and mo e a iable age a ma u a ion a e also known (G ege sen e al. 2011), and he spawning expe i- ence and age o a emale may ha e a signi ican e ec on he ep oduc i e success o ish (Kamle e al. 1982; E ans e al. 1996; Busch e al. 2009). We s udied he e ec o plas ici y in g ow h and ecun- di y o endace emales on hei ep oduc i e ai s using a se ies o long- e m incuba ion expe imen s. The wild pa - en al ish o igina ed om ou sepa a e popula ions wi h ma kedly di e en densi ies and hence lake-speci ic na u- ally induced di e ences in hei g ow h and ecundi y. We in e p e ed he e ha he la ge di e ences in g ow h o pa - en al ish we e mainly caused by hei pheno ypic plas ici y in esponse o he popula ion densi y, and al hough local adap a ions may e ol e in ish popula ions apidly (Eiza- gui e e al. 2012; Wes ley e al. 2012), hei ole in ou analysis is insigni ican . The ene gy alloca ion o soma ic issues and eggs p io o spawning ( a iable P S+R ) se ed as a p oxy o o al esou ce a ailabili y o an indi idual emale, and i s e ec s on o sp ing su i al and g ow h we e analyzed. Fi s ly, we examined whe he he P S+R di - e ed be ween he emales om he di e en lakes, he ela- i e p io i y o g ow h and eggs in he ene gy alloca ion by young and old spawne s, and how he P S+R a ec ed he egg p oduc ion ( ela i e ecundi y, egg size) o emales. Sec- ondly, di e ences in su i al and g ow h o o sp ing om young and old spawne s wi h di e en P S+R we e s udied using common ga den e iliza ion and incuba ion expe i- men s. Sho age in ma e nal ood esou ces has been obse ed o inc ease he egg size ia a ade-o be ween egg numbe and size and subsequen ly o a ec o sp ing size and su i al (Hu chings 1991). Finally, we explo ed whe he he alloca ion mechanisms e ealed by ou expe i- men s can explain he high a iabili y in he ec ui men o endace in ou s udy lakes. Ma e ials and Me hods S udy lakes and popula ions Ma u e endace o p o ide eggs and mil we e caugh om ou Finnish lakes: Pyh€ aj€ a i (I), Pyh€ aselk€ a (II), Puula esi (III), and Sou he n Konne esi (IV); he ea e he lakes a e e e ed o by hei oman nume als I–IV. The lakes we e selec ed o ep esen he wide ange o Figu e 1. Ou s udy species endace (Co egonus albula (L.)). Pho og aph by Jussi Mu osaa i. 780 ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Pheno ypic Plas ici y in G ow h o Female Fish J. Ka jalainen e al. endace popula ion densi ies in Finland: The densi y o he ma u e adul (age ≥1+) endace in 2004–2006 anged om abou 20 o 4000 indi iduals ha 1 in au umn, and he o al biomass o endace in au umn anged om 12 o 42 kgha 1 (Table 1). The densi y es ima es we e based on he emo al me hod (Helminen e al. 1993) and coho analysis (Pope 1972) in he in ensi ely ished Lake I and combined esul s o echo-sounding and CPUE in Lakes III and IV (Ma jom€ aki & Huolila 1995; Valkeaj€ a i and Ma jom€ aki 2013). Da a o adul ish om Lake II we e sca ce, and age-speci ic densi y es ima es we e no a ailable o he s udy yea s. Howe e , om o al ca ch s a is ics i s densi y was oughly es ima ed o be he sec- ond lowes . Fishing mo ali y di e ed be ween lakes being highes in Lake I (Table 1, Appendix S1), which is shal- lowe and mo e p oduc i e han he o he lakes and has a longe open-wa e pe iod han he o he lakes (Appendix S1). The annual popula ion ecundi y was calcula ed by mul iplying emale (sex a io assumed o be 1:1) spawn- ing s ock densi y (ind. ha 1 ) by he mean ecundi y (eggs emale 1 ) o ma u e emales in each lake. The densi ies o newly ha ched la ae o endace (ind. ha 1 ) in sp ing 2004–2007 we e es ima ed o all lakes using a s a i ied sampling design (U panen e al. 2009). Ene gy alloca ion by emale endace o soma ic g ow h and eggs To al ene gy alloca ed o soma ic g ow h and eggs (P S+R ) du ing he g owing season (kJg 1 o ish we mass) was es ima ed om andomly sampled ipe emales (eggs no unning). To al leng h (mm) and we mass (g) o hese emales we e measu ed. Bo h o a ies we e emo ed and weighed o he nea es 0.01 g. Th ee subsamples aken om di e en pa s o he gonads we e weighed, and he numbe o eggs pe subsample was coun ed unde a dis- sec ing mic oscope. Mean egg numbe o he subsamples was mul iplied by he a io o he o al sample mass o he mean