<%BANNER%>

Effect of pulsed gastric lavage on apparent survival of a juvenile fish in a natural system

www.elsevier.com ( Publisher's URL )
University of Florida Institutional Repository
Permanent Link: http://ufdc.ufl.edu/IR00003292/00001
 Material Information
Title: Effect of pulsed gastric lavage on apparent survival of a juvenile fish in a natural system
Series Title: Journal of Experimental Marine Biology and Ecology 422-423 (2012) 107–113
Physical Description: Journal Article
Creator: Barbour, Andrew
Publisher: Elsevier
Place of Publication: Journal of Experimental Marine Biology and Ecology
 Notes
Abstract: Dietary studies are essential for studying trophic dynamics, and are often based on analysis of stomach con- tents. A popular method to collect stomach contents is the use of pulsed gastric lavage (PGL), wherein a stream of pressurized water forces an individual to regurgitate food items. Most past experimental studies have shown no effect of PGL on survival, but these studies are limited to laboratory or cage experiments, thereby controlling for natural effects such as predation or emigration. Using a mark–recapture/resighting approach, we determined the effect of PGL on apparent survival (␣ = 1 ␣ mortality ␣ emigration) in a natural system. In two study sites, we marked a total of 200 juvenile common snook, Centropomus undecimalis (Bloch 1792) (mean=251.7mm standard length, sd=30.7, range=202–320mm) with PIT tags, lavaged 89 of these snook, and resighted 90% of marked ␣sh at least once with a telemetry array. Using the Barker survival model, we determined a signi␣cant effect of PGL on apparent survival through QAICc model selection, 95% con␣dence intervals of parameter estimates, and likelihood ratio testing (P=0.017). The PGL effect reduced QAICc model averaged maximum likelihood estimates of apparent survival by 12.0–17.4%. Since we estimat- ed apparent as opposed to true survival, we could not fully partition lethal and sublethal (emigration) effects; however, a lower incidence of emigration in lavaged individuals suggests that emigration did not drive the declines in apparent survival. Regardless of the mechanism, we found that PGL affected individuals, which is contrary to most previous controlled studies. Future researchers using PGL must consider the in␣uence of potential lethal/sublethal effects in natural settings.
Acquisition: Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Andrew Barbour.
Publication Status: Published
 Record Information
Source Institution: University of Florida Institutional Repository
Holding Location: University of Florida
Rights Management: All rights reserved by the submitter.
System ID: IR00003292:00001

Permanent Link: http://ufdc.ufl.edu/IR00003292/00001
 Material Information
Title: Effect of pulsed gastric lavage on apparent survival of a juvenile fish in a natural system
Series Title: Journal of Experimental Marine Biology and Ecology 422-423 (2012) 107–113
Physical Description: Journal Article
Creator: Barbour, Andrew
Publisher: Elsevier
Place of Publication: Journal of Experimental Marine Biology and Ecology
 Notes
Abstract: Dietary studies are essential for studying trophic dynamics, and are often based on analysis of stomach con- tents. A popular method to collect stomach contents is the use of pulsed gastric lavage (PGL), wherein a stream of pressurized water forces an individual to regurgitate food items. Most past experimental studies have shown no effect of PGL on survival, but these studies are limited to laboratory or cage experiments, thereby controlling for natural effects such as predation or emigration. Using a mark–recapture/resighting approach, we determined the effect of PGL on apparent survival (␣ = 1 ␣ mortality ␣ emigration) in a natural system. In two study sites, we marked a total of 200 juvenile common snook, Centropomus undecimalis (Bloch 1792) (mean=251.7mm standard length, sd=30.7, range=202–320mm) with PIT tags, lavaged 89 of these snook, and resighted 90% of marked ␣sh at least once with a telemetry array. Using the Barker survival model, we determined a signi␣cant effect of PGL on apparent survival through QAICc model selection, 95% con␣dence intervals of parameter estimates, and likelihood ratio testing (P=0.017). The PGL effect reduced QAICc model averaged maximum likelihood estimates of apparent survival by 12.0–17.4%. Since we estimat- ed apparent as opposed to true survival, we could not fully partition lethal and sublethal (emigration) effects; however, a lower incidence of emigration in lavaged individuals suggests that emigration did not drive the declines in apparent survival. Regardless of the mechanism, we found that PGL affected individuals, which is contrary to most previous controlled studies. Future researchers using PGL must consider the in␣uence of potential lethal/sublethal effects in natural settings.
Acquisition: Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Andrew Barbour.
Publication Status: Published
 Record Information
Source Institution: University of Florida Institutional Repository
Holding Location: University of Florida
Rights Management: All rights reserved by the submitter.
System ID: IR00003292:00001


This item is only available as the following downloads:

Barbour_et_al_2012_JEMBE_lavage ( PDF )


