..•

ICES STATIJTORY MEETING 1993 C.M. 1993/F:34Sess. R

Mariculture CommitteelRef. Theme Session R

•

ARE REARED JUVENll...ES FIT FOR RELEASE INTO THE WILD?

by

Terje Sväsand

Institute of Marine Research, Department of Aquaculture,P.O. Box 1870 Nordnes, 5024 Bergen, Norway

ABSTRACT

The main question that has to be raised when planning to enhance natural fish populations iswhether reared fish are fitted for a life in the wild, or more specifically, whether there aredifferences between reared and indigenous fish. There are many published examples of observeddifferences between reared and wild individuals. The causes and effects of these differences are

discussed. Results from the Norwegian enhancement programme, comparing reared and wildindividuals of Atlantic salmon, Salmo salar L., Atlantic cod, Gadus morhua L., and Europeanlobster, Homarus gammarus L. are discussed in the light of other results in the literature.

..•

Iritroduction

A critical phase in sea ranching programmes is the transition from the hatchery to thewild. In the wild, animal will try to optimise their feeding depending on the availability of prey,and the probability of being eaten. An important question is therefore if reared cod or lobstershave developed normal feeding and anti-predator behäviour compared to their wild counterparts.Releases of individuais with imrtatirral behaviour, may result in reduced suivival rates. Ir the aimof a release programme is to enhance recruitmerit to overfished populations, the released animals

must also have natural spawning behaviour, so that it cim contribute to reproduction.Several factors, such as the price and qu~ity of juveniles and the value of recaptured fish,

will decide whether a sea ranching project can be economicallyfeasible or not Usually, there willbe a close correlation between the quality of the juveniles and their prospects of survival in thesea. The quality of the juveniles should then be a key factor in the development of ecologicallyand economically based sea ranching programme~. it is imponailt to stress here that gOod qualityfry, fit for sea ranching, are not identical to good quality ti)' for net pen rearing. Some qualitiesmight be the same (e.g. resistance to disease), while many will be different. This is easy tounderstand when we think of the completely different environments of the hatchery and the wild.

In this paper I wÜI start by lookirig at morphological and behavioural differences betweenreared and wild individuals, and at the possibie causes of these differences. I then go into moredetails regarding results with Atlantic salmon, Salmo salar L., Atlantic cod, Gadus morhua L.,

and European lobster, Homarus gammarus L. I will discuss results from the Norwegian

enharicement programme (Anon. 1989) in the light of other resuIts in the literature.

Differences between wild and reared fish

•

A survey of the literature reveals inariy weiI~documented examples of morphological and

behavioural differences between Wild and reared fish and crustaceans (e.g., Blaxter 1970,1975,Browman 1989). Differences caused by the rearing environment should, however, be separated ,.from differences caused by local adaptation, where populations have evolved different lifehistory •strategies that can be attributed to inter- cr intra-species interactions, and enviromrientalheterogeneity. Comparative studies should therefore always be carried out using animals from thesame area.

Morphological differencesMorphometric characters often change with exposure to different environmental conditions

(Browrnan 1989). It is therefore expected that there will be rnorphological differences between

reared and wild individuals as shown in Table 1. For sea ranching it is of the greatest interest tofocus on rnorphological changes that may affect feeding and anti-predator behavioUr. Reared

individuals are also often characterised by unusual pigmentation, and altered coloration may

2

increase the risk of predation. Landeau and Terborgh (1986) showed iliiu dyed silvery minnows,

Hybognathus nuchalis, in schools of undyed fish were eaten more often than their schoolmates.

Although most of the literature deals With morjJhological differences that may have

negative effects on the reared individuals after release, there are few documentated examples of

increased mortality that can be attributed to such changes. This may be due to problems ofmethOd, and should not be taken as a proof of "no effect". In sea ranching programmes, efforts

should, on the contririy, be put into diminishing unnatural morphological development and

coloration.

Table 1. Examples of morphological differences between reared and Wild individuals.

Species

Lined sole (Achiruslineatus)

Several species

Ayu(P/ecoglossus altivelis)

Red Sea Bram (Pagrusmajor)

Morphological characterS

Incomplete eye migration, Finabnormalities, ambicoloration, partialalbinism.

Several morphological characters

Morphological defoImities

Body height, eye diameter, upper jawlength, and other morphometricmeasures.

Ref.

Houde 1971

Blaxter 1975

Komada 1980

Matsumiya et al.1984

Masu salmon Structure of the head skeleton Romanov 1984(Oncorhynchus masou)

Haplochromis Premaxillae Wilte 1984squamipinnis (Pisces,Cichlidae)

• American lobster Differenthition of the claws Govind and Pearce(Homarus americanus) 1986

Jack Mackerei Several morphometric measures Suda et al. 1987(Trachurus japonicus)

Europeari lobster Differentiation of the claws Wickins 1986(Homarus garrimarus)

Red Sea Bream Numbers of vertebrae, pleural and dorsal Matsuoka 1987(Pagrus major) ribs and fm rays. Bone abnormality.

