Recording strategies and selection potential of feed intake measured using the X-ray method in rainbow trout

Gene . Sel. E ol. 38 (2006) 389–409 389 c INRA, EDP Sciences, 2006 DOI: 10.1051/gse:2006011 O iginal a icle Reco ding s a egies and selec ion po en ial o eed in ake measu ed using he X- ay me hod in ainbow ou An i Ka∗, Declan Tb, Alexand e Db∗∗, Dominic Hb,SamMb,EsaA.M ¨ a, Ossi Rc,Ka iRd aMTT Ag i ood Resea ch Finland, Animal P oduc ion Resea ch, Animal B eeding, FIN-31600 Jokioinen, Finland bSchool o Biological Sciences, Depa men o Zoology, Uni e si y o Abe deen, Abe deen, AB24 2TZ, Uni ed Kingdom cFinnish Game and Fishe ies Resea ch Ins i u e, Te o Fishe ies Resea ch and Aquacul u e, FIN-72210 Te o, Finland dFinnish Game and Fishe ies Resea ch Ins i u e, Tu ku Game and Fishe ies Resea ch, FIN-20520 Tu ku, Finland (Recei ed 4 No embe 2005; accep ed 27 Ma ch 2006) Abs ac – This s udy examines he way long- e m eed in ake should be eco ded accu a ely o selec i e b eeding pu poses, and es ima es selec ion po en ial in eed in ake using he X- ay me hod o eco d indi idual daily eed in ake in ainbow ou (Onco hynchus mykiss). The analysis showed ha he poin es ima es o daily eed in ake displayed low epea abili- ies ( =0.09−0.32). This indica es ha a minimum o h ee epea ed eco ds we e needed o accu a ely eco d a e age eed in ake a a ixed age. To effec i ely b eed o eed in ake o e he whole g owing pe iod, i is necessa y o de e mine a e age eed in ake a diffe en ages, since he e we e only mode a e pheno ypic and gene ic co ela ions be ween a e age daily eed in ake eco ded a 140 g, 750 g and 2000 g we mass. He i abili y o a e age daily eed in ake was low (a e age h2=0.10), indica ing ha modes gene ic changes can be ob ained in esponse o selec ion. I was concluded ha selec ion o gene ically change long- e m eed in ake can be success ul, ye epea ed obse a ions a se e al li e s ages a e needed o ensu e he accu acy o eed in ake es ima es and he efficiency o selec ion. eed efficiency /X- adiog aphy /he i abili y / epea abili y /salmonids ∗Co esponding au ho : [email p o ec ed] ∗∗ Cu en add ess: Uni o Social Ecology, Uni e si é Lib e de B uxelles, CP 231, Bd du T iomphe, 1050, Belgium. A icle published by EDP Sciences and a ailable a h p://www.edpsciences.o g/gse o h p://dx.doi.o g/10.1051/gse:2006011 390 A. Kause e al. 1. INTRODUCTION Feed is one o he majo cos s in a m animal p oduc ion. In aquacul u e, imp o ing eed efficiency, a a io o we mass gain o eed in ake, would ha e he effec o educing eed cos s and minimising en i onmen al loading. Selec- i e b eeding is a po en ial ool o imp o ing eed efficiency. Feed efficiency can be imp o ed by simul aneous selec ion o apid g ow h and agains eed in ake, o , when g ow h a e and eed efficiency a e a ou ably gene ically co - ela ed, by sole selec ion o inc eased g ow h a e [21]. Since eed in ake is a undamen al componen o eed efficiency, eliable means o eco ding eed in ake a e equi ed. This s udy examined he way long- e m eed in ake should be eco ded in ainbow ou (Onco hynchus mykiss Walbaum) o selec i e b eeding pu poses, and es ima ed selec ion po en ial in eed in ake. In o de o be able o es ima e he gene ic po en ial o selec ion, indi idual eed in ake needs o be eco ded in a amily-s uc u ed popula ion. Fo h ee easons, an accu a e eco ding o indi idual eed in ake o e he whole g ow- ing pe iod is challenging in g oups o ish. Fi s , eed in ake o indi iduals a ies g ea ly om day- o-day in salmonids e.g. [15, 28]. The obse ed high a iabili y educes he abili y o p edic long- e m eed in ake o an indi idual using a single daily eco d [16]. Because i is known ha measu emen accu- acy can be inc eased by eco ding he same indi iduals epea edly [9], i is o in e es o quan i y he numbe o epea ed daily eed in ake eco ds needed a each ixed age poin o accu a ely es ima e a e age eed in ake. Second, indi- idual anking ac oss ages may change. Fo b eeding pu poses, i is o in e es o know whe he o no ea ly eed in ake measu emen s can be used as eli- able gene ic p edic o s o e he whole g owing pe iod. I would be emp ing o measu e eed in ake om young ish because eed in ake is especially la- bo ious o eco d om la ge ish bu alida ion o his me hod is equi ed. So a , es ima es o gene ic co ela ions be ween eed in ake eco ded a diffe - en