Full Text

PAGE 1

Effectofpulsedgastriclavageonapparentsurvivalofajuvenile shina naturalsystem AndrewB.Barbour a ,RossE.Boucek b ,AaronJ.Adams c a SchoolofForestResourcesandConservation,ProgramofFisheriesandAquaticSciences,7922NW71stSt,TheUniversityofFlorida,Gainesville,FL32653,USA b DepartmentofEarthandEnvironmentalStudies,FloridaInternationalUniversity,UniversityPark,ECS337,11200SW8thSt,Miami,FL33199,USA c MoteMarineLaboratory,CharlotteHarborFieldStation,P.O.Box2197,Pineland,FL33945,USA abstract articleinfo Articlehistory: Received21October2011 Receivedinrevisedform17April2012 Accepted20April2012 Availableonlinexxxx Keywords: Barkermodel Centropomusundecimalis Fishdiet Mark recapturemark resighting PITtagantenna Telemetry Dietarystudiesareessentialforstudyingtrophicdynamics,andareoftenbasedonanalysisofstomachcontents.Apopularmethodtocollectstomachcontentsistheuseofpulsedgastriclavage(PGL),whereina streamofpressurizedwaterforcesanindividualtoregurgitatefooditems.Mostpastexperimentalstudies haveshownnoeffectofPGLonsurvival,butthesestudiesarelimitedtolaboratoryorcageexperiments, therebycontrollingfornaturaleffectssuchaspredationoremigration.Usingamark recapture/resighting approach,wedeterminedtheeffectofPGLonapparentsurvival( =1 mortality emigration)inanatural system.Intwostudysites,wemarkedatotalof200juvenilecommonsnook, Centropomusundecimalis (Bloch 1792)(mean=251.7mmstandardlength,sd=30.7,range=202 320mm)withPITtags,lavaged89of thesesnook,andresighted90%ofmarked shatleastoncewithatelemetryarray.UsingtheBarkersurvival model,wedeterminedasigni canteffectofPGLonapparentsurvivalthroughQAIC c modelselection,95% con denceintervalsofparameterestimates,andlikelihoodratiotesting( P =0.017).ThePGLeffectreduced QAIC c modelaveragedmaximumlikelihoodestimatesofapparentsurvivalby12.0 17.4%.Sinceweestimatedapparentasopposedtotruesurvival,wecouldnotfullypartitionlethalandsublethal(emigration)effects; however,alowerincidenceofemigrationinlavagedindividualssuggeststhatemigrationdidnotdrivethe declinesinapparentsurvival.Regardlessofthemechanism,wefoundthatPGLaffectedindividuals,which iscontrarytomostpreviouscontrolledstudies.FutureresearchersusingPGLmustconsiderthein uence ofpotentiallethal/sublethaleffectsinnaturalsettings. 2012ElsevierB.V.Allrightsreserved. 1.Introduction Thestudyoftrophicdynamicsisfundamentaltoecology.Quantifyingtrophicinteractionshasenabledecologiststounderstandenergeticpathwayswithinnaturalsystems( Matichetal.,2011;Richoux etal.,2010;RosenblattandHeithaus,2011 ),measureeffectsoffragmentationandanthropogenicdegradation( Laymanetal.,2007 ), identifybasalevolutionarydrivers( Bolnicketal.,2003 ),andmaintain sheries( Wip iandBaxter,2010 ).Ideally,trophicinteractionsare quanti edbydirectlyobservingconsumption,butdirectobservation isoftendif cultorimpossible.Thus,manystudiesquantifytrophic interactionsbyexaminingrecentlyconsumedpreyitemsinaspecies' digestivetract( Dehnetal.,2007;Hyslop,1980;OlsonandBoggs, 1986;Richouxetal.,2010 ).Frequently,stomachcontentexamination occursbysacri cinganddissectingindividualstoensurenoprey itemsaremissed;however,thisisinfeasiblewithendangeredspecies, eliminatestheopportunitytoresampletheindividual,canimpact ecologicalstudies,andcaninstigatecon ictswithstakeholders. Thus,researchershavedevelopedanarrayofminimallylethaltechniquestostudydiet( Lightetal.,1983;Miranda,1986;Murphyand Willis,1996 ). Onesuchtechniqueistheuseofisotopicchemicaltracers,which isanon-lethalmethodfrequentlyusedtoidentifyanindividual'srelativetrophicpositionaswellastherelativecontributionsofdistinct basalresourcepools.However,theuseofstableisotopescanbeimpracticableinsomesystemsandoftenonlyprovidesbroaddescriptionsofdiet( Laymanetal.,inpress;Post,2002 ).Additionally,the useofstableisotopesoftenrequiresanexaminationofaspecies' diettoguideandvalidateresource-usetrends( Laymanetal.,in press ),thusstillrequiringstomachcontentanalysis. Anothernon-lethaltechniqueistheuseofemeticchemicals,such ashydrogenperoxide,toinduceregurgitation( Miranda,1986 ). Emeticchemicalsallowdirectexaminationofpreyitems,butef ciencyvariesamongspeciesandtheprocessiscumbersome( Miranda, 1986 ).Amoreconvenientmethodofdirectstomachcontentexaminationistheuseofacrylictubestoperformpulsedgastriclavage ( MurphyandWillis,1996 ).Pulsedgastriclavageforcestheindividual toregurgitateconsumedfooditemsby ushingthestomachwith JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113 Correspondingauthor.Tel.:+17034070251;fax:+13523923672. E-mailaddresses: snook@u .edu (A.B.Barbour), rbouc003@ u.edu (R.E.Boucek), aadams@mote.org (A.J.Adams). 0022-0981/$ seefrontmatter2012ElsevierB.V.Allrightsreserved. doi: 10.1016/j.jembe.2012.04.011 Contentslistsavailableat SciVerseScienceDirect JournalofExperimentalMarineBiologyandEcology journalhomepage:www.elsevier.com/locate/jembe