3

. Behavioural differences , .Tbe development of any partieular item of t>ehaviour is detennined by both geneue arid

environmental faetofs (Huntingford 1986). Throughout their life eyeles, animaIs must rriOdify theirbehaviour aeeoroing to the risk of predation and the availability of prey organisms. Recentlyreleased fish show different feedirig behaviour from wild fish in the first period after release, andit can take several months before their ability to eapture wild prey is as effective as that of wildiridividuals. Tbis has beeri documented both for salmonids (Sosiaket al. 1979, Shustov et al.1981, O'Grady 1983, Baehman 1984, Johnsen & Ugedal 1986; 1989), and fOf Atlantie cod(Nordeide and Salvanes 1991, Kristiansen and Sväsand 1992).

Normal anti-predator behaviour after release is imponant in stock enhaneementprogrammes. Several studies. have shown that most of the major eomponents of the agonistierepenoire appear iri isolated fish at the norinal age, aIthough these are orten poorly eoordinatedand wrongly oriented (Huntingford 1886). Experience with predators will often strengthenantipredator behaviour, as has been shown for sockeye salmon (Oncorhynchus nerka), Ginetz andLarkin (1976); zebra danio (Brachydanio rerio) , Dill (1974); EirrOpean minnow (PluJxinusphoxinus), Magurran (1990), eoho salmon (Oncorhynchus kisutch), Patten (1977), Olla and Davis(1989), and others. Predator training ean therefore t>e a way of increasing the suivival of releasedfish.

Tbe seleetion pressure for some traits is often higher and differently direeted in naturethan in a laboratory tank. Reared individuals might therefore be less weIl adapted to life afterrelease than their wild eonspeeifies (Wales 1954). ,..

Another differenee is the use of formulated feed in the hatchery. Artificial feed is oftendifferent from ihe natural prey organisms. Romanov (1984) showed that significant differencesin the strUeture of the head skeleton of culnifed and wild juveniles. of masu salmon,Oncorhynchus masou. He attributed these differenees to the eulttire eonditions, and in panicularto the diet of anificial feeds used.

Blaxter (1970) suggested that marine larvae reared in tanks are almost cenainly deprivedof suffieient sensory input for the proper development of their sense organs and their assOciatedareas in the central nervous system, arid that social stress in a cTowded situauon may 1<~ad ioabnormal behaviour. and undesirable morphologieal eonsequenees. This was followed up byBrowman (1989) who discussed ontogenetic critieal periods in fish. Critical periods refer torestrieted periods during development vvhen the organism is especially sensitive to missing orunnatural types of stimulus. Ir these natural stimuli are not given during a critieal period, lastingchanges in the phenotype or behavioUr can take , piace. Browrnan (1989) provides severalexamples from critieal periods in embryology, ethology and ecology, from both fish and otheranimal groups.

Conclusion general partFrom the above data, we can be eonclude that an artifieial rearing environment will

produce individuals that differ in some traits from wild individuaIs. However, it is still unclearto what degree these differences affect the survival potential of iridividuals after release, and howlong this effeet may laSt. It is also unelear what is normal and what is unnatural. Nature showsa wide range of diversity within and between populations, and different life history strategiesmight have similar fitness. A general charaetenstic of marine organisms is their ability to modifytheir behaviour aecording to changes in their environment.

4

•"

•

•

.­J

•

• f. ~ • _... . .'.:.,:

Atlaritic salmori (Salmo salar L.)

Tbc brooostOck used for enharicemerit pwposes consists usually of Wild salmon origmatirlg fromthe river to be enhanced, or of ranched fish returning to the site of release. Deperiding on theecological traits of the stock inquestion, the eggs are stripped frOm October to De~ember. Thenewly fertilized eggs are thereafter traitsferred to the haichery where they hateh after abOut 400daY-degrees. When the yolk sac riearly is absorbed, the fry are transferred to tanks where theyare fed until the time of smoltification,1.5-2.5 years after stripping. The fry or smoIts arethereat"ter releasoo upriver (fry or presmoIts), in the rlvermouths, in the estuanes or in the sea(smoIts). . . .