s ages o he li e cycle ha e no been a ailable o any ish species. Thi d, eco ding indi idual eed in ake on a la ge numbe o ish held in a common ank is echnically difficul . A poin es ima e o indi idual eed in ake om a la ge numbe o ish can be eco ded by he X- ay me hod [14,16,17,30]. To measu e eed in ake using he X- ay me hod, all ish held in a ank a e i s ed wi h eed con aining small glass beads con aining lead oxide. The ea e , he ish a e X- ayed. The num- be o glass beads consumed can be coun ed om he X- ay ilms. Because he glass bead con en o he eed is known, i is possible o calcula e he quan- i y o eed ha each ish consumed [16, 30]. The ques ion emains, howe e , Feed in ake in ish b eeding 391 whe he he X- ay me hod would suffice o selec ion pu poses ha a e aimed a changing long- e m eed in ake las ing o e he whole on-g owing pe iod. The X- ay me hod was applied he e o a la ge numbe o pedig eed ainbow ou Onco hynchus mykiss (Walbaum) o examine he ollowing ou opics. Fi s , i was assessed how many daily eed in ake measu emen s a e needed o accu a ely es ima e a e age eed in ake o an indi idual a a ixed age. Second, i was es ed whe he eed in ake displays non-ze o he i abili ies ha would al- low i s gene ic imp o emen h ough selec ion. Thi d, o s udy whe he eco d- ing o eed in ake a a single age can be used as a p edic o o e he whole g owing pe iod up o 2 kg ish, i was assessed whe he a e age eed in ake measu ed a diffe en ages co ela e posi i ely. Finally, he ish in his s udy we e ed wi h wo diffe en die s, ha ing diffe en p o ein and lipid con en , o examine he consis ency o pheno ypic and gene ic pa ame e s ac oss dis inc nu i ional en i onmen s. Feed in ake o salmonids changes ac oss nu i ional en i onmen s [7], and in esponse o changes in body composi ion and mass, e.g.in he o m o compensa o y eeding [15]. Hence, i is possible ha phe- no ypic and gene ic pa ame e s o eed in ake diffe on al e na i e die s. The wo die s, a no mal p o ein die (NP) ep esen ing mode n eeds, and an ex- pe imen al high p o ein die (HP), we e es ablished o speci ically assess he in luence o die s on pheno ypic and gene ic (co) a ia ion o composi ion ai s (de ailed by Tobin e al., in p ep.) bu he e he same expe imen al se up was applied o eed in ake. 2. MATERIALS AND METHODS 2.1. Popula ion s uc u e The ish used o igina ed om he Finnish na ional b eeding p og amme (The Selec), ca ied ou a he Te o Fishe ies Resea ch and Aquacul u e s a- ion (a eshwa e s a ion) o he Finnish Game and Fishe ies Resea ch Ins i- u e. The ini ial popula ion was es ablished in he la e 1980s by c ossing ou high quali y s ains. The b oods ock managemen and selec ion p ocedu es a e desc ibed by Kause e al. [20]. The ish o he p esen s udy we e kep in eshwa e a he Te o s a ion du ing he whole expe imen (June 2001 un il No embe 2004). In o de o s udy he po en ial o selec i e b eeding o eed in ake, he 2001 gene a ion was exposed o wo diffe en die ea men s in a spli - amily design. In Ap il 2001, a o al o 89 si es and 109 dams we e ma ed in a ac o ial design o p oduce 210 ull/hal -sib amily g oups. Each si e was ma ed o an a e age 392 A. Kause e al. Table I. Popula ion s uc u e and a e age amily size o a o al o 210 amilies sam- pled a h ee eco ding imes. 165 amilies 45 amilies Numbe o si es and dams 81, 99 34, 40 Family size, ime 1 (1s X-Ray) 6.6 30.9 Family size, ime 2 (2nd X-Ray) 5.4 26.6 Family size, ime 3 (3 d X-Ray) 4.4 13.4 o 2.3 dams ( ange: 1–5) and each dam o 1.9 si es ( ange: 1–3). The ma ings las ed h ee days. The pedig ee o e e y ish was known om he gene a ions bo n in 1998, 1995, 1992 and 1989. The eggs we e incuba ed in he same incuba o , bu amilies we e kep sep- a a ed. The eye-s aged eggs we e ans e ed o indoo 150 L amily anks in June 2001. Each o he 210 amily was held in hei own ank. In Feb ua y 2002, a e 8 mon hs o g owing in he amily anks, 2931 inge lings we e emo ed om he anks and indi idually agged wi h PIT- ansponde s (T o an L d, Ge many) o enable indi idual iden i ica ion. Fo 165 o he 210 amilies, an a e age o 7.0 ish pe amily we e agged ( ange 4–7 ish). F om he emaining 45 amilies, an a e age o 39.6 ish pe amily ( ange 37–40 ish) we e agged. The la ge ini ial amily size was chosen because hese amilies we e des uc i ely sampled o muscle and body com- posi ion o o he pu poses han his s udy. The change in he a e age amily size ac oss he h ee sampling imes is shown in Table I. Be o e he ini ia ion o he die ea men s, all ish we e ed wi h comme cial ainbow ou d y eed (Nu a S a e and Nu a Pa , Rehu aisio Inc., Finland). 