PAGE 2

pressurizedwaterfromatubeinsertedpasttheesophagus( Fowler andMorris,2008;HartlebandMoring,1995;Lightetal.,1983 ). Gastriclavageisawidelyusedmethodin sheriesscienceand ecology,andisusedonendangeredspecies( MurphyandWillis, 1996;ShumanandPeters,2007 )duetothebeliefthatgastriclavage isanon-lethaltechnique.Totestthedetrimentaleffectsofgastriclavage,severalstudieshavequanti edpost-lavagesurvivalinalaboratorysettingorbycagingindividualsinthewild( Hafsetal.,2011; HakalaandJohnson,2004;Lightetal.,1983;ShumanandPeters, 2007 ).However,thesecontrolledexperimentseliminateadditiveeffectsfoundinnaturalsystemsthatmayincreasemortalityrates(e.g. viapredation)orchangebehavior.Thisisofparticularimportance whenstudyingrareandendangeredspecies,orinstudiesthatrequire shtoberecapturedordetectedatalaterdateforestimation ofpopulationsize,survival,movement,orothercharacteristics. Also,manylavage-effectstudiesarecompletedovertimescalestoo short(days)toobservepossiblelonger-term(months)effects.Therefore,thereisanexigentneedtounderstandhowgastriclavageaffectsindividualsinanuncontrollednaturalsettingoveraprolonged period. Inthisstudywedeterminedtheeffectofgastriclavageontheapparentsurvivalofjuvenilecommonsnook, Centropomusundecimalis inanaturalestuarinesystem.Itwaspreviouslyestablishedthatlavageisaneffectivenon-lethalsamplingmethodforsnook( Adams andWolfe,2006 ),buttheeffectswereonlyexaminedundercontrolledconditionsviacaging.Here,weemployedamark recapture/ resightingapproachtotracklavagedandnon-lavagedindividuals. Wecalculatedmaximumlikelihoodestimatesofapparentsurvival fromthemark recapture/resightingdataforcomparisonbetween lavagedandnon-lavagedpopulations.Thisapproachallowedarare, ifnotthe rst,studyofpost-lavageeffectsfora shinanaturalsystem,inwhichnaturaleffectswerenotcontrolled. Fig.1. StudylocationsinsouthwestFlorida,USA.MarkingoccurredinSouthSilcoxandCulvertCreekonly,butrecapture/resightingeffortoccurredinfourcreeks(NorthSilcox, SouthSilcox,YuccaPen,andCulvertCreek)toreducepotentialsurvivalbiascausedbyemigration. 108 A.B.Barbouretal./JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113

PAGE 3

2.Methods 2.1.Capture recaptureandlavage ThisexperimentwasconductedfromOctober2010toApril2011 onwildcaughtjuvenilecommonsnook, C.undecimalis ,fromCharlotte Harbor,FL,USA.Allcaptureandrecapture/resightingoccurredinfour tidal-mangrovecreeksofapproximatelyequallength( 1.6km)on theeasternshorelineofCharlotteHarbor:NorthSilcox,SouthSilcox, YuccaPen,andCulvertCreek( Fig.1 ).MarkingoccurredinSouth SilcoxandCulvertCreekonly,butrecapture/resightingeffortoccurred inallcreekstoreducepotentialsurvivalbiascausedbyemigration. Forthisstudy,wefocusedonjuvenilesnook201 320mmstandardlength(SL),whichapproximatedage-1individuals( Barbour andAdams,inpress;McMichealetal.,1989;Stevensetal.,2007; Tayloretal.,2000 ).Capture recaptureofsnookwasconductedwith acenter-bagseinenet(30.5!1.8m,6.3mmmesh)andhook-andline( Table1 ).Aftercapture,all shwerescannedwithahandheld passiveintegratedtransponder(PIT)tagreader(modelno.RS601, All ex¨).AllindividualsweremeasuredforSLtothenearestmillimeter,andunmarkedindividualsweremarkedwithuniquelycoded half-duplex(HDX)PITtags(23mmlength!3.4mmdiameter,0.6g inair;TexasInstrumentsTIRFIDS-2000). BeforesurgicallyimplantingthePITtag,weperformedpulsedgastriclavage(PGL)onapproximatelyhalf(44.5%)ofthecapturedindividuals.Previousandongoingresearchindicatethatjuvenilesnook movementsandapparentsurvivalaresimilarthroughouttheseason inwhichthestudywasconducted(October April)( Barbourand Adams,inpress;Barbouretal.,2012 );therefore,wegenerallydedicatedentiresamplingdaystolavaging sh,ornot( Table1 ).Onthe fewdayswhenwedidnotlavageall sh,werandomlyselectedindividualstolavage.Thelavagedeviceconsistedofa12Vbattery,a 22.7Lperminutebilgepump(Rule¨),2.1mof2.54cmdiameter