. .Before they are reIeased in the wild, the salmon have thus speni a considerable. time inariificial rearirig. In comparison, rearoo cod are released from four to Cight months. afterfenilization. With regaro to fish densities and feedirig regimes, the hiltchery environment differsgreatly from the natural habitats for salmon. The fry are kept in densities of thousands of fish perm3 arid startfed on formulated feed, arid,only exceptionally zooplankton may eriter the.t8.nks Withthe iricoming water supply. The survival from ova to SInolt in hateheries rnay surpass 50%, whileit is less than 1 % in the wild (e.g. Piggiris and Mills 1985). The shelteroo hatehery environmentthus riillY. piOduce individuaIs with low potential for survival in the wild.

Wild salmon spend two to five years in the river before smoltification and migration tothe sea (Hanseri 1987). At time of smoltification, several physiological chänges occUr, and thebehaviour ofthe salinon changes (c.f. Haar 1976). Wrule wild imon parr have a positivelyrheotacticand strongly territorial behaviour (Keenleyside and Yamamoto 1962), at smoltificationthey develop a negative rheotacrlc response, rCduce their aggressive activity and start to shoaI(Kalleberg 1958), and usually migi-ate to the sea in striaIl shoals (Holm and Skilbrei pers. comm.)

ComparaÜve studies .Most of the companitive studies of wild arid reared salinon have been perfoimed on

salmon pm in river habitats. Laboratory experiments have reveaied thai high densities andcontinuous stress iri the rearing phase rriight inhibit developmeni of anormal social behaviour(Feriderson & Carpenter 1971). In a study of the temtorial behaviour of a wild and a cultivatedpm population, a significantly higher proportion of dominant fish was found in the pm fromwild parents (Norman 1987). The difference was explained by different selection pressure in thehatchery comparoo to the natural habitat of salmon. .

Several stUdies have documenied thai re3.red saImon pari have.differeni feeding behaviourfor seveciJ. moriths arter release, compared with wild Salmon (E.g., Sosiak et al.; 1979; Shustover al., 1981).

Allti.preWitor behavioiIrOuring their seaward migration sea-ranched salmon are subject to heilVY mortaiity from

prooation (e.g. Larssori 1985, Hvidsten arid M~kkeIgjerd 1987). High mortality of rcared sinoltmigration to theb sea might be attiibuied to a synergistic effect causoo by osmotic stress andfright reaction from an tinkßown predator (Järvi 1989). The physiological stress response wasrooucCd ",hen the fish were acclimatized either to the osmotic conditions or habiimitecI topre<Iatofs (Järvi 1990). Acclirriaiizauon to seawater änd predator training .before release mighttherefore help survival; Pol'gen et al. (1989) suggests thai sWimriiing training arfects survivalrates., arter release. Handling stress might inhibit ,normal anti-predator behaviour, and beforerelease, the smolts should therefore be given possibility to restrain after uinspori or handling.

'"

5

Suggestion for improvement. . , . .. .In summary, there are many documented differences betWeen reared and wild salmon,

especially in the river phase before migration to the sea. However,we .must be aware that mostof the comparative studies have been conductCd in the rivers,. arid there is corisequently rioevidence for fewer differences between wild and reared post smolts durlng migration to thefeeding grounds. or in the sea phase in general. .

Relating to the described studies, I will put forward the folloWing three suggestioris forimprovement in the rearing period; . . .. .1. Production metlwds that ensure good smolt quality for release (Stefansson 1991):

Productiori of smolt for stock enhancement deinands different strategies thm forproduction of smolt for intensive farming. Environmerital factorsmay.be manipulated tocontrol growth rate arid to influence the timing of the parr-smolt transformation. However.these manipulations may interfere with the quality of the smolts. Reaiing. conditionsshould also be optimal. as suooptlmal rearing conditions. e.g. high densities. poor feemngand poor water quality, might reduce the success of striohs arter release.

2. Stamina enforcement - Use of tanks with sufficiently strong w~lter circulation to enhancethe swimming ability of the smolt by simulatirig river conditions.

3 AccUmatization - Seawater acclimatization and predator träining prior to release.Acclimatization after handling stress.

European lobster (Homarus gammarus L.)

In Norway,lobsters for stock enhancement purposes are produced at Kyrksreter~ra, whichhas a yearly capacity of about 80.000 smalllobsters (Grimsen et al. 1987). Wild berrled femaJesare captured in the sea and hatching occurs iri individual containers with flowing seawater. Athatching the pelagic larvae are transferred to srriall tanks ("Hughes kreisels") where they Me keptuntil metarnorphosis ( 20°C <ternp.< 24°C). The length of this periOd is about two weeks. afterwhich the postlarvae are collected and transferred to separate ongrowing compartmerits wherethey live from six months to a year.