subsample mass o ob ain an es ima e o abso- lu e ecundi y o each emale. Vendace is a de e mina e and o al spawne species: The whole clu ch o de eloped oocy es is laid down du ing a single pe iod wi h no u - he ec ui men o oocy es be o e he s a o spawning (Dlugosz and Wo niallo 1985; Mu ua e al. 2003). Thus, he oocy e numbe p io o spawning gi es an adequa e es ima e o he o al indi idual ecundi y (Kjesbu e al. 2010) which makes i possible o es ima e he o al ene gy alloca ed o ep oduc i e issue. In ou analysis, P S+R was assumed o e lec he le el o ma e nal ood esou ces (indi ec ly) and g ow h condi ions (di ec ly), and he in e annual a ia ion in popula ion densi y induced he a ia ion in P S+R. Rela i e ecundi y o emales was calcula ed by di iding absolu e numbe o eggs by o al mass o ish including gonads (W T , we mass). One subsample o un e ilized eggs om each emale was ozen a 80°C o he de e - mina ion o he ene gy con en o eggs by hei ca bon con en . Th ee pooled ( i e eggs) eplica es om each emale we e weighed and eeze-d ied and hen g ound o a ine powde using a mo a and pes le. App oxima ely 0.6 mg o powde ed sample was weighed in o a small in cup o he ca bon analysis. Analyses we e ca ied ou using a FlashEA 1112 elemen al analyze connec ed o a Table 1. Cha ac e is ics o endace popula ions [mean (SD)] in he s udy lakes in 2004–2007. We mass o emales a e i s yea in second sp ing was es ima ed om he scale by he Monas y sky’s equa ion (Monas y sky 1930). Lake I Lake II Lake III Lake IV SW Pyh€ aj€ a i Pyh€ aselk€ a Puula esi S Konne esi To al ca ch, kgha 1 13 (1.0) 1 (0.2) 4 (0.5) 3 (0.5) To al mo ali y, % >90 –50 50 To al densi y o spawne s, ind. ha 1 157 (116) –1072 (353) 2770 (834) To al biomass o spawne s, kgha 1 10 (6) –15 (4) 18 (5) Densi y 0+in au umn, ind. ha 1 1063 (265) –953 (715) 1541 (1610) Biomass 0+in au umn, kgha 1 20 (9) –4 (3) 4 (4) To al biomass in au umn, kgha 1 30 (10) –19 (5) 22 (9) We mass o 1-yea old emales, g 29 (8) 8 (3) 4 (1) 3 (2) We mass o no ice emales, g 78 (22) 25 (6) 11 (2) 9 (5) We mass o epea ed spawne s, g 127(48) 34 (10) 19 (5) 10 (3) To al leng h o all ma u e emales, mm 216 (24) 155 (18) 124 (14) 112 (23) Fecundi y o ma u e emales, eggs 9791 (3966) 3593 (1317) 1739 (714) 999 (333) % o ma u e indi iduals a age 1+99 (1) 98 (2) 100 100 % o no ice spawne s 84 (9) 58 (36) 40 (45) 45 (32) Densi y o ha ched la ae, ind. ha 1 5865 (4590) 1520 (1021) 8726 (7038) 22105 (26036) –=no da a a ailable. ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 781 J. Ka jalainen e al. Pheno ypic Plas ici y in G ow h o Female Fish The mo Finnigan DELTAplus Ad an age mass spec ome- e . Ca bon con en was ans o med o he ene gy con- en using a coe icien o 44.9 kJg 1 C (M. Paala uo, P. Muje & J. Ka jalainen, unpublished da a). Age o emale endace was de e mined om he scales and o oli hs (sagi ae), and he adius o each annulus in he scales was measu ed. Leng h a age 1, he beginning o he las g owing season, was back-calcula ed using Monas y sky’s equa ion (Monas y sky 1930), and he alue o he exponen (b=0.607) was de e mined om ou own da a (n=817). To al ene gy alloca ed o soma ic g ow h and eggs (P S+R, i ,kJg 1 ) was es ima ed indi idually o each emale (i) P S+R, i =(W T, , i  ^ W S, 1, i )*E i /Ẅ 1 , i whe e W T, , i =obse ed o al we mass a ime o emale i, ^ W S, 1, i =es ima ed soma ic we mass a he begin- ning o he las g owing season, a ime 1, E i = he g oss ene gy con en es ima e (kJg 1 )o emale i. Ẅ 1 , i =mean weigh o he emale idu ing he las g owing season. Es ima ed soma ic we mass a ime 1is ^ W S, 1, i =a*  L 1, ib *[W T, , i /(c*L , id )] whe e  L 1, i =back-calcula ed leng h o emale ia 1, L , i =i s obse ed leng h a , aand ba e he in e cep and he slope o log-linea eg ession o soma ic we mass (W S, ) on o al leng h (L , ) o emale endace (a=0.0000044, b=3.06, R 2 =0.98, n=461), espec i ely, and cand da e he in e cep and he slope o log-linea eg ession o o al we mass (W T, ) on o al leng h (L )o emale endace (c=0.0000051, d=3.09, R 2 =0.98, n=461), espec i ely. The mean mass o he emale idu ing he las g owing season is Ẅ 1 , i =(W T, , i + ^ W S, 1, i )/2 E i = he g oss ene gy con en (kJg 1 ) o emale iis based on i s Ẅ 1 , i . E i was es ima ed om log-linea eg ession be ween g oss ene gy con en and W T o age >0+ endace (R 2 =0.937, n=861, M. Paala uo, P. Muje & J. Ka jalainen, unpublished da a): E i =4.231 *Ẅ 1 , i0.112 Ene gy alloca ion o eggs (P R,i ,kJg 1 ) is es ima ed om P R,i =(F i *E a )/Ẅ 1 , i whe e F i =indi idual absolu e ecundi y and E a =mean ene gy con en o egg a gi en age g oup a, and was de e mined o each age g oup om e e y lake and yea om he ca bon con en o he eggs. Fe iliza ion and incuba ion expe imen s The expe imen s we e ca ied ou simila ly in all pe iods 2004/2005, 2005/2006, and 2006/2007. All ish we e caugh by local ishe men in he middle o he spawning season om he las week o Oc obe o he i s week o No em- be . Fish we e caugh wi h ap ne s (Lake I), gill ne s (Lake II), and seine ne s (Lakes III and IV) and be o e he e iliza ion ish we e kep a he same empe a u e as he wa e empe a u e in he spawning a ea (2–3°C). Only li e indi iduals om he ca ch wi h unning mil and eggs we e used o he e iliza ion expe imen . Annually in each popula ion, 9 eady- o-spawn emales and 27 males we e picked om a andom sample o ca. 100 ma u e ish (Appendix S2). Bo h young (1+) and old spawne s (>1+) we e included in he e iliza ions. L T and W T o each ish we e measu ed be o e emo ing he eggs and mil o he e iliza ion. Fe iliza ion was pe o med immedia ely a e ish we e caugh in he well-s ocked ield labo a o y es ab- lished in a an. F om each emale, h ee se s o a mini- mum o 100 eggs we e s ipped in o plas ic Pe i dishes. Each se was e ilized wi h a mix u e o mil om h ee males o ensu e su icien mil and o simula e he com- munal spawning in na u e. A e adding he mil , a small olume o lake wa e was added o he dishes o ac i a e he spe m and he dishes we e gen ly wa ed o sp ead he spe m uni o mly. The dishes we e placed in one-li e plas- ic bags wi h lake wa e a 2–5°C and anspo ed wi hin 2–4 h o a ha che y a he Konne esi Resea ch S a ion o he Uni e si y o Jy € askyl€ a, whe e he se s o eggs we e dis ibu ed in o incuba ion pla es in andomized o de and incuba ed in ac ylic plas ic pla es (5 95 cm) wi h gen le con inuous e ical wa e low h ough hei mesh bo om. Fe iliza ion success o eggs, ha is, he p opo ion o e ilized eggs o all eggs ( andom sample o 50–100 eggs pe se ), was de e mined 24–48 h a e e iliza ion unde a dissec ing mic oscope. Fe ilized eggs we e iden i ied by isible cell di ision. Fungal g ow h was con olled by weekly ea men wi h malachi e g een un il he emb yos eached he eyed s age. Dead eggs we e coun ed and emo ed weekly and wa e empe a u e eco ded daily h oughou he incuba ion pe iod. Mean (SD) wa e empe a u es du ing incuba ions we e 3.0 1.1, 3.2 1.0 and 2.7 0.7°C du ing he incuba ion pe iod in 2004/2005, 2005/2006, and 2006/2007, espec i ely. The o al du a ion o he incuba ion pe iod in di e en yea s, om e iliza ion o 90% ha ching, was 170–196, 164– 184, 171–194, and 171–195 days o eggs om Lakes I, II, III, and IV, espec i ely. The ins an aneous o al mo ali y (Z) o emb yos and la ae du ing expe imen s was calcula ed sepa a ely o each egg and ea ing se : 782 ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Pheno ypic Plas ici y in G ow h o Female Fish J. Ka jalainen e al. Z=ln (N e /N b ), whe e N e = he numbe o li e emb yos o la ae a he end o a pe iod and N b = hei numbe a he beginning o ha pe iod. A mean Z o each emale was calcula ed as a mean o he h ee eplica es (egg se s), and popula- ion-speci ic es ima es we e calcula ed as a mean o he nine emales. Mo ali y a es we e calcula ed sepa a ely o wo pe iods: Z F = e iliza ion ailu e and Z E = o al emb yonic mo ali y om e iliza ion o 100% ha ching (including Z F and Z L ). Values o p opo ional o al mo - ali y (A, %) we e calcula ed om A=1–exp(-Z). The p opo ion o ha ched la ae ali e (H%) was also eco ded. The mean egg we mass o each o he nine emales pe lake was measu ed om 50 eggs (Appendix S2). Du - ing he ha ching pe iod, he numbe o dead eggs and ha ched la ae was eco ded daily. One subse (n=10) o ha ched la ae om each emale was p ese ed in 70% e hanol o he measu emen s o L T and W T . La al g ow h, su i al, and swimming pe o mance expe imen s Ha ched la ae om he labo a o y incuba ions we e mo ed