hose,anda6.35mmdiameternozzle( Fig.2 ).WeperformedPGLby forcingastreamofwaterpasttheesophagusintoanindividual's stomach.Oncethestomachofthesnookwassatiated,weexternally massagedthestomachforcingwaterandrecentlyconsumedprey itemsontoameshscreen( Fig.2 ).Werepeatedthisprocessuntil threeconsecutivetreatmentsfailedtoproduceanynewitems.The treatmenttimeforeachsnookrangedfrom30sto2min.Postlavage,weplacedthesnookina1!1!1mholdingpenforanapproximateminimumof30sbeforemarkingwithaPITtag. WeinsertedPITtagsintotheabdominalcavityofallunmarked shthrougha3mmincisionposteriorandventraltothepectoral n.Forthismark,previousstudiesfound100%tagretentionwith nomortalityfor C undecimalis of>120mmSLwithnoneedforsuturestoclosetheincision( Adamsetal.,2006;BoucekandAdams, 2011 ).All shwerereleasedwithin100moftheircapturelocation withinthecreekofcapture. Toincreaseresightingprobabilities,weemployedanarrayofelevenautonomousPITtagantennae( Adamsetal.,2006;Barbouretal., 2012 ).Acompletedescriptionofantennaeunitsandconstructionis availableelsewhere( Barbouretal.,2011;2012 ).Asamarked sh passedanantenna,therebyenteringitsmagnetic eld, thePITtag's uniquenumberwasread,andtheantenna'scomputerrecorded timeanddateofdetection.Weplacedanantennainthelower,middle,andupperstrataofeachcreekwiththeexceptionofYuccaPen upper(dueto nancialconstraints)(Fig.2in Barbouretal.,2012 ). Eachstratumwasapproximately0.5kminlength.Placingantennae outsideoftheprimarystudycreeksallowedquanti cationofemigration,whichwede nedasbeingdetectedoutsideofanindividual's markingcreek. 2.2.Modelselectionanddataanalysis WeestimatedsurvivalwiththeBarkersurvivalmodel( Barker, 1997,1999 )withinProgramMARK( WhiteandBurnham,1999 ). Thismodelallowsforincorporationofinformationfromphysicalcapture recaptureduringdiscreteprimaryperiods,andcontinuous resightingfromallantennaeduringsecondaryperiods(theintervals betweenprimaryperiods).Wede nedcapture recaptureeventsas theprimaryperiods,andcodedforuneventimeintervalsbyscaling 31daystoequalanintervaloflength1( Table1 ).Weaddedtwo dummy primaryperiods( Table1 )afterthecompletionofphysical capture recapturetocontinuemonthlyestimationofsurvivaluntil theendofthecontinuousresightingdataonApril10,2011.We createdcapturehistoriesforeachindividual,andassignedeachindividualtooneoffourgroups.Groupswerecreatedtoclassifyeach individualas:(1)lavaged,markedinCulvertCreek;(2)lavaged, markedinSouthSilcox;(3)notlavaged,markedinCulvertCreek; or(4)notlavaged,markedinSouthSilcox. Table1 Summaryofcommonsnook( Centropomusundecimalis )mark recapture/resightingdatafromOctober14th,2010toApril10th,2011. PrimaryperiodstartdateOct-14Oct-15Nov-9Nov-10Nov-26Nov-27Dec-6Jan-17Jan-18Jan-31Mar-3Apr-3 CapturegeartypeSeineSeineSeineSeineSeineSeineSeineSeineHook-and-lineHook-and-lineDummyperiodDummyperiod Timeinterval0.850.550.311.370.44110.19 TaggingcreekSSCCSSCCSSCCCCSSSSSS Averagewatertemp(¡C)26.127.0 24.925.716.618.820.121.6 Numberlavaged121500171900260 Numbernotlavaged0041430139212 Numberphysicallyrecaptured 01002000 Numberresighted2210192738411110055 Geartyperepresentsthetypeofgear(seineorhook-and-line)usedforcapture recapture,and dummy periodsrepresentperiodswhennocapture recaptureoccurred.Timeintervalswerecodedinrelationtoa31-dayperiodequalinganintervallengthof1.Timeintervalsrepresentthelengthoftimefromtheprimaryperiodtothebeginningofthefollowingprimaryperiod,and numberresighted representsthenumberofmarkedindividualsresightedbyPITtagantennaeduringtheinterval.TaggingcreekswereSouthSilcox (SS)andCulvertCreek(CC),see Fig.1 forcreeklocations. Fig.2. Photographoflavageprocedurewithajuvenilecommonsnook( Centropomus undecimalis )regurgitatingpreyitemsontoameshscreen.Tworegurgitatedprey shes canbeseenonthemeshscreen. 109 A.B.Barbouretal./JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113