Comparative studiesThe environment in the rearing periOd is very different from the natural habitat for wild

lobster. After settlement in the wild, juvenilelobsters (Homarus americanus) will hide in the •sediment or in crevices (Wahle and Steneck 1992). Here they grow to a size of about 12-15 cm(3-4 years). In this period they prey on the fauna in the sediment and by filtering pelagic plankton(Lavalli and Barshaw 1989). In the rearing unit, the lobsters live singly in srriall ooxes. TIieirmain feed is frozen brine shrimp (Artemia salina). An important question is whether this type ofartificially rearing affects the ontogenetic development of the lobsters.

In reared lobsters, differences in morphological development and unnatural behaviour afterrelease have been documented. Earlier, most of the rearCd lobsters possessed two rriorphologicaIlysimilar cutter claws (Wickins 1986, Uglem pers. corn.). and their colour has been different fromthat of wild lobsters. Different coloration may increase the predation risk, and a missing crusherclaw may influerice feecllng behaviour and the social status of the lobsters.

In the first scientific release experiments cam6d out in Norway, more than 10% of theanimals released were enten within a few hours of release, primarily by labridae. After release

6

•'.

•. ,

some lobsters swam up and down, while some remained imnlobile ori the sea bottom foi- severaiminutes (vän der Meeren 1991). This behavioUr ceitainly increased the possibility of heing eaten.. The reasons for this behaviouT wert~ investigated by van der Meeren (1993)., Sheinvestigated the effects .of environmeniaI parameters, the presence of predators. and theiraßsporiation methoo. The results revealed that acCIimaiised lobsters whichare riöt stressedshowed funcriorial behaviour in relation to light. temperature and prerlators. and to some degreealso to each other.

Iri later releases in Norway. when the lobsters had been acciimaiised to normal sea waterfor about 15-30 miri; the lobsters showed nearly normal behaviouT (van derMeeren et al.; 1990).Iß contras~ in eäriier releases of l~bsters in Norway, they were taken directly from a nearlyfrozen siaie in the transporültion boxes, and released without fUrther acclimatisaiion.

Improvement of the rearirig methods arid new studiesResults from Wales showed that wheri live oyster spat were addCd to the rearing

chambers; most of the rearerl lobsters developed a crUsher claw (Wickiris 1986). This was alsodone at Kyrksreterimi. arid the use of shellsand gave promising results (Uglem pers. com.). Theobserved differences in coloration may also be riormalised by manipulating the coritent of pigmentin the fOod, and such experiments are now beiilg carried out (Uglem pers. com.). .

Small changes in the hatchery practice may then reduce several of the observed differenceshetweeri reared and wild lobsters. However, there is a need for more detailed comparativeinvestigations including studies of the. anti-predator behaviour of reared lobsters in differeritsiniations. interactions with other species. density-dependent effects. arid selection of the bestnursery grounds in the release area.

Atlailtlc cod (Gadus morlzua L.)

In contrast with salmon and iobster. cod are reared in semi-natui-al coriditions. Theproduciion of coo in enclosed seawater ponds has beeil documented by 0iestad er al. 1985; BIomet al. 1991. In mid-February the broodstock are transferred to spawnirig pens where they spawnriaturally. ,The fertilised eggs are then transferred to the hatchery where they hatch arter two iothree weeks dependirig on the temperature. Several million coo larvae. are released into large seaenclosure1

, where. they are scirtfed on riilturiUly, occurring plärikton. This environmerit has mariysimilarities with the natural environment of wild cod larvae. Differences are larger densities ofboth larVae and prey. In November theb year before. potential predators were killed with rothenon(ci plant poison). The reared coo accept aruficial feed from a size of about 0.5 grams. From thissize until the coo are large enough for tagging arid release between July and October (size: 10-50gram). the maiD diet of the cod consists of formulated feed.

. Due to high mortality rate (primarily thI-ough cannibalism) less than 5%.survive the pondperioo. The mortality is probably highest for the less, fit larvae. arid 'tbe pond period may thus actas aselection mechanism. This may resuIt in only the most vigorous larvae surviving.

10ther production units such as large net pens, basin and intensiveproduction methods have been tried.

7

Comparative studies ,Sirirllar feooing preferences (Sväsand and Kristiansen 1985; Kristiansen 1987) arid

differences in the weight of the stomach contents of wild and rearecI cod (Kristiansen 1987) havebeen reported, based on stomach content analyses of 1+ cOd (releasoo as O-group). Release<!rearoo cod had learnoo to catch the same prey types and in the same proportions as wild coo thesecorid week after release, but mean weight of the stoinaehs eontents were smaller (Kristiansenand Sväsand 1992). Nordeide and Salvaries (1991) found differerices in feeding behaviourbetween wild and newly releasoo reared O-group c<>d. This indicates that reared fish need sometime after release to aecliritatise to their new environment With regard to predation, Nordeide andSalvanes (1991) also found that densely stockoo O-group cod were heavily preyoo uponjtist afterrelease, and Nordeide and Sväsand (1990) reporterl differences in the behaviour ofjuvenile rearooand wild cod towards a potential predator.