daily o low- h ough anks (26 li e s) and la ae om each lake and om 3 o 5 emales we e kep sepa a ely in he collec ing anks o a maximum o 3 days wi hou ood in o de o collec enough la ae o ea ing expe i- men s. Subsequen ly, newly ha ched la ae we e di ided in o ae a ed low- h ough aqua ia (leng h 40 cm 9wid h 25 cm 9heigh 20 cm), and 65 la ae pe aqua ium wi h 2–5 eplica es pe lake we e ea ed o 23 days o moni o hei g ow h and mo ali y. The pho ope iod du ing ea - ing expe imen s was se a 18-h ligh : 6-h da k, and he la - ae we e ed ad libi um on ha ched A emia nauplii wi hou cys s (daily a ion adjus ed o 100% o he W T o he la - ae). The aqua ia we e cleaned daily by siphoning ou eces and unea en A emia. The numbe o dead la ae and wa e empe a u e we e eco ded daily. The ins an aneous mo - ali y (Z L ) du ing he la al ea ing pe iod was calcula ed sepa a ely o each aqua ium, and he lake-speci ic Z L was calcula ed as a mean o aqua ia o a gi en popula ion. The mean (SD) ea ing empe a u es we e 12.9 0.4, 13.5 0.6 and 15.0 0.4°C in sp ing 2005, 2006, and 2007, espec i ely. A andom sample o i e la ae om each aqua ium was aken o a challenge es . Challenge es s we e pe o med in a low- h ough swimming chambe wi h a olume o 18 ml. The chambe was connec ed by ubbe hoses o a wa e ci cula ion pump, and he whole sys em was imme sed in an aqua ium. Wa e eloci y in he chambe was calib a ed sepa a ely in each ial acco ding o he a e age L T o la ae a a gi en ime. Fi e la ae we e placed in he swimming chambe , and wa e low was adjus ed o he le el o 1 L T s 1 o 1 min, a e which wa e eloci y was inc eased o 5 L T s 1 (SD =1). Swimming abili y o la ae agains low was eco ded as he ime (s) un il he median la a ( hi d o he i e) was exhaus ed and was unable o keep i s posi ion in he swimming unnel. A e ha , la ae we e emo ed om he chambe and anaes- he ized be o e measu emen o indi idual L T and W T . Da a analysis A gene al linea model (GLM) wi h a main e ec model s uc u e was used o es o he e ec o lake, s udy yea , and age o spawne s (young s. old spawne s) on he o al ene gy alloca ed o g ow h and eggs (P S+R , kJg 1 ) o o ep oduc ion (P R ,kJg 1 ) o emales in he summe be o e spawning. We body mass o he emales was used as a co a ia e in he models. GLM, wi h P S+R as a co a ia e, was used o es o he e ec o he spawning his o y on he ela i e ecundi y (eggs g 1 ) and we mass o eggs (mg) o emales. In hese analyses, he P S+R o indi idual ish om he s udy lakes was assumed o e lec he lake- and yea -speci ic s a us including he he mal e ec s, so lake and yea ac o s we e no longe included in he models. GLM, wi h he o al biomass o endace (kgha 1 )in summe be o e he spawning season as a co a ia e, was used o es o he e ec o he emale age (young s. old) on P S+R. The di e ence in he e iliza ion ailu e (Z F ), o al mo ali y o emb yos du ing he expe imen al incuba ion pe iod (Z E ), p opo ion o ha ched la ae ali e (H%), and L T and W T o he ha ched la ae we e compa ed be ween young and old spawne s by he GLMs wi h P S+R as a co a ia e. In la al expe imen s, he ea ing empe a u e di e ed be ween he s udy yea s due o p oblems wi h he hea ing sys em. Thus, he empe a u e e ec on he g ow h o he la ae was co ec ed by a empe a u e-speci ic g ow h unc ion de i ed om he expe imen al da a o Luczynski (1991). The L T and W T o la ae in each aqua ium we e adjus ed o co espond o g ow h a 13.5°C which was he mean ea ing empe a u e in 2006. Because he la ae o all emales om a lake we e pooled in he ea ing expe imen s (no possible o iden i y mo he s o indi id- ual la ae), he P S+R o he analysis was calcula ed as an annual mean o all emales in speci ic yea and lake. Da a we e di ided u he in o wo P S+R classes lowe and highe han he o e all mean o he emales in all yea s (o e all mean =4.6 kJg 1 ). The di e ences in empe a- u e-co ec ed L T and W T we e es ed by - es be ween he wo P S+R classes. GLM model wi h he ln- ans o med ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 783 J. Ka jalainen e al. Pheno ypic Plas ici y in G ow h o Female Fish ea ing empe a u e as a co a ia e was used o es he e ec s o P S+R on he Z L and swimming pe o mance o la ae a e he 23-days ea ing expe imen . In his