2.2. Fish managemen Du ing agging, ish om each amily we e spli in o wo g oups o be ea ed on diffe en die s. In May 2002 (week 20) wo die a y ea men s we e ini ia ed. The die s we e a mode n die wi h no mal p o ein (40–45%) and high lipid con- en (30–33%) (NP die ), and an al e na i e die wi h high p o ein (50–56%) and low lipid con en (15–24%) (HP die ) (Tab. II). The die s consis ed o ish meal, ish oil, whea meal and whea s a ch, and we e supplemen ed o mine - als and i amins acco ding o he NRC, Na ional Resea ch Council [31]. The expe imen al HP die was designed so ha he capaci y o inefficien ish o deposi diges ed p o ein as p o ein g ow h would be g ea ly exceeded, leading o hei slowe g ow h and highe lipid deposi ion compa ed o he efficiency o indi iduals. This was hypo hesised o lead o an enhanced selec ion po en ial Feed in ake in ish b eeding 393 Table II. P oxima e composi ionao no mal p o ein (NP) and high p o ein die s (HP), including glass-bead labelled X- ay die s ed a measu emen imes 1 (pelle size 3 mm), 2 (size 6 mm) and 3 (size 7 mm). Feed, pelle size Wa e (%) Ash (%) CL (%) CF (%) CP (%) NFE (%) P (g ·Kg −1)Ene gy(KJ·g−1) NP, 3 mm 2.1 7.4 30.5 0.73 44.9 15.1 12.1 23.6 NP, 6 mm 1.5 7.7 30.3 0.72 44.6 15.9 12.7 26.0 NP, 7 mm 4.0 6.7 33.4 1.00 39.5 15.4 9.60 25.9 HP, 3 mm 2.3 9.6 20.7 0.44 56.4 11.1 15.5 25.6 HP, 6 mm 1.7 9.5 20.6 0.48 56.3 11.9 15.4 24.0 HP, 7 mm 7.2 8.3 23.8 1.30 49.4 10.0 11.8 23.3 NP, 3 mm (labelled) 6.9 7.5 29.4 0.74 41.1 15.1 11.3 24.3 NP, 6 mm (labelled) 5.6 7.4 31.4 0.72 40.1 15.5 11.2 24.9 NP, 7 mm (labelled) 3.8 6.3 34.3 1.00 40.5 14.1 10.4 25.6 HP, 3 mm (labelled) 7.6 10.4 15.4 0.62 54.2 12.4 15.2 21.1 HP, 6 mm (labelled) 9.4 10.6 14.1 0.55 54.1 11.8 15.1 20.7 HP, 7 mm (labelled) 3.9 9.1 20.6 0.90 55.8 9.70 13.8 22.4 aCL =c ude lipid; CF =c ude ib e; CP =c ude p o ein; NFE =ni ogen- ee ex ac s; P =phospho us. 394 A. Kause e al. on he HP die , in e ms o inc eased he i abili ies o lipid body composi ion and mo e a ou able co ela ions be ween g ow h and lipid deposi ion (Tobin e al., in p ep.). The ish spli be ween he wo die s we e o simila mass a agging (mean ± SD: NP =62.4 ±19.9 g, n=1355 and HP =62.3 ±19.4 g, n=1335). Each die g oup was placed in ou eplica e ib eglass es anks and housed indoo s. Fish we e kep in 3 m3 anks un il week 24 and ans e ed o 20 m3 anks he ea e . The amilies we e equally dis ibu ed among he anks. Fish densi y in each ank emained unde 20 kg ·m−3. Feeding was au oma ed using compu e -con olled pneuma ic eede s (A o-Tec Inc., Finland). Daily eeding a ion was based on con inuous em- pe a u e measu emen s and he g ow h o mula o ainbow ou based on an op imum g ow h empe a u e o 16 ◦C. Sligh o e eeding wi h 10–20% eed gi en o e he ecommenda ions was applied du ing he expe imen o p e en a es ic ion o g ow h po en ial o he ish. Wa e empe a u e du - ing he expe imen was ambien and exposed o seasonal luc ua ions ( ange = 0.4−22.9◦C). Oxygen concen a ion was kep o e 6.0 mg ·L−1and a cons an pho ope iod o 16L:8 D was used. 2.3. Reco ding ai s Daily eed in ake and body mass we e eco ded on h ee occasions; May 2002 ( ime 1, a 140 g), Oc obe 2002 ( ime 2, a 750 g), and Sep embe 2003 ( ime 3, a 2000 g). The iming o a ai eco ding is indica ed by a subsc ip o he ai name. The esul s o body mass a e b ie ly summa ised he e o compa e he eed in ake esul s wi h a ai o which beha iou is al eady well unde s ood and documen ed [19,20]. Each eco ding ime, 1, 2 and 3, consis ed o a 3-week session wi h h ee epea ed weekly measu emen s o bo h body mass and daily eed in ake. Du - ing a 3-week session, each ank was measu ed once a week, measu ing wo anks pe day in a consis en o de each week. Due o he measu emen con- s ain s o eed in ake ials, he es anks we e di ided in o wo eeding g oups ( wo NP and wo HP anks pe g oup), one g oup being ed om 0400 o 0800 h and he o he g oup om 0800 o 1200 h. On successi e days he eco ding o de o NP and HP anks was e e sed, o a oid he effec s o sys- ema ic eeding hy hms. In o he wo ds, on Tuesday mo nings one NP and one HP ank we e measu ed