Stengrund (1993) compared the feeding behavioUr of wild and reared juvenile cod towardsgobies (Gobiusculus jlavescens) in tank experiments. This study revealed differences. in thebehavioUr of the two groups: the reared fish tried to exhaust their prey, while the wild cod usedan "ambush" strategy. Similar number of prey were eaten by wild and rearoo cod, although thereared cod usoo more energy to capture their prey. The first period after transfer from theproduction unit io the release ärea may, therefore be critical. , . a

However, these differences are sinall compared with what has been reported for other ,.species. What can be the reason for this? One is obvious; the released cod were produced in asemi-natural production system (0iestad et al. 1985) and were startfed on the same riaiirrallyoccurring zooplankton as wild cod. The reared cod fed on natural zooplankton from sWt feedinguntil after metamorphosis, arid they were only offered artificial feed after about a morith pastmetamorphosis. The production pond also has many siriülarities with the natuI"al enviromrierit ",ith

. regard to predators and vegetation. Most other species used for enhancement purposes areprodueed in smaller Wtificial systems, and they are usually only offered artificial feed.Furthennore, the production environment is often tanks that have little or n6 resemblance to thenatural environment.

Suggestions ror improvement and new studies . .To summarize, none of the observed differences between rearoo and Wild cod are critical

for the further development of the cod enhancement prograrmrie. On the other hand, ii isimportant not to ignore these observations. The development of methods that result in niore rapidacclimatisation to, and dispersal within the release environment, should be investigated in futureenhancement programmes. Thus, the effects of training are confmned by several authors. Further •comparative investigations should also be carried out, arid there is a special need for moredetailed studies of the behaviour of cOd (miero-seale, diurnal behaviour, anti-predator andfeeding). .

Summary arid perspectives

I conclude that exposure to an artificial rearing environment during ontogeny can affectboth the phenotype and the behaviour of the reared individuals, and thereby also their potentialfor survival after release into the wild. Browrnan (1989) foeused on critical periods and suggestedthat spatial and temporal overlap between the developing organism and specific environmentalinputs is essential. From this point of view, the production environment should be as natural as

8

•

•

.,",.

e

•••••.•• : l \ •

possible iri order to ensirre that the flsh receive the necessary stimuli at the right time (crlticalwindow). All rearing of fish roust be ariificial to some extent, bui the application of a semi­natural prOducuon regime and the use of live prey may be an important explanation of why onlya few differences were detectoo in the cod release experiments.

It is, however, important to distinguish betweeri: . , .' .... ..' .' .a) sh()rt term differences in behaviour caitsed by lack 0/ acclimansation aTuJ stress

caUsed by transportation anti release inio a new environment,and

b) iaSiing differences in pheTwrjpe and behaviour caUsed by the rearing environment.For bcitb groups, something Carl be done to reduce 01- diIriinish the differences betWeen

reafed. and wild iridividuals. .. , Fo!' .the first type, acclimatising to the release environment, and feewng, arid prCdator

training will often give positive Cffects, and this might be a gOd investnierit in sea ranchingprogrammes. Learning is' suggested äs an important mechanism thllt provides behaviouralflexibility (Dill 1983), and that may increase. the abilitY of salmonids to avoid predatol-s (Patten

. 1977; aUa arid Davis 1989). Moreover, it has been shown iri a study with Europeannürinow,Phoxinw phoxinus, that anti-predator behaviour is inheriied, but that it caß be inodified byearlyexperience .(Magurran 1990). '. . . . ' '.'

Lasting (type b) differences might. be caused by a lack of, 01- unnätitral; SiirilUli dUrlngperiOds of oritogeny. To nomiallse these differences, we roust look at the reanngenvironment.One example is the reanng of lobs1ers, where the use of pä.rticular substrates in the rearlng bOxes~prOduces lobster fry with norinaI claws. .

The possible effectS of releasiiig deviant individuals depends on, the species and thesirategy for the release programme. Ir the airii is a "put and take fishery", ii is not a strlct demandthat the reaied individuals should resemble Wild conspecifics. In such cases, it is more imjlOrtarn10 ensure ä high return rate. . . .

In releases where the goal is to enhance oveifished populations~ as with lobster stocks iriNorway, the consequences of releasing diverging lobsters might be serious. Thus, if the aim iserihancement or restocking, strict demaiids must be. made of the individuals that are to bereleasCd. The brOodstock shöuld be taken frOm tbe release area, arid the size of the bfOodstockshould be large eriough to ensure that rare alleles also are iricorporated in the biOOdstcick. Thereared lobster inust be as siinilar ä.s possible to Wild animals bOth in the juvenile period, andwheri recruitirig to the spawnirig stOCk. Measures of reproouctive success are therefore impoItant.