analy- sis, P S+R class was a ixed ac o . Resul s Plas ici y in g ow h and ecundi y The la ge di e ences in he popula ion densi y be ween lakes caused high a ia ion in he mean size o bo h ju e- nile and ma u e ish, as well as in he ecundi y o emales. The in e annual a ia ion in he we mass o 0+ ish a e he i s g owing season was high, and e en a ou old di e ence be ween successi e yea s has been obse ed in Lake I (Fig. 2). The majo i y o endace emales spawned o he i s ime in hei second au umn (Table 1, % o ma u e indi iduals a age o 1+) and he p opo ion young spawne s a ied be ween 40 and 84% o all spawne s. Thus, e e y yea he e we e also olde and likely epea spawne s (age 2+o olde ) p esen in all lakes, bu hei p opo ion and o al numbe we e clea ly lowes in Lake I wi h he lowes popula ion densi y and highes adul mo ali y due o in ensi e ishing. To al ene gy (kJ) alloca ed by endace emales o soma ic g ow h and eggs di e ed conside ably be ween lakes (Fig. 3A) demons a ing he ex eme a iabili y o hei g ow h and ecundi y. Young spawne s (age g oup 1+) in es ed signi ican ly mo e (F=659.6, P<0.001, Fig. 3B) ene gy in o g ow h and eggs (P S+R ,kJg 1 ) han old spawne s (age g oup ≥2+). The P S+R also di e ed signi ican ly be ween lakes (F=7.2, P<0.001), being la - ges in he spa ses popula ion wi h ex emely as -g ow- ing indi iduals in Lake I, and also be ween yea s (F=28.1, P<0.001). The e ec o he co a ia e we mass was no signi ican (F=0.01, P>0.05). The ene gy alloca ed o eggs (P R ,kJg 1 ) also di e ed signi ican ly be ween young and old spawne s (F=25.3, P<0.001). The P R o young spawne s was highe han ha o old spawne s (Fig. 3B), bu old spawne s alloca ed a highe p opo ion (69%) o o al ene gy o eggs han young emales (49%), and hus, a smalle sha e emained o soma ic g ow h. Al oge he , he a ia ion in P R was e y low compa ed o ha o P S .The P R di e ed signi i- can ly be ween yea s (F=3.2, P<0.05) bu did no di - e be ween lakes (F=1.0, P>0.05). Again, he e ec o he co a ia e mass was no signi ican (F=1.1, P>0.05). The la ge di e ences in he age-speci ic size o emales be ween lakes yielded la ge be ween-lakes di e ences in hei absolu e ecundi y (eggs emale 1 , Table 1). A he same size, he old spawne s had on a e age 10% lowe Figu e 2. In e annual a ia ion in he mean we mass (g) o 0+ endace in 1973–2013 (Helminen e al. 1992) in Lake I and Lake IV (P. Valkeaj€ a i, unpublished da a). Figu e 3. (A) To al ene gy alloca ion o soma ic g ow h (kJ) and eggs (kJ) by young (1) and old (2) spawne s in Lakes I–IV. (B) Rela i e ene gy alloca ion o soma ic g ow h (P S, kJg 1 o emale we mass) and egg p oduc ion (P R ,kJg 1 o emale we mass) by young and old spawne s in Lakes I-IV. The means o e all lakes and yea s o o al ene gy (P S+R ,kJg 1 o emale we mass) used o g ow h and egg p oduc ion du ing g owing season and o mean ene gy used o eggs (P R ) a e shown by dashed and solid lines, espec i ely. 784 ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Pheno ypic Plas ici y in G ow h o Female Fish J. Ka jalainen e al. absolu e ecundi y han he young spawne s. We mass o eggs (mg) and ela i e ecundi y (eggs g 1 emale we mass) di e ed signi ican ly be ween hese spawning his- o y classes (F=3.9, P<0.001; and F=25.9, P<0.001, Fig. 4). The egg mass o old spawne s was highe bu hei ela i e ecundi y was lowe han hose o young spawne s. In he GLM model, he co a ia e P S+R signi i- can ly a ec ed he egg we mass (F=11.5, P<0.01) and ela i e ecundi y (F=3.6, P=0.06) o emales: In bo h young and old spawne s, highe P S+R (and hus mo e ene gy a ailable) was associa ed wi h highe egg mass bu lowe ela i e ecundi y. E ec s o P s+ on e iliza ion ailu e and egg mo ali y P S+R had no e ec on he e iliza ion ailu e Z F (F=0.15, P>0.05) o he o al emb yonic mo ali y om e iliza ion o 100% ha ching Z T (F=0.15, P>0.05). Nei he Z F (F=0.15, P>0.05) no Z T (F=0.7, P>0.05) di e ed signi ican ly be ween young and old spawne s. In he incuba ion expe imen s, Z T was gene ally high and a iable be ween yea s and popula- ions; he ela i e o al mo ali ies (A%, mean SD) we e 82 9%, 53 23%, 25 16%, and 76 36% in Lakes I, II, III, and IV, espec i ely. Fe iliza ion ailu e cons i u ed mos o he o al mo ali y. E