PAGE 4

FortheBarkermodel,ProgramMARKestimatessevenparameters ( Table2 )( WhiteandBurnham,1999 ).Sinceresightingsdidnotoccur overtheentiregeographicrangeofourstudyspecies,wehereafter refertosurvival( s )asapparentsurvival( ),whichincorporatesthe effectsofmortalityandemigration( =1 mortality emigration). Fornumericaloptimization,wetransformedrealparametervalues [0,1]to -values[ !# # ]withthelogit-linkfunction.Additionally, duetoconcernsoverpotentialconvergenceissueswiththeBarker modelatcertainparametervalues(A.B.B.,manuscriptinpreparation),weveri edmodelresultswithanalternativeoptimization method(simulatedannealing). TheBarkermodelisheavilyparameterized,allowingforwellover 1000basemodelcombinations.Therefore,wemadeaprioriassumptionsto x r p F ,and F ( Table2 )astimeandgroup-independent(.) parameterssincewehadnodeadrecoveriesandfewphysicalrecaptures,andthereforelackedsuf cientdatatoinformtime-dependent orgroup-wiseestimates.Next,wecreatedaseriesofmodelsthat allowed R ,and R ( Table2 )tovarywithtimedependence(t)orindependence(.),bygroup(G)orsubgroup(L=lavagedvs.not lavaged;C=CulvertCreekvs.SouthSilcoxCreek),andbyanadditive (+)ormultiplicative( )combinationoftime/group.Forexample, (G t)estimatedaunique foreachofthefourgroupsforeach timeperiod,resultingin32distinct estimates; (L1)estimateda single,time-independent(.) -valueinterceptfor sharedbyall groups,andestimatedasecond,additive -valuetorepresentadifferencein forlavagedindividuals; (L2)wasidentical,excepta third -valuerepresentedadifferencein lavageeffectbycreek; and [C+L1]hadacommon !" -valueinterceptforallfourgroups, a second !" -valuetodifferentiatecreeks,andathird !" -valueto representasharedlavageeffectforthecreeks. Toassessoverdispersioninthemodel,weconductedamedian goodness-of! ttest( WhiteandBurnham,1999 )ontheglobal model: (G t)R(G t)R (G t)r(G t)p(G t)F(G t)F (G t).Wethen appliedtheestimated c (variancein ationfactor)tothecompleted modelsettoadjustsubsequentestimatesofvarianceandmodelselectioncriteria( BurnhamandAnderson,2002 ).Valuesoutsideof 1 $ c $ 4indicateaseverestructurallackof tforthegivendataset andoverallmodel( BurnhamandAnderson,2002 ).Weranked modelsusingAkaike'sInformationCriterion(AIC)values( Akaike, 1973 ).ModelswithlowerAICvaluesareassumedtobetterexplain thevariationinthedatawithmaximumparsimony.Inpractice,we usedthesmall-sample, # correctedversionofAIC,QAIC c .Models withQAIC c valuesdifferingbylessthan2wereconsideredequivalent ( BurnhamandAnderson,2004;FeareandDoherty,2004 ). Aftermodelselection,weconductedhypothesistestingbytwo methods.First,wetestedforsigni canceofthelavageeffectusing theQAIC c selectedmodel.Weexamined95%CIsofthe !" -values representingthelavageeffect.Ifthe95%CIsofthese -valuesdid notinclude0.0,asigni cantlavageeffectwasdeterminedatthe # =0.05level.Second,weconductedalikelihoodratiotestbetween theQAIC c selectedmodelandanestedmodelwithnolavageeffect ( # =0.05).Afterthehypothesistesting,weusedmodelaveraging basedonQAIC c weightingtoderive nalvaluesfromrealparameter estimates( BurnhamandAnderson,2004 ).Thesevalueswereused todeterminethemagnitudeofthelavageeffect. 3.Results Duringprimarysamplingperiods,watertemperaturesranged 15.3 27.4¡C( Table1 ),wellwithinnormalphysiologicaltolerances forjuvenilesnook( Sha andandFoote,1983 ).Intotal,wemarked 200snook(mean=251.7mmSL,sd=30.7,range=202 320mm) andlavaged89( Table1 ).Wemarked81snookinCulvertCreek(34 lavaged)and119inSouthSilcox(55lavaged).Werecaptured3markedsnookbyseinenetandresighted180(90%)atleastoncewiththe PITtagantennaearray.Resightingpercentagesbygroupwereasfollows:lavagedCulvertCreek(85.3%),notlavagedCulvertCreek (95.7%),lavagedSouthSilcox(78.2%),andnotlavagedSouthSilcox (95.3%).Antennaedetectedalowerpercentageoflavagedsnookas emigrating.Oftheresightedsnook,emigrationpercentagesbygroup wereasfollows:lavagedCulvertCreek(24.1%),notlavagedCulvert Creek(35.6%),lavagedSouthSilcox(46.5%),andnotlavagedSouth Silcox(65.6%). Median goodness-of! ttestingwithinProgramMARKresulted ina =1.74,wellwithintheacceptablethresholdof1 $ c $ 4 ( BurnhamandAnderson,2002 ).Allreportedresultswereadjusted withthisvalue.QAIC c supportedamodelwithtime-independent estimatesthatdifferedbycreekandincludedasingle,sharedlavage effect( [C+L1]R[G t]R [t])( Table3 ).Thismodelhadacommon -valueinterceptforallfourgroups,asecond !" -valuetodifferentiatecreeks,andathird !" -valuetorepresentasharedlavageeffect forthecreeks.Thesecondmostparsimoniousmodelwasidentical, exceptthesurvivalterm( [G])hadanadditional !" -valueto allowaseparatelavageeffectineachcreek,thusgivingeachgroup adistinct estimate.Thethirdmodelhadacreek,butnolavageeffect ( [C]),whilethefourth( [L2])and fthmodels( [L1])hadlavage effectson .AsdeterminedbyQAIC c weighting,thesewerethe ve modelsusedformodelaveraging.Giventhehighnumberofapriori modelcombinations,weonlyreportedarepresentativesubsetof modelsrun,includingallmodelsusedinmodelaveraging( Table3 ). -valuesfromtheQAIC c selectedmodel( Table3 )signi cantly supportedalavageeffecton .Ninety! vepercentcon denceintervals onthelogit-linked -valuerepresentingthelavageeffectdidnotoverlapzero(95%CI=0.12 1.39),andneitherdidthe -valuerepresenting Table2 Barkerjointdatamodelparameterde nitions. ParameterDe nition s i Theprobabilitythatananimalaliveat i isaliveat i +1. p i Theprobabilityananimalatriskofcaptureat i iscapturedat i r i Theprobabilityanthatanimaldiesin i i +1isfounddead. R i Theprobabilityananimalthatsurvivesfrom i to i +1isresighted (alive)sometimebetween i and i +1. R i Theprobabilityananimalthatdiesin i i +1withoutbeingfound deadisresightedalivein i i +1beforeitdied. F i Theprobabilityananimalatriskofcaptureat i isatriskof captureat i +1. F i Theprobabilityananimalnotatriskofcaptureat i isatriskof captureat i +1(thisde nition,asusedinProgramMARK,differs fromthede nitionin Barker(1997) inordertoforceprobability driveninternalconstraints)( WhiteandBurnham,1999 ). Parameters r p F ,and F were xedastimeandgroup-independent(.)throughapriori assumptions. Table3 QAIC c tablefromBarkersurvivalmodelresults. Model QAIC c $ QAIC c QAIC c weight k (C+L1)R(G t)R (t)668.430.000.57 29 (G)R(G t)R (t) 670.502.080.20 30 (C)R(G t)R (t) 671.823.390.10 28 (L2)R(G t)R (t) 672.964.530.06 29 (L1)R(G t)R (t) 673.034.600.06 28 (C+L1)R(G t)R (.)677.799.360.00 26 (G)R(G t)R (.) 679.4411.020.00 27 (G)R(G t)R (G t)680.2011.770.00 37 (C)R(G t)R (.) 682.7014.280.00 26 (L1)R(G t)R (.) 686.8418.410.00 26 (G)R(t)R (t) 697.1328.710.00 18 Formodelcreation,werepresentedtime-dependence(t),groupeffect(G),creekeffect (C),andlavageeffect(L)withtheassociatedsymbols.Theparameterofprimaryinterestwasapparentsurvival( ),andmodelsthatdifferedby $ QAIC c b 2wereconsidered equivalent.Numberofestimatedparameters( k )islistedforeachmodel. 110 A.B.Barbouretal./JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113