,. RepOOducuvesuccess my be studied using genetically inarkoo individuals. Iri Norway, suchstudies häs been conducted on bfown trout (Skaala, 1992). In addition, nearly i50.000 geneticallymarked coo are released in NOrWay (J~rstäd et al., in press), and several thousand geneticallyroarked salmon have been released in a river in western Norway (Skaä.Ia~ 1993). We are alsotrying to find a genetic tag that can b6 USOO 10 identify released lobsters (J~rstad et al. 1992).

Firially, I wish to stress that the development of methOds for prciduction of high qualityjuveniles (for sea ranching) is the first step in the development of a large scale sea ranching... , (

programme.

ACknowledgmerits

i appreciate cntical commerits lind suggestions of Marianne Holm, Gm van der Meeren andIngebrigt Ugleni.

9

Literature. .

Anon. 1989. St.prp. nr. 95 (1989-90). Gjennomfpring av det fmanspolitiske opplegget for 1990.Baehman, R A. 1984. Foraging behavior of free-rariging wild and hatehery brown trout in a

stream. Trans. Am. Fish. Soc.: 113: 1-32.Blaxter, J. H. S. 1970. Sensory deprivation and serisory input in rearing experiments. Helgoländer

wiss. Meeresunters., 20: 642-654. .Blaxter, J.H.S. 1975. Reared and wild fish - how do they eompare? In Proc. 10th Eur. Mar. biol.

Symp. Vol. 1, pp. 11-26. ~. by G. Persone arid E. Jaspers. Universal Press, Wettern,

Belgium. . ..BIom, G., Otterä, H., Sväsand, T., Kristiansen, T.S., and Serlgstad, B. 1991. The relationship

between feeding eonditions and produetion of cod fry (Gadus morhiui L.) iri a marin~semi-enclosed ecosystem in western Norway, illustrated by use of a consumption mOdel.lCES mar. Sei. Symp., 192: 176-189 . .

Browman, H. 1. 1989. Embryology, ethology and ecology of ontogenetie critieal periods in fish.Brain, Behavior and Evolution, 34: 5-12.

Dill, L.M. 1974. The eseape response of ~e zebra danio (Brachydanio rerio) II. The effeet of •experienee. Anim. Behav., 22: 723-730. .

Dill, L.M. 1983. Adaptive flexibiÜiy in the foraging behavior of fishes. Can. J. Fish. Aquat. Sci.,40: 398-408.

Fenderson, O. C., & Carpenter, M. R 1971. Effects of crowding on the behaviour of juveriilehatchery and wild landlocked Atlantie salmon (Salmo salar L.). Anim. Behav., 19:439-447.

Ginetz, R M.; and Larkin, P. A; 1976. Faetors affeeting rainbOw trout (Salmo gairdneri)predation on migrant fry of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. BoardCan., 33: 19-24.

Grimsen, S., Jaques, RN., Erenst, V. and Balehen, J.G. 1987. Aspects of automation in a lobsterfanning plant. Mooelling, Identifieation and Control, 8(1):61-68.

Govind, C. K., and Pearce, J. 1986. Differeniiai reflex ,aetivity deterrilines claw and closer muscleasymmetry in developing lobsters. Scienee, 233: 354-356.

Hansen, L.P., 1987. Viktige fiskearter - Laks, pp. 50-66, In. Fisk i Fersvann, 0kologi ogressursforvatling. Ed. by R Borgstrpm og L.P. Harisen. LaridbIuksforlaget Oslo. •

Hoar, W.S. 1976. Smolt transformation: evolution, behaviour and physiology. J. Fish. Res. BoardCan., 33:1234-1252.

Houde, E.D. 1971. Developmental abnormalities of the flatfish AChirus lineatus reared in thelabOratory. Fish. BuH. 69(3): 537-544

Huntingford, F. A. 1986. Development of behaviour in fish. pp. 47-68. In The behaviour ofteleost fishes. Ed. by Tony J. Pitcher. Croom Helm, London & Sydney.

Hvidsten, N.A., & Mpkkelgjerd, P.l. 1987. Predation on salmon smolts, Salmo salar L., in theestuary of the River Surna, Norway. J. Fish Biol., 30: 273-280.

Johnsen, B. 0., & Ugedal, O. 1986. Feeding by hatchery-reared and wild brown trout, Salmotrutra, in a Norwegian stream. Aquacultw-e arid Fisheries Management, 17: 281-287.

Johnsen, B. 0., & Ugedal, O. 1989. Feeding by hatchery-reared brown traut, Salmo trutra L.released in lakes. Aquaculture and Fisheries Management, 20: 97-104.