successi ely, on Wednesday mo nings one HP and one NP ank we e measu ed successi ely, e c. To ini ia e a session, all ish we e weighed du ing he i s week o each session, and p ede e mined andomly Feed in ake in ish b eeding 395 chosen ish om each amily we e eco ded o eed in ake using he X- ay me hod. Du ing he second and he hi d weeks, only he ish X- ayed du ing he i s week we e weighed and X- ayed again. The X- ays we e pe o med in he same way o all anks and du ing all sessions. Indi idual eed in ake was eco ded by X- adiog aphy using a po able X- ay uni (Todd Resea ch 80/20, UK), as desc ibed by Talbo and Higgins [38] and McCa hy e al. [30]. P io o X- ay, all ish om a gi en ank we e ed as usual bu he die was labelled wi h adio-opaque ballo ini glass beads (Jencons Scien i ic L d, Leigh on Buzza d, UK) (Tab. II). The la- belled pelle s used a imes 1, 2 and 3 consis ed o 1%, 0.5% and 0.3% beads wi h a diame e o 400–600 µm, espec i ely (i.e.43–45 beads ·g−1o d y eed a ime 1, 12–17 beads ·g−1o d y eed a ime 2, 7–12 beads ·g−1o d y eed a ime 3). A minimum o 2 h a e he eede s had s opped p o id- ing eed, he ish o be eco ded we e se ially placed in o anaes he ic solu ion (buffe ed MS-222), weighed and X- ayed. A e X- aying, he ish we e e- u ned o hei espec i e anks. The ea e , he beads we e coun ed om he ilms, and he mass o eed wi hin a s omach was es ima ed using a calib a ion eg ession equa ion. A calib a ion line was cons uc ed o each session and each die sepa a ely. This was done by X- aying diffe en bu known amoun s o eed (n=minimum o 8 eed samples pe die ), and hen by eg essing he numbe o beads coun ed agains he known eed mass (R2=0.90–0.99). Mo eo e , sepa a e expe imen s we e pe o med o ind an app op ia e du a- ion o he eeding pe iod and o he iming o he X- ays o a oid e acua ion o he eed wi h beads be o e X- aying he ish (da a no shown). The ish we e eco ded o sex and ma u i y in May 2003 h ough isual inspec ion o seconda y sexual cha ac e s, and in No embe 2003 by he ex- amina ion o gonads om slaugh e ed ish. The ish we e classi ied in o six sex/ma u i y g oups; males we e classi ied as ma u e a 2, 3, o la e yea s, e- males as ma u e a 3 o la e yea s, and all o he ish as indi iduals o unknown sex and ma u i y age. 2.4. Repea abili y and accu acy o eco ds Repea abili ies we e calcula ed o es ablish he consis ency o daily eed in ake eco ds in each die and each h ee-week X- ay session. When e- pea ed eco ds o each indi idual a e a ailable, pheno ypic a iance (VP)is composed o VP=VG+VEg+VEs,whe eVG e e s o gene ic a iance, VEg o gene al en i onmen a iance, and VEs o he wi hin-indi idual a i- ance a ising om epea ed measu emen s. Repea abili y was calcula ed as 396 A. Kause e al. =1−VEs/VP[9] and i s s anda d e o as desc ibed by Becke [3]. The accu acy o obse a ions is inc eased by calcula ing he mean o n epea ed eco ds measu ed o each indi idual. This educes VEs by a ac o o 1/n and consequen ly educes he pheno ypic a iance o he mean alues (Vmean). Thus, he a io o Vmean o VPis a measu e o a gain in he accu acy om epea ed eco ds. The change in accu acy (in uni s o %) was calcula ed as 100−100[1 + (n−1)]/n[9]. Fu he mo e, a epea abili y o he mean o n epea ed eco ds was calcula ed as mean =n /[1 +(n−1) ]. 2.5. Leas squa es means The ish we e eco ded o body mass and daily eed in ake one o h ee imes du ing each 3-week X- ay session. Because some ish lacked one o wo o he h ee eco ds and because he e we e diffe ences among he a e - age weekly ai means, he aw means o each indi idual would ha e been unsui able o he subsequen calcula ions. The e o e, weigh ed leas squa es means o each indi idual we e ob ained by accoun ing o he es ank-wise weekly pe o mance. This was pe o med o each ank sepa a ely by i ing an analysis o a iance o he longi udinal eed in ake da a wi h a model in- cluding indi idual ish (all ish wi hin a gi en es ank) and es week (weeks 1, 2, and 3) as ac o s, and calcula ing leas squa es means o each indi idual (LSMeans op ion, SAS So wa e, Ca y, NC). These weigh ed means o he indi iduals we e used as obse a ions in he subsequen s a is ical and gene ic analyses. 