ec s o P s+ on la al g ow h, mo ali y, and swimming pe o mance P S+R had a signi ican posi i e e ec on he o al leng h (L T ) o ha ched la ae (F=4.3, P<0.05) bu no on he p opo ion o ha ched li e la ae ali e (H %, F=2.7, P>0.05). L T o newly ha ched la ae (F=5.6, P<0.05) and H % (F=7.5, P<0.05) di e ed signi ican ly be ween young and old spawne s. The mean L T was 8.5 mm (SD =0.7, n=31) and 8.7 mm (SD =0.4, n=48) o he young and old emales, espec i ely. The mean H % was 93% (SD =15, n=38) and 98% (SD =7, n=59) o he young and old emales, espec i ely. A e he 23-days ea ing expe imen , he la al pe iod mo ali y Z L (Fig. 5A) di e ed be ween he low and high P S+R classes (F=20.3, P<0.001) and he empe a u e co a ia e was also signi ican (F=13.8, P<0.01). Mo - ali y in he low P S+R class was highe han in he high P S+R class. The swimming pe o mance (Fig. 5B) o he la ae did no di e signi ican ly be ween P S+R classes (F=1.3, P>0.05), and he e ec o empe a u e co a i- a e was insigni ican (F=0.5, P>0.05). The e was no signi ican di e ence be ween P S+R classes (Fig. 5C, D) in la al L T o W T ( =1.75, P>0.05; and =0.24, P>0.05, espec i ely). Implica ions o popula ion le el In he s udy yea s, he es ima es o mean densi y o spawne s, egg p oduc ion, and la al p oduc ion di e ed conside ably be ween he lakes (Fig. 6A–C): The o sp ing p oduc ion was highes in Lake IV wi h he highes o al Figu e 4. (A) Rela ionship be ween egg we mass (g) and o al ene gy alloca ed o g ow h and egg p oduc ion (P S+R ,kJg 1 o emale we mass) du ing he g owing season. Reg ession lines (y=a+bx) o young (a=1.81 0.12 (SE), b=0.039 0.021) and old spawne s (a=1.83 0.13, b=0.11 0.04) a e gi en sepa a ely. (B) Rela ionship be ween ela i e ecundi y o emales (egg g 1 emale we mass) and P S+R . Reg ession lines o young (a=130.3 11.1 (SE), b=2.53 1.93) and old spawne s (a=121.7 8.52, b=3.31 2.39) a e gi en sepa a ely. Figu e 5. (A) Mo ali y (Z L ), (B) swimming pe o mance (seconds), (C) o al leng h (L T , mm), and (D) we mass (W T , mg) o endace la ae a he end o he 23-day ea ing expe imen s in sp ing 2005, 2006, and 2007. Ba s indica e he mean alues o la ae ha ched om he eggs o emales wi h low o high P S+R . E o ba s ep esen he s anda d e o o he mean, and ***indica es s a is ically signi ican di e ence be ween he low and high P S+R . ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 785 J. Ka jalainen e al. Pheno ypic Plas ici y in G ow h o Female Fish densi y es ima e, while he lowes egg and la al p oduc- ion was obse ed in Lake I wi h he lowes popula ion densi y and he highes annual mo ali y. Ins ead, he mean annual densi y o 0+ ish in au umn showed low a ia ion be ween he lakes (Fig. 6D), and he o al annual biomass o he popula ion showed li le a ia ion be ween he lakes o e en an opposi e end compa ed o he densi y, wi h Lake I ha ing he highes mean biomass among he s udy lakes (Fig. 6E). Discussion In ou s udy lakes, endace emales alloca ed a a he con- s an and high p opo ion o hei ene gy in es men s o eggs (kJg 1 ) despi e e y di e en g ow h pa e ns depending on he esou ces a ailable in he di e en lakes (e.g., ange o mean W T and ecundi y a age o i s ep o- duc ion om 5 o 132 g and om 530 o 21,000 eggs e- male 1 , espec i ely). The in es men s in o eggs hus dic a ed he sha e emaining o g ow h. Egg p oduc ion by la ge emales and old spawne s ook a la ge p opo ion o hei o al ene gy alloca ed o g ow h and eggs (P S+R ) han he egg in es men s by smalle and young spawne s, which led o a end o dec easing g ow h a e a highe size and age. Such educed g ow h a e ma u a ion is a phe- nomenon widely obse ed among ishes (Ro 1983; Folk- o d e al. 2014), and al hough some ish species wi h limi ed ood supply may pa ly o ully sac i ice egg p o- duc ion o p ese e body condi ion, mos main ain in es - men in ep oduc ion a he expense o body ese es (Sabo ido-Rey and Kjesbu 2005). The ene gy alloca ion o eggs pe body mass was highe in young han in old spawne s, and he egg size and con- sequen ly he ela i e ecundi y also di e ed, wi h young emales p oducing mo e bu smalle eggs so ha he newly ha ched la ae o young emales we e smalle han he la ae o old spawne s. The ecognized high alue o la ge ma u e