10

•

•

Järvi, T. 1989. Synergisuc Cffect Oll mortaIity in Atlantic saimon~ Salmo salar, smolt causoo byosmouc stress änd presence of predators. Env. Bio!. Fish., 26: 149-162.

Järvi, T. 1990. Cumulative acute physiologieal stress in Atlantic sahnon smolts: the effect ofosmorlc imbalance arid the presellce of prCdators. Aquaculture; 89: 337-350. .

j~rstad, K.E., ~arestveii, R, van der Meeren, G. 1992. Generlc studies in .connecti.on wi~ theNOrWegian stock enhancement prograriune of lobster (Homarus ganunarus). Posterpreseniated at ;;Geneti.cs arid Evohition of Aquatic Organisms", Bangor, September 1992.

J~I"Stad, K.E., Nrevda1, G, Paulsen, 0.1., and Thorkildsen, S. in press. Release arid iecapture of'genetically taggoo fry in a Norwegian fjord system. In: Gerietics ~d Evolution of Aquatic

Kallleberg,H. 1958. Obse,rvation in a stream tank of terriotarialiiyari~ competitio~ ,injuvenile salmon arid tt~ut (seilmo salar L. arid Salmo lrutta L.)~ Rep. Inst. Freswh. Res.Drotmingholni 39:55-98.Orgarusrns, Ed. hy A.R. Beaumomt. Chapman & Hall.

Keerileyside, M.H.A, arid Yamamoto, F.T. 1962. Territorial behavioUr ofjuvenile Atlantic Salmon,. (Salmo salar 1.). BehavioUr, 19:139-169. " 'Komada; N. 1980 Incidence of gross malfonnations and vertebral anomalies of natural and

, hatchery Plecoglossus altivelis. Copeia 1:29-35. . '.. .. ' ,Kristiansen, T.S. 1987. Vekst og ernrering til utsait, oppdrettet torskeyngel og vill torsk (Gadus

mOrhua L.) i HeimarkSpollen; Austevoll. Cand. seient thesis, Departrnerit of Fisheries, .,' riiology, Uriiversity of Bergen, Norway. 134 pp. (In NorWegian). . .' .". '

Kristianseri, T.S., & Sväsand, T. 1992. Compänitive analysis of stomaeh eontents of culnired andwild cod, Gcldus TnOrhua L. Aquaculture arid Fisherles Mariagement, 23:661-668.

Lärideau, L.; and Terl>9rgh, J. 1986. OdditY arid the 'confusion effeet' in predaUon. Anim.Behav., 34: 1372-1380. '

Larsscin; P.O. 1985., i>redaticiri on migration smolt as a reguhiting f.letar in Haltie salmon, Salmo. ' salar L, populations. J. Fish Bio1.,26: 391-3~7.Lavalli; K.L~ änd Barsh.lw, D.E. 1989. Post-larVal Ameriean lo~sters (Homar~ americanw)

living in buriöws may bC suspension feewng. Mar. Behav. Physiol. 15:255-264.Magtirra1i; A.E. 1990.The inheritanee arid deveiopment of riUnriow ailti-predator behaviour.

Anun. Behäv.: 39:834-842.Matsumiya, Y., Kariämuni, iL Ob; M.; Tateishi, M~ Morphometrie comparison bCtween

arti?cially-releaSoo Red Sea Breäin and O-age wild fish. Bull. iap. Soe. Sei. Fish.' 50:1173-1178. '

Matsuob, M. 1987. Development of the skCIetal tissues and skeldal muscles in the roo seabream. BuH. Seikai. Reg. Fish. Res. Lab.; 65: 1-112.

Nordeide, i T.; and Salvanes, A.G.V~ 1991. Observations on reared newlyreleaSoo and wild coo_. , ' ,(GadifS morhua L~)~d their potenti~pr~iors. lCES max:. Sei. Synip.; 192: 139-146.Nordeide, J.T.; arid Sväsand, T. 1990. The behavioUr of wild and reared juvenile cod, GaduS

,;wrhiuz L., iowai:ds apotential pfedator. Aquae. fish. ~anage., 21: 317-325.' ,Nonnan, L: 1987. AkVarleoDservationer av revirhävdandet hos laxunger (Sa/mo salar L.) av vild

och odlad härstari11Ilriing. LaxforSknißgsinstitutet Meddeiande 1987:2.O'Giady, M. F. 1983. Observation on the dictiij habitS of wild and siocked hrown irou~ Salmo

salar L., in Irish lakes. J. Fish Biol.; i2: 593-&H.

11

OHa, B.L. & Davis M.W. 1989. Tbe role of Ieaming and stress in predator avoidance ofhatehery-reared eoho salnlOn (Oncorhynchus kisutch) juveniles. Aquaeulture; 76: 209-214.