2.6. S a is ical es s o die diffe ences In o de o examine he diffe ences be ween he die s in body mass and eed in ake, pa ame ic analyses o a iance we e pe o med on he leas squa es means o indi idual eco ds (p ocedu e P oc Mixed, SAS So wa e). The ixed effec s included in he model we e die , sex/ma u i y class, and in e ac- ion o die wi h sex/ma u i y. The andom ac o s included we e amily ( his consis s o bo h common en i onmen effec and gene ic effec o he 210 am- ilies), es ank nes ed wi hin die , in e ac ion o amily wi h die , in e ac ion o amily wi h sex/ma u i y class, and in e ac ion o es ank wi h sex/ma u i y. Tes ank was conside ed o be a andom ac o in he Ano a o be able o gen- e alise he esul s o include o he es en i onmen s. The me hod o Kenwa d and Roge [22] was used o calcula e co ec F- es s and hei deg ees o ee- dom o he ixed effec s. A e age eed in akes we e squa e oo - ans o med o Feed in ake in ish b eeding 397 ob ain no mally dis ibu ed esiduals o he models. When calcula ing s a is- ical es s and leas squa es means o ela i e eed in ake (a e age eed in ake co ec ed o body mass), body mass was included as a co a ia e in he models o a e age eed in ake. This ai is he e e e ed as eed in ake %. 2.7. Gene ic analyses (Co) a iance componen s we e es ima ed om he leas -squa es means o indi idual eco ds using a e age in o ma ion (AI) es ic ed maximum likeli- hood me hod o he DMU so wa e [13]. A ai measu ed on he wo die s was ega ded as wo diffe en ai s. The model o he die -speci ic eed in akes and body mass was he ollowing: yijkl =animi+ am ank j+SEXMATk+TESTTANKl+εijkl, whe e animiis a andom gene ic effec o an animal (i=1... numbe o obse - a ions), am ank jis a andom amily ank effec (j=1–210), SEXMATkis a ixed sex and ma u i y effec (k=1–6), TESTTANKlisa ixed es ankeffec (l=1–4 anks), εijkl is he esidual, and yijkl is an obse a ion o an indi id- ual. Tes ank is conside ed o be a ixed ac o in he gene ic analysis because we wan o emo e i s effec on pheno ypic a iance and because o he low numbe o es anks. Fo eed in akes eco ded a imes 2 and 3, he andom amily ank effec was negligible, and i was emo ed om he model. To cal- cula e gene ic pa ame e s o ela i e eed in ake ( eed in ake %), body mass was included as a co a ia e o he models o a e age eed in ake. Using hese models, gene ic (VG), common en i onmen (V am ank), esid- ual (VR) and pheno ypic a iances (VP=VG+VR+V am ank), as well as pheno ypic ( P) and gene ic co ela ions be ween ai s ( G) we e ob ained. He i abili y was calcula ed as h2=VG/VPand common en i onmen a io as c2=V am ank/VP. Common en i onmen (co) a iance includes effec s due o sepa a e ea ing o ull-sib amilies un il agging, bu also pa s o dominance effec s. Gene ic (co) a iances, es ima ed by he animal effec , include addi i e gene ic (co) a iance bu also pa s o dominance (co) a iance. Consequen ly, he s a is ical model used does no allow us o ge a clean (isola ed) es ima e o he addi i e gene ic (co) a iances. Ra he , gene ic a iances, he i abili ies and gene ic co ela ions es ima ed he e a e b oad-sense es ima es. To scale he pheno ypic a iance o ai s wi h diffe en means (x), coefficien o pheno- ypic a ia ion was calcula ed as CVP=√VP/x. App oxima e s anda d e o s o (co) a iances we e ob ained by a i s -o de Taylo se ies expansion o he AI ma ix o he es ima ed (co) a iances [13]. The s anda d e o s o he i abil- i ies, gene ic co ela ions, and CV we e calcula ed ollowing Becke [3]. 404 A. Kause e al. conside able eco ding noise o he long- e m eed in ake, he me hod is s ill accu a e o he pu poses o sho - e m physiological and beha iou al s ud- ies las ing a ew days. I is well es ablished ha he X- ay me hod accu a ely desc ibes he ue daily eed in ake [28,30]. I would be bene icial o ob ain highe epea abili ies o eed in ake, e.g. using imp o ed eco ding me hods, o by eco ding eed in ake h oughou on-g owing, as is done in a m animals. A ace elemen me hod in which a ma ke included in eed accumula es in he ish body has been used in ish and may p o e o be a use ul al e na i e o eco d in ake o longe pe iods o ime e.g.