ish (high ecundi y, la ge eggs, highe su - i al o o sp ing) o he popula ions o se e al species has led o he ecommenda ion in hei ishe ies manage- men o conse e la ge, ypically old indi iduals (Dua e and Alca az 1989; Be keley e al. 2004; Sabo ido-Rey and Kjesbu 2005; Enbe g e al. 2011). Howe e , in ou en- dace expe imen s, he e iliza ion a e and he su i al o emb yos we e no a ec ed by he age o spawne s o he o al ene gy alloca ion o he eggs. This con as s wi h he esul s o ea lie expe imen s by Kamle e al. (1982) who epo ed ha endace emales spawning o he i s ime p oduced he “wo s ” eggs, middle-aged emales p oduced he bes eggs, and old spawne s laid eggs wi h again lowe quali y. The emb yonic mo ali y in ou expe imen s was highes du ing ea ly emb yogenesis un il he eyed s age, which has also been epo ed in p e ious s udies (Wilkonska and Zu omska 1982; Kamle 2005; Ka - jalainen e al. 2014). Ou e iliza ion p ocedu e simula ed he communal spawning whe eby se e al males elease hei spe m simul aneously du ing spawning, and he e- o e, he pa e nal e ec s on emb yonic su i al emain un esol ed. In many ish species, he impo an ep oduc i e ai s o age and size a ma u a ion a y wi h changes in popu- la ion densi y (Adams 1980; T ippel 1995; Sabo ido-Rey and Kjesbu 2005; Roos e al. 2006). As he age a ma u a- ion o endace in Finnish lakes is p ac ically ixed o hei second au umn (age g oup 1+yea s), he size a ma u a ion a ies ma kedly wi h he indi idual g ow h a e o ish, which again depends on popula ion densi y. In ou s udy lakes, luc ua ion be ween high and low s ock densi ies a e ypical and caused by in e annual a i- a ion in en i onmen al condi ions (Ka jalainen e al. 2000; Ma jom€ aki 2003) as well as in e s age e ec s o en- dace popula ions (Ma jom€ aki e al. 2014). In some Fin- nish lakes, in ensi e ishing o he endace popula ion app eciably shapes he spawning s ock s uc u e (Sa ala and Helminen 1995; Huusko and Hy € a inen 2005) and ep oduc i e ai s o he species. Especially in Lake I, a spa se popula ion ollowing high ishing mo ali y esponds o high esou ce a ailabili y by high g ow h a e p io o spawning and subsequen high indi idual ecun- di y. A simila esponse o g ow h (Fig. 2) and ecundi y has been seen in many o he lakes, including Lake IV whe e in he 1990s a long and p obably en i onmen al- Figu e 6. Densi y o spawning s ock (A, ma u e indi iduals ha 1 ), popula ion ecundi y (B, eggs ha 1 ), la al densi y (C, la ae ha 1 ), densi y o 0+ endace in au umn (D, indi iduals ha 1 ), and o al biomass o endace in au umn (E, biomass o 0+and ma u e ish kgha 1 ) in he s udy lakes in 2003–2007. Ba s ep esen annual means and e ical lines s anda d de ia ions. 786 ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Pheno ypic Plas ici y in G ow h o Female Fish J. Ka jalainen e al. induced phase o low endace s ock p e ailed (Valkeaj€ a i and Ma jom€ aki 2013). In e es ingly, he ela i e ecundi y o endace emales in he spa se popula ion was lowe and egg size la ge han hose o ish g owing in he dense s ock wi h e y much lowe ood a io pe capi a. This esponse o egg size o ood sca ci y was opposi e o he mechanisms epo ed among pe iodic salmonids, which end o p o- duce la ge eggs i emales unde go ood sca ci y p io o spawning (Tho pe e al. 1984; Hu chings 1991; Bu on e al. 2013). This ade-o o in es ing in highe quali y, la ge la ae while educing hei quan i y unde ood limi a ion has been conside ed a bene icial adap a ion in low esou ce en i onmen al condi ions (Hu chings 1991; G ege sen e al. 2011; V ilek and Reicha d 2014). How- e e , s udies ha e e ealed ha he exp ession o ma e nal e ec s is con ex -dependen and es ic ed ma e nal ood a ailabili y has been shown o ei he dec ease (Gagliano and McCo mick 2007) o inc ease (Hu chings 1991; Gui- sande e al. 1996; Allen e al. 2008) o sp ing size and su - i al. Plais ow and Ben on (2009) concluded ha ma e nal e ec s in high compe i ion popula ion (low ood a ion) impac mo e on ju enile su i al compa ed wi h low compe i ion popula ions (high ood a ion), whe e ma e nal e ec s ha e mo e impac on he popula- ion g ow h a e. In addi ion o he compensa o y densi y-dependen inc ease in p e ec ui men su i al o endace along wi h he dec