PAGE 5

thecreekeffect(95%CI=0.19 1.48).Likelihoodratiotestingbetween theQAIC c -selectedlavageeffect (C+L1)R(G t)R (t)andnested, creekeffectmodel (C)R(G t)R (t)indicatedasigni cantimprovementinmodel twiththeadditionofthesinglelavage-effect -value(chi-square=5.75,df=1, P =0.017).Modelaveragedapparentsurvivalestimatesattributeda17.4%reductioninthemaximum likelihoodestimatesof inCulvertCreekanda12.0%reductionof inSouthSilcoxtothelavageeffect( Fig.3 ). 4.Discussion Ourstudyofjuvenile C undecimalis revealedasigni canteffectof pulsedgastriclavage(PGL)onapparentsurvival.Thisresultcon icts withpreviouslyreported ndingsthatPGLdoesnotreducesurvival ( Hafsetal.,2011;HakalaandJohnson,2004;Lightetal.,1983; ShumanandPeters,2007 ),butthosestudieswereincontrolledsettingsanddidnotallowfornaturaleffectssuchaspredation,competition,oremigration.Ourstudyallowedfortheseadditivenatural effects(thoughthisstudydidnotaddresstheeffectmechanism) anddemonstratedasigni canteffectofPGLonapparentsurvival, whichsuggeststhat eldstudiesonotherspeciesareneededinadditiontolaboratoryandcagingstudies. Thedifferenceinapparentsurvivalmayhaveresultedfromlethal orsublethaleffects.Inthisstudy,lethaleffectsarethosethatcaused directmortalityoftheindividual,whereassublethaleffectscauseda changeinsomeaspectofthesubjectthatresultedinnon-detection inthesystem. 4.1.Lethaleffects Lethaleffectsmayhavebeencausedbyanincreaseinpredation susceptibilityfollowinglavageorbydamagefromthelavageprocedureitself.Increasedmortalityfrompredationcouldhaveoccurred duetoriskyforagingbehaviorfollowinglossofthelavagedmeal,or throughalossofequilibrium,whichwouldalterthebodysystems thatnormallyfunctiontoavoidlife-threateningsituations( Beitinger etal.,2000;Danylchuketal.,2007 ).Inthisstudy,alossofequilibrium wasunlikely,asweheld shinarecoverypenpre-releaseforseveralminutes areasonabletimeforrecovery( MundayandWilson, 1997 ) andthenreleased shbyhandwithoutnoticinganyapparent effects.Additionally,werarelyencounteredpredators,suchaslarger shandpiscivorouswadingbirds( MirandaandCollazo,1997 ),capableofpreyinguponthesize-classofthesnookwestudied.Moreover, previousresearchoncannibalismbyadultsnook,themostcommon largepredatory shinthesesystems,showedthatalthoughsmall juvenilesnookwerecannibalized,juvenilesinthe201 320mmSL werenot( AdamsandWolfe,2006 ). Thus,itisunlikelythatpredation wasamajordriverofmortality. Fig.3. BarkerQAIC c modelaveragedapparentsurvivalestimatesforjuvenilecommonsnook( Centropomusundecimalis )markedandlavagedvs.markedandnotlavagedintwo coastalmangrovecreeksplottedwith95%con denceintervals.InCulvertCreek,81 shweremarkedand34lavaged.InSouthSilcox,119 shweremarkedand55lavaged.Model resultswereadjustedwithavariancein ationfactor( )of1.74.Substantialoverlapof95%CIsexistsduetomodelaveraging,the adjustment,theuseofasingle # valuetorepresentthelavageeffect,andtransformationofoptimized # valuesfromthelogit-linkfunction. 111 A.B.Barbouretal./JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113