[10, 24]. In a m animals, a e age daily eed in ake is ypically eco ded du ing each day o a long es ing pe iod, in some cases o e he whole p oduc- ion ime (e.g., in pigs). Fo ins ance, he es ing pe iod has been 35–100 days o pigs [33,35], 70–119 days o ca le [1,2], 42–56 o lambs [4,11,37], and 23–182 days o chicken [5, 34]. In compa ison, in he p esen s udy eed in- ake was eco ded a each age only du ing h ee days, ye o e a 21-day es ing pe iod. A che e al. [1] showed how e o a iance is d ama ically dec eased when he es ing pe iod is p olonged om 7 days o 119 days in ca le. An im- p o ed eed in ake eco ding in ish would lead o mo e eliable es ima es o ea men means, indi idual eed in akes, gene ic pa ame e s, a educed need o se e al epea ed eco ds, and an inc eased powe o s a is ical es s due o educed e o a iance. Pheno ypic and gene ic co ela ions be ween a e age eed in akes eco ded a diffe en ages we e only mode a e. The highes co ela ions we e be ween successi e ages, indica ing ha some a ou able co ela ed gene ic esponses a e expec ed when selec ing only o one ai . I a e age eed in ake is o be eco ded a a single occasion, hen ma ke -sized ish (>750 g) should be p e- e ed. This is because ea ly measu emen s aken om 140 g ish canno be used as eliable p edic o s o e he whole g owing pe iod. The esul s show u he ha i one a emp s o effec i ely b eed o long- e m eed in ake, se - e al eco ds a e needed om ish o diffe en ages. In line wi h ou inding on he e- anking o amilies ac oss ages, Thodesen e al. [39] showed ha ela i e eed in ake o selec i ely b ed A lan ic salmon is highe han ha o wild salmon only du ing he ea ly s ages o g ow h (when g own om 814 g o 1455 g), e ealing a s ain-by-age in e ac ion. 4.2. He i abili y o eed in ake The he i abili y o a e age eed in ake was low (a e age h2=0.10), he highes indi idual alue being 0.17. Consequen ly, he e is sufficien Feed in ake in ish b eeding 405 gene ic a ia ion o be used o selec i e b eeding, bu clea ly less han o body mass [19, his s udy]. He i abili ies o eed in ake in poul y (a e age 0.45, [32]), pigs (0.29, [6]), sheep (a e age 0.36, [4,11,37]), and ca le (a e - age 0.41, [26]) a e highe han epo ed he e, p obably e lec ing he ac ha eed in ake displays less unexplained esidual a ia ion and eed in ake eco ds a e mo e eliable in a m animals. Wi h he excep ion o Sil e s ein e al. [36], p e ious s udies on he ge- ne ics o eed in ake in ish ha e been based on amily mean eco ds. Each ull-sib amily has been kep in hei own (o eplica ed) anks, and he ank- mean eed in ake has been eco ded. As s essed by Doupé and Lymbe y [8], he d awback o his app oach is ha by calcula ing he amily means, he g ea wi hin- amily a ia ion exis ing in ish is excluded om he da a. He - i abili ies and p opo ion o a ia ion due o amily om such s udies ange om 0.31–0.84 [23,25,40] bu hese a e o e es ima es due o he amily mean me hod. Sil e s ein e al. [36] ound a he i abili y o 0.37–0.41 o indi idual eed in ake eco ds o ca ish. The highe he i abili y compa ed o he p esen s udy may be due o species diffe ence o diffe ences in expe imen al design. Fo ins ance, Sil e s ein e al. used a ull-sib design in which case en i on- men al effec s common o ull sibs end o aise he i abili y es ima es. In pigs, a he ea ly s ages o g ow h when pigs mainly deposi p o ein, low eed in ake may be es ic ing hei g ow h. In con as , as pigs age, olun a y eed in ake g ea ly exceeds ene gy equi emen s, and hus, educes eed effi- ciency and inc eases lipid deposi ion [35,41]. Acco dingly, he aim o gene i- cally changing eed in ake is no always o educe i . In a m animals, andom eg ession me hods ha e been applied o simul aneously analyse all sepa a e eed in ake eco ds o e he whole li e span, in o de o con ol he gene ic changes occu ing a diffe en ages [33, 35]. Applying he andom eg ession me hod o indi idual daily eco ds appea s difficul in salmonids. This is due o he high day- o-day a iabili y o eed in ake, in which case, o example, eed in ake o some indi iduals has nega i e eg ession slopes e en hough hese indi iduals a e g owing s eadily du ing a es pe iod. 4.3. Die effec s on eed consump ion and epea abili ies Fish on he HP die had highe a e age eed consump ion han ish on he NP die wi h wo, o mo e, explana ions. Fi s , ish ed he HP die may be de icien in lipid o ene gy (and hey a e o smalle size) and he e may ha e been an a emp o compensa e o his by eeding mo e. I has p e iously been shown ha ish ac i ely compensa e o educed lipid s o es and small body size by 406 A. Kause e al. inc easing eed and ene gy in ake e.g.