ease in spawning s ock (Ma jom€ aki 2003), excep- ionally high compensa o y plas ici y in g ow h and sub- sequen changes in ecundi y and egg size inc ease he capaci y o endace popula ions o wi hs and high ishing mo ali y. In a spa se s ock, emales p oduce pe capi a mo e and la ge eggs and la ae wi h highe su i al: In Lake I, he spawning s ock densi y was much lowe han in he o he lakes, bu he es ima e o su i al om newly ha ched la ae o 0+ ish in au umn was as high as 32 30% o lake I (mean SD) compa ed o 11 7% o Lakes III and IV. Mo eo e , he ex emely high g ow h a e o ish in he Lake I p io o spawning esul ed in o al biomass o endace in au umn as high as in he o he lakes (Fig. 6E). The longe g owing season and highe p oduc i i y in Lake I han in he o he lakes also con ibu e o he high g ow h and egg p oduc ion o endace emales, al hough he me a-analysis by Viljanen (1986) showed ha di e ences in he g ow h a e o en- dace popula ions be ween sou he n and no he n lakes was much smalle han he in e annual di e ences in g ow h o ish in same lake and e en in successi e yea s. Size-selec i e ishing is commonly in e p e ed as one o he majo causes behind obse ed changes in age a ma u- a ion by a ec ing g ow h a es and inducing e olu ion- a y egime shi s (Sabo ido-Rey and Kjesbu 2005; Roos e al. 2006). In endace popula ions, ishing is e lec ed in o he ep oduc i e ai s, and in ensi e ha es ing has also been obse ed o simpli y he spawning popula ion s uc- u e causing changes in he popula ion dynamics (Huusko and Hy € a inen 2005). Al hough he la ge in e annual di e ences in g ow h o pa en al ish wi hin a lake we e caused by hei ex eme pheno ypic plas ici y in esponse o he changes in he popula ion densi y, we canno o ally o e ule he e ec o local adap a ion, which has been sugges ed o be com- mon in salmonids (Taylo 1991; P imme 2011). Local adap a ion o ish can be apid, p og essi e p ocess d i en by en i onmen -induced selec ion p ocess (Eizagui e e al. 2012; Wes ley e al. 2012). Plas ici y i sel “is he i a- ble and popula ion di e ences in eac ion no ms can e lec adap i e esponses, by na u al selec ion, o local en i onmen s” (Hu chings 2011), and hus, locally adap ed popula ions may a y in hei plas ici y. Ou s udy species has high in e annual a ia ion in g ow h ha ing la ge di e ences in mass o 0+ ish e en be ween successi e yea s, and he po en ial o plas ici y seems o be simila in he s udy lakes (see Fig. 2). Also s ock ans- e s o endace by ishe ies manage s om lakes wi h dense popula ion and slow-g owing indi iduals o ano he lake wi h spa se popula ion has caused ema k- able changes in g ow h a e o ans e ed ish ma ching wi h he ampli ude o a iabili y in ou s udy lakes (Huuskonen, H., Ma jom€ aki, T. J., unpublished esul s). In he incuba ion expe imen s, he emb yos o all pop- ula ions de eloped a simila a es and hei ha ching suc- cess was high. The newly ha ched la ae o old spawne s we e sligh ly la ge han hose o young spawne s. Acco ding o he concep o “ apid g ow h leads o educed suscep ibili y o size-selec i e mo ali y and enables high su i al and high i ness” (Mille e al. 1988) e en small di e ences in he ha ching size may signi i- can ly a ec he u u e su i al. Al hough U panen e al. (2005) obse ed no clea size-dependen mo ali y du ing he ea ly li e o endace in Finnish lakes, Au inen (1995) and Sa ala and Helminen (1995) ha e epo ed ha he yea -class s eng h o endace in hei s udy lakes seemed o be posi i ely associa ed wi h egg size. In ou ea ing expe imen s, he mean lake-speci ic ins an aneous mo al- i y a e (Z L ) anged om 0.10 o 0.34 and mo ali y did no di e among popula ions bu was signi ican ly di e - en be ween he o sp ing o emales wi h low and high P S+R . Howe e , he g ow h and swimming pe o mance o he la ae unde expe imen al condi ions did no di e be ween P S+R classes. Al hough ou expe imen al and ield da a basically suppo he asymme ic ood compe i ion hypo hesis o Ham in and Pe sson (1986) as a mechanism explaining he 2-yea cyclici y o endace popula ions, he di e ences in la al su i al and p ope ies a e a oo ª2016 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 787 J. Ka jalainen e al. Pheno ypic Plas ici y in G ow h o Female Fish