PAGE 6

Alternatively,damagetointernalorgansduringlavagemayhave causeddeathorinabilitytofeedeffectively.Internalorgandamage couldbecausedby:(1)theuseofhighlypressurizedwaterinducing hemorrhagingofthestomachlining,(2)thetail! rstremovalofprey shesallowingdorsaloranalspinestopuncturethestomachor esophagusofthestudysubject,or(3)stomachdistentionduetoa rapidincreaseofstomachvolumeduringhigh-pressure ushing. Suchinternalinjuriesarerarelyreportedinotherspeciesfollowing lavage( Brosseetal.,2002;HakalaandJohnson,2004;Wanner, 2006 ),butmayhaveoccurredinthisstudyduetoacombinationof PITtagging,thelavageprocedure,andsmallbodysize.Sinceapreviouscagingstudy( Adamsetal.,2006 )demonstrated100%survivalfor thestudyspeciesafterPITtaggingofjuvenile s nookof>120mmSL, itisunlikelythatthisprocedurehascausedmortalityorothernegativeeffectsassociatedwithintracoelomictagging( Cookeetal., 2011 ).Althoughitispossiblethatthecombinedeffectoftagging andlavagecausedinternaldamage,basedonextensivetaggingand lavageexperiencewithjuvenileandadultsnook( Adamsetal., 2006,2012;BarbourandAdams,inpress;BoucekandAdams, 2011 ),wefeelthatthisisunlikely,andthatalavage-speci ceffect causedthedeclineinapparentsurvival. Althoughinternalinjuriesarerarelydocumentedafterlavage, internalhemorrhagingofthestomachliningwasreportedinlargemouthbass( Micropterussalmoides )ofsimilarbodysizetothejuvenile shinourstudy( HakalaandJohnson,2004 ).Thissuggestsan increasedsusceptibilitytointernalinjuryfromlavageatlowbody size,whichwassupportedby HartlebandMoring(1995) 's nding ofanimmediate60%mortalityforthesmallestspeciestheylavaged, goldenshiners( Notemigonuscrysoleucas ),andnomortalityforthe other,largerspeciestested.Furthermore,snookingestsprey shes wholeandtypicallyhead rst(R.E.B.,personalobservation),increasingtheprobabilityofcomplicationsfromremovingprey shes tail rstduetothedefensivecharacteristicsof n-raysandspines ( Hooglandetal.,1956;L'AbŽe-Lundetal.,1996;Reimchen,1991 ). Also,sincecommonsnookemployssuctionfeedingmechanics ( Wainwrightetal.,2006 ),PGLmayhavedamagedthestructuresresponsibleforgeneratingsuctionfeedingpressure( Wainwrightet al.,2006 ),reducingpredationef ciency. Damagefromthelavageproceduremayhaveledtoimmediate mortalityafterreleaseduetoseverehemorrhaging,ortodelayed mortalityduetoaninabilitytosuccessfullyconsumeanddigest prey.However,QAIC c modelselectionresultedintime-independent survivalestimates,makinginferenceonthistopicdif cult.Timeindependentestimateswerelikelyduetoaninsuf cientnumberof markedindividualspergroupasopposedtobiologicalrealism,as wehavepreviouslyfoundaneedfortime-dependentestimatesin thesamesystemwithsimilardata( Adamsetal.,2006;Barbourand Adams,inpress;Barbouretal.,2012 ).Additionally,the11 18%reductioninresightingsforlavagedindividualssuggestsanimmediate effectofPGLonapparentsurvival,sincelavagedindividualsmust haveimmediatelyleftthesystem(throughmortalityoremigration) inordertoreducetheirprobabilityofdetection. 4.2.Sublethaleffects Thedecreaseinapparentsurvivalmayhavebeenduetosublethal (e.g.behavioral)effects.Sincewecalculatedapparentasopposedto truesurvival,wewereunabletodistinguishemigrationfrommortality.Thetraumaoftheproceduremayhavecausedlavagedindividuals toidentifythestudyareaasahigh-riskhabitat,andemigrateinsearchofhigherqualityareas( DahlgrenandEggleston,2000 ).Ifthelavageprocedureincreasedemigration,thisalterationofthestudy subject'snaturalbehaviormighthavesubstantialimplications.For example,iflavagedindividualswereusedinastudyofdemographic parameters(survivalorpopulationsize),increasedemigrationwould negativelybiasestimates( Pradeletal.,1997 ),leadingtoincorrect inferencesondemographictrendsortomisinformedmanagement decisions.Also,thedatasetwouldlikelybeoverdispersediflavaged andnon-lavagedindividualsweretreatedasasinglegroup,asthe twogroupswouldhaveheterogeneousprobabilitiesofmovement ( Schmidtetal.,2002 ).Furthermore,migrationisaninherentlyhighriskbehavior( Adamsetal.,1994 ),meaninganincreaseinemigration wouldincreasepredationriskandpotentiallyincreasemortality. However,alowerincidenceofemigrationinlavagedindividualsin thisstudysuggestsemigrationdidnotdrivethedeclinesinapparent survival. 5.Conclusions AlthoughanassumptionexiststhatPGLisaneffectivemeansof non-destructivesamplingtostudydiet,thismaynotalwaysbethe case.Moststudiesonlavageeffectshavebeenconductedincontrolledsituations,whichdonotexposetheorganismstoconditions theyfacewhenlavagedandreleasedinthe eld.Thus,thesestudies mayunderestimatetheeffectsoflavage.Our eldstudydemonstratedthatlavagehadanegativeeffectonjuvenilesnook,andprovided directiontofurtherexaminetheimplicationsofPGL.Forexample, doeslavageimpactalllifestagesofsnook,orjustjuveniles?More workisneededtodeterminetowhatextentour ndingsweredue tolethalorsublethaleffectssinceeachhasdifferentimplicationsfor thepopulationandforparameterestimatesresultingfrommark recapture/resightingdata.Finally,our ndingsshouldbeacautionto otherstouselavageasanon-destructivesamplingtoolonlyafter fullytestingforeffectsinbothcontrolledand eldexperiments, sincelavageeffectsmaybespeciesorlife-stagespeci c.Clearly,lavageispreferredcomparedtodestructivesamplingfordiet,buteffectsdoneedtoberecognizedandaccountedfor. Acknowledgments WethankJ.HillandJ.Kingfor eldassistance.Thispublication wassupportedbygrantsfrom:theNationalSeaGrantCollegeProgramoftheU.S.DepartmentofCommerce'sNationalOceanicandAtmosphericAdministration(NOAA),grantno.NA06-OAR4170014,and aMoteScienti cFoundationgranttoAJA;andaNationalScience FoundationGraduateResearchFellowshipundergrantno.DGE0802270toABB.Taggingwasconductedaccordingtostandardsset bytheMoteMarineLaboratoryInstitutionalAnimalCareandUse Committee(IACUCpermitno.0704aa1).Theviewsexpressedin thisstudyarethoseoftheauthorsonlyanddonotnecessarilyre ect theviewoftheseorganizations.[ST] References Adams,A.J.,Wolfe,R.K.,2006.Cannibalismofjuvenilesbyadultcommonsnook ( Centropomusundecimalis ).GulfMex.Sci.24,11 13. Adams,C.E.,Brown,D.W.,Keay,L.,1994.Elevatedpredationriskassociatedwithinshoremigrationsof shinalargelake,LochLomond,Scotland.Hydrobiologia 290,135 138. Adams,A.J.,Wolfe,R.K.,Pine,W.E.,Thornton,B.L.,2006.Ef cacyofPITtagsandanautonomousantennasystemtostudythejuvenilelifestageofanestuarinedependent sh.EstuariesCoasts29,311 317. Adams,A.J.,Hill,J.E.,Kurth,B.N.,Barbour,A.B.,2012.Effectsofaseverecoldeventon thesubtropical,estuarine-dependentcommonsnook, Centropomusundecimalis GulfCaribb.Res.24,13 21. Akaike,H.,1973.Informationtheoryasanextensionofthemaximumlikelihoodprinciple.In:Petrov,B.N.,Csaki,F.(Eds.),SecondInternationalSymposiumonInformationTheory.AkademiaiKiado,Budapest,pp.267 281. Barbour,A.B.,Adams,A.J.,inpress.Biologgingtoexaminemultiplelifestagesofan estuarine-dependent sh,commonsnook, Centropomusundecimalis .Mar.Ecol. Prog.Ser. Barbour,A.B.,Adams,A.J.,Behringer,D.C.,Yess,T.,Wolfe,R.K.,2011.PITtagantennae arraysas sherymonitoringtoolsintropicalenvironments.Proceedingsofthe 63rdGulfCaribbeanFisheriesInstitute,pp.118 124. Barbour,A.B.,Adams,A.J.,Yess,T.,Behringer,D.C.,Wolfe,R.K.,2012.Comparisonand cost bene tanalysisofPITtagantennaeresightingandseine-netrecapture 112 A.B.Barbouretal./JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113