[17,18]. Second, wi h he NP die he e was no equi emen o inc ease eed in ake o mee he ex a p o ein demand associa ed wi h g ow h. This is because he excess lipid in he die may ul il ene gy equi emen s, hus allowing ish o spa e p o ein o he pu poses o g ow h, leading o a p o ein spa ing effec [12]. Mo eo e , epea abili ies we e highe wi h he HP die . This may be a esul o he highe o e all consump ion o ha die . To conclude, he analyses o he daily eed in ake da a eco ded using he X- ay me hod p o ed ha selec ion o gene ically change long- e m eed in ake can be success ul, ye epea ed obse a ions a se e al li e s ages a e needed o ensu e he accu acy o eed in ake es ima es and he efficiency o selec ion. ACKNOWLEDGEMENTS Tuija Paananen, Heli Wahl oos, Ángel Ruiz Bañón and Césa Luis Bo egue o helped in da a collec ion and p epa a ion. No alab L d. conduc ed he chemical analysis o he eeds. Rob Doupé and an anonymous e e ee p o- ided help ul commen s on he p e ious e sions o he ex . You a e all g ea ly acknowledged. This wo k has been ca ied ou wi h he inancial suppo om he Commission o he Eu opean Communi ies, Quali y o Li e and Manage- men o Li ing Resou ces P og amme, p ojec Q5RS-2001-0994 “P o ein and G ow h Efficiency in Salmonid Selec ion (PROGRESS)”. I does no e lec i s iews and in no way an icipa es he Commission’s u u e policy in his a ea. REFERENCES [1] A che J.A., A hu P.F., He d R.M., Pa nell P.F., Pi ch o d W.S., Op imum pos - weaning es o measu emen o g ow h a e, eed in ake, and eed efficiency in B i ish b eed ca le, J. Anim. Sci. 75 (1997) 2024–2032. [2] A hu P.F., A che J.A., Johns on D.J.,He d R.M., Richa dson E.C., Pa nell P.F., Gene ic and pheno ypic a iance and co a iance componen s o eed in ake, eed efficiency, and o he pos weaning ai s in Angus ca le, J. Anim. Sci. 79 (2001) 2805–2811. [3] Becke W.A., Manual o Quan i a i e Gene ics, Academic En e p ises, Pullman, 1985. [4] Cammack K.M., Leymas e K.A., Jenkins T.G., Nielsen M.K., Es ima es o ge- ne ic pa ame e s o eed in ake, eeding beha io , and daily gain in composi e am lambs, J. Anim. Sci. 83 (2005) 777–785. [5] Chambe s J.R., Wang L., McMillan I., Gene ic a ia ion o b oile eed con- sump ion and efficiency co ec ed o diffe ences in es body weigh s, Poul y Sci. 73 (1994) 1196–1203. Feed in ake in ish b eeding 407 [6] Clu e A.C., B ascamp E.W., Gene ics o pe o mance ai s, in: Ro hschild M.F., Ru insky A. (Eds.), The Gene ics o he Pig, CAB In e na ional, Walling o d, 1998, pp. 427–462. [7] de la Higue a M., Effec s o nu i ional ac o s and eed cha ac e is ics on eed in ake, in: Houlihan D., Bouja d T., Jobling M. (Eds.), Food In ake in Fish, Blackwell, Ox o d, 2001, pp. 250–268. [8] Doupé R.G., Lymbe y A.J., Indica o s o gene ic a ia ion o eed con e sion efficiency in black b eam, Aquac. Res. 35 (2004) 1305–1309. [9] Falcone D.S., Mackay T.F.C., In oduc ion o Quan i a i e Gene ics, Longman G oup L d, Essex, 1996. [10] Fo se h T., Jonsson B., Næumann R., Ugedal O., Radioiso ope me hod o es- ima ing b own ou ood consump ion, Can. J. Fish. Aqua . Sci. 49 (1992) 1328–1335. [11] F ançois D., Bibé B., B unel J.B., Weisbecke J.L., Rica d E., Gene ic pa ame- e s o eeding ai s in mea sheep, in: P oceedings o he 7 h Wo ld Cong ess on Gene ics Applied o Li es ock P oduc ion, 19–23 Augus 2002, Mon pellie , F ance, CD-ROM Communica ion No. 10-10. [12] G ay on R.D., Beamish F.W.H., Effec s o eeding equency on ood in ake, g ow h and body composi ion o ainbow ou (Salmo gai dne i), Aquacul u e 11 (1977) 159–172. [13] Jensen J., Madsen P., A use ’s guide o DMU. A package o analyzing mul i a i- a e mixed models, Resea ch Cen e Foulum, Danish Ins i u e o Animal Science, Tjele, 2000. [14] Jobling M., Baa d ik B.M., The in luence o en i onmen al manipula ions on in e - and in a-indi idual a ia ion in ood acquisi ion and g ow h pe o mance o A c ic cha , Sal elinus alpinus, J. Fish Biol. 44 (1994) 1069–1087. [15] Jobling M., Koskela J., In e indi idual a ia ions in eeding and g ow h in ain- bow ou du ing es ic ed eeding and a subsequen pe iod o compensa o y g ow h, J. Fish Biol. 49 (1996) 658–667. [16] Jobling M., Baa d ik B.M., Jø gensen E.H., In es iga ion o ood-g ow h ela- ionships o A c ic cha , Sal elinus alpinus L., using adiog aphy, Aquacul u e 81 (1989) 367–372. [17] Jobling M., Co ès D., Damsgå d B., K is iansen H.R., Koskela J., Pe u sdo i T.E., Kad i S., Gudmundsson O., Techniques o measu ing eed in ake, in: Houlihan D., Bouja d T., Jobling