Patten, B.G. 1977. Booy size and leamed avoidanee as factors affeeting predation on eohosalmon, Oncorhynchus kisutch, fry by Torrent seulpin, Cottus rhotheus. Fish. Bull. U.S.75: 457-459. .

Piggins, D.J.; arid Mills, C.P.R., 1985. Comparative aspeets of the biology of naturally produeedand llatehery-reared Artlantic salmon smolts (Salmo salaT L.). Aquaeulture, 45:321~333.

Pol'gen, A.V., Kueherov· O.A., Popov A.P., Vitvitskaya, L.V., Nikonorov, S.I. 1989.Expenmental rearlng of juvenile Atlantie salmon jointly with predators. Rybn. Khoz., 11:21-23

Romanov, N. S. 1984. Effeet of eulture eonditions on skuH morphology in smolts of the masusalmon, Oncorhyncus masou (Brevoort). AquaeuIiure, 41:147-153.

Shustov, Y. A., Shchurov, I. L.; & Smirnov, Y. A. 1981. Adaption rlmes of hatehery salmon,Salmo salar, to river conditions. J. Iehtyol. 20:156-159.

Skaala, 0. 1992. Genetie variation in brown trout Salnio trutta L., and application of genetiemarkers in studies on gene flow from cultured populations. Dr. Scient thesis, departinentof Fisheries and Maririe Biology, University of Bergen, Norway., 96 pp. " •

Skaala, 0. 1993. R~mming, havbeite og GMO; Unders~kingar av genetiske og ~kologiske

effektar pA viHe bestander. HaVforskningsinstituttet, senter for havbruk nr 11., 1993. 16pp. (in Norwegian).

Sosiak, A. J., RandalI, R. G., & MaeKenZie, J. A. 1979. Feeding by hatehery-reared and wildAtlantic salmon (Salmo salar) parr in streams. J. Fish. Res. Boared Can., 36: 1408-1412.

Stefansson, S. 1991. Faetors affecting the quality of Atlaritie salmon smolts for use in searanehing. pp. 82-84 In Sea Rariehing - Seienufic experienees and challenges. Proeeedingsfrom the symposium and workshop 21-23 Detober 1990, Bergen Norway, Ed. by T.N.

. 'Pedersen ,and E. Kjprsvik.Stengrund, P. 1993. Forskjeller i beiteatferd mellom oppdrettet og vill torsk (Gadus morhua L.)

og effekt av hering: To ulike beitestategier pA tangkutling (Gobiusculus flavescl!nsFabricius).Cand. scient. thesis, Departrilent of Fisheries and Marine Biology, Universityof Bergen, Norway. 154 pp. (In Norwegian). , .

Suda, Y., Shimizu; M., Nose, Y. 1987. Morhologieal differenee benyecn cultivated arid wildjaekmaekerel Trachurus japonicus. Nippon Suisan Gakkasihi 53: 59-61. •

SvAsand, T., and Kristiarisen, T.S. 1985. Release of aitificiaiIy reared O-group coastal coo (Gadwniorhua L.) in a Iandlocked fjord in western Norway. ICES CM 1985/F:1O, 12 p.

varider Meeren, G.I. 1991. Out-of-Water trarisponation effects on the ~haviour in. newlyreleased juvenile Atlantic lobsters Homa"is gammarus. Aquaeulture Eng., 10: 44-64.

van der Meeren, G.I. 1993. Initial response to physieal and biological conditioris in naive lobstersHomarus gammarus. Mar. Behav. Phys., xx-xxx-xxx.

van der Meeren, G. 1., SvAsand, T., Grimsen, S., Kristiansen, A.; and Farestveit, E. 1990. Largeseale release experirilents ofjuvenile lobsters, Homarus gammarus, in Norway. ICES C.M.19901K:2, 9 pp.

12

•

l

•

J

Wahle, R.A. and Steneck. R.S. 1992. Habitat restrietions inn early benthic life: experiments onhabitat selection and in situ predation with the American lobster. 1. Exp. Mar. Bio. Ecol.,

157:91-114.Wales, J.H. 1954. Relative survival of hatchery and wild trout. Prog. Fish-Cult 16:125-127.Wickins, J. F. 1986. Stimulation of crusher claw development in cultured lobsters, Homarus

gammarus (L.). Aquaculture and Fisheries Management, 17: 267-273.Witte, F. 1984. Consistency and functional significance of morphological differences between

wild-caught and domestic Haplochromis squamipinnis (Pisces, Cichlidae).0iestad, V., Kvenseth, P.G., and Folkvord, A. 1985. Mass production of Atlantic cod juveniles

(Gadus morhua L.) in a Norwegian salt water pond. Trans. Am. Fish. Soc., 114: 590-595.

'.

13