PAGE 7

techniquesforsurvivalanalysisofanestuarine-dependent sh.Fish.Res.121 122, 153 160, http://dx.doi.org/10.1016/j. shres.2012.01.013 Barker,R.J.,1997.Jointmodelingoflive-recapture,tag-resight,andtag-recoverydata. Biometrics53,666 677. Barker,R.J.,1999.Jointanalysisofmark recapture,resightingandinrecoverydata withage-dependenceandmarking-effect.BirdStudy46,82 91. Beitinger,T.L.,Bennett,W.A.,McCauley,R.W.,2000.TemperaturetolerancesofNorth Americanfreshwater shesexposedtodynamicschangesintemperature.Environ. Biol.Fishes58,237 275. Bolnick,D.I.,SvanbŠck,R.,Fordyce,J.A.,Yang,L.H.,Davis,J.M.,Hulsey,C.D.,Forister, M.L.,2003.Theecologyofindividuals:incidenceandimplicationsofindividual specialization.Am.Nat.161,1 28. Boucek,R.E.,Adams,A.J.,2011.Comparisonofretentionratesformultipletagtypesin commonsnook.N.Am.J.Fish.Manag.31(4),693 699. Brosse,L.,Dumont,P.,Lepage,M.,Rochard,E.,2002.Evaluationofagastriclavage methodforsturgeons.N.Am.J.Fish.Manag.22,955 960. Burnham,K.P.,Anderson,D.R.,2002.ModelSelectionandMultimodelInference:A PracticalInformation TheoreticApproach,Seconded.Springer,NewYork. Burnham,K.P.,Anderson,D.R.,2004.Multimodelinference:understandingAICandBIC inmodelselection.Soc.Meth.Res.33,261 304. Cooke,S.J.,Woodley,C.M.,Eppard,M.B.,Brown,R.S.,Nielsen,J.L.,2011.Advancingthesurgicalimplantationofelectronictagsin sh:agapanalysisandresearchagendabasedona reviewoftrendsinintracoelomictaggingstudies.Rev.Fish.Biol.Fish.21,127 151. Dahlgren,C.,Eggleston,D.,2000.Ecologicalprocessesunderlyingontogenetichabitat shiftsinacoralreef sh.Ecology81,2227 2240. Danylchuk,S.,Danylchuk,A.,Cooke,S.,Goldberg,T.,Koppelman,J.,Philipp,D.,2007.Effectsofrecreationalanglingonthepost-releasebehaviorandpredationofbone sh ( Albulavulpes ):theroleofequilibriumstatusatthetimeofrelease.J.Exp.Mar. Biol.Ecol.346,127 133. Dehn,L.A.,Shef eld,G.G.,Follmann,E.H.,Duffy,L.K.,Thomas,D.L.,O'Hara,T.M.,2007.Feeding ecologyofphocidsealsandsomewalrusint heAlaskanandCanadianArcticasdeterminedbystomachcontentsandstableisotopeanalysis.PolarBiol.30,167 181. Feare,C.J.,Doherty,P.F.,2004.Survivalestimatesofadultsootyterns Sternafuscata fromBirdIsland,Seychelles.Ibis146,475 480. Fowler,A.L.,Morris,J.E.,2008.Anewstomachsamplecollectiondevicefor shgastric lavagesamples.N.Am.J.Fish.Manag.28,1775 1778. Hafs,A.W.,Niles,J.M.,Hartman,K.J.,2011.Ef ciencyofgastriclavageonage-0brook troutandthein uenceongrowthandsurvival.N.Am.J.Fish.Manag.31,530 534. Hakala,J.P.,Johnson,F.D.,2004.Evaluationofagastriclavagemethodforuseon largemouthbass.N.Am.J.Fish.Manag.24,1398 1403. Hartleb, C.F.,Moring,J.R.,1995.Animprovedgastriclavagedeviceforremovingstomachcontentsfromlive sh.Fish.Res