M. (Eds.), Food In ake in Fish, Blackwell, Ox o d, 2001, pp. 49–87. [18] Johansen S.J.S., Ekli M., S angnes B., Jobling M., Weigh gain and lipid deposi- ion in A lan ic salmon, Salmo sala , du ing compensa o y g ow h: e idence o lipos a ic egula ion?, Aquac. Res. 32 (2001) 963–974. [19] Kause A., Ri ola O., Paananen T., Män ysaa i E., Eskelinen U., Selec ion agains ea ly ma u i y in la ge ainbow ou Onco hynchus mykiss: he quan i- a i e gene ics o sexual dimo phism and geno ype-by-en i onmen in e ac ions, Aquacul u e 228 (2003) 53–68. [20] Kause A., Ri ola O., Paananen T., Wahl oos H., Män ysaa i E.A., Gene ic ends in g ow h, sexual ma u i y and skele al de o ma ions, and a e o inb eeding in a b eeding p og amme o ainbow ou , Aquacul u e 247 (2005) 177–187. 408 A. Kause e al. [21] Kennedy B.W., an de We J.H.J., Meuwissen T.H.E., Gene ic and s a is ical p ope ies o esidual eed in ake, J. Anim. Sci. 71 (1993) 3239–3250. [22] Kenwa d M.G., Roge J.H., Small sample in e ence o ixed effec s om e- s ic ed maximum likelihood, Biome ics 53 (1997) 983–997. [23] Kingho n B., Gene ic a ia ion in ood con e sion efficiency and g ow h in ain- bow ou , Aquacul u e 32 (1983) 141–155. [24] Kolehmainen S.E., Daily eeding a es o bluegill (Lepomis mac ochi us) de e - mined by a e ined adioiso opic me hod, J. Fish. Res. Boa d Can. 31 (1974) 67–74. [25] Kols ad K., G isdale-Helland B., Gje de B., Family diffe ences in eed efficiency in A lan ic salmon (Salmo sala ), Aquacul u e 241 (2004) 169–177. [26] Koo s K.R., Gibson J.P., Smi h C., Wil on J.W., Analyses o published gene ic pa ame e es ima es o bee p oduc ion ai s. 1. He i abili y, Anim. B eed. Abs . 62 (1994) 309–338. [27] Lynch M., Walsh B., Gene ics and Analysis o Quan i a i e T ai s, Sinaue Associa es, Sunde land, 1998. [28] McCa hy I.D., Ca e C.G., Houlihan D.F., The effec o eeding hie a chy on indi idual a iabili y in daily eeding o ainbow ou , Onco hynchus mykiss (Walbaum), J. Fish Biol. 41 (1992) 257–263. [29] McCa hy I.D., Houlihan D.F., Ca e C.G., Mou ou K., Va ia ion in indi idual consump ion a es o ish and i s implica ions o he s udy o ish nu i ion and physiology, P. Nu . Soc. 52 (1993) 427–436. [30] McCa hy I.D., Houlihan D.F., Ca e C.G., Indi idual a ia ion in p o ein u no e and g ow h efficiency in ainbow ou , Onco hynchus mykiss.P.Roy. Soc. Lond. B Bio. 257 (1994) 141–147. [31] NRC Na ional Resea ch Council, Nu ien Requi emen s o Fish, Na ional Academic P ess, Washing on DC, 1993. [32] Pym R.A.E., Nu i ional gene ics, in: C aw o d R.D. (Ed.), Poul y Gene ics and B eeding, Else ie , New Yo k, 1990, pp. 847–876. [33] Schnyde U., Ho e A., Lab oue F., Künzi N., Mul iple ai model combining andom eg essions o daily eed in ake wi h single measu ed pe o mance ai s o g owing pigs, Gene . Sel. E ol. 34 (2002) 61–81. [34] Schulman N., Tuiskula-Haa is o M., Sii onen L., Män ysaa i E.A., Gene ic a i- a ion o esidual eed consump ion in a selec ed Finnish egg-laye popula ion, Poul y Sci. 73 (1994) 1479–1484. [35] Schulze V., Roche R., Lo enzo Be mejo J., Loo H., Kalm E., Gene ic associ- a ions be ween obse ed eed in ake measu emen s du ing g ow h, eed in ake cu e pa ame e s and g owing- inishing pe o mances o cen al es ed boa s, Li es . P od. Sci. 73 (2002) 199–211. [36] Sil e s ein J.T., Boswo h B.G., Waldbiese G.C., Wol e s W.R., Feed in ake in channel ca ish: is he e a gene ic componen ?, Aquac. Res. 32 (2001) 199–205. [37] Snowde G.D., Van Vleck L.D., Es ima es o gene ic pa ame e s and selec ion s a egies o imp o e he economic efficiency o pos weaning g ow h in lambs, J. Anim. Sci. 81 (2003) 2704–2713. [38] Talbo C., Higgins P.J., A adiog aphic me hod o eeding s udies on ish using me allic i on powde as ma ke , J. Fish Biol. 23 (1983) 211–220. Feed in ake in ish b eeding 409 [39] Thodesen J., G isdale-Helland B., Helland S.J., Gje de B., Feed in ake, g ow h and eed u iliza ion o offsp ing om wild and selec ed A lan ic salmon (Salmo sala ), Aquacul u e 180 (1999) 237–246. [40] Thodesen J., Gje de B., G isdale-Helland B., S o ebakken T., Gene ic a ia ion in eed in ake, g ow h and eed u iliza ion in A lan ic salmon (Salmo sala ), Aquacul u e 194 (2001) 273–281. [41] Whi emo e C.T., Tullis J.B., Emmans G.C., P o ein g ow h in pigs, Anim. P od. 46 (1988) 437–445. To access his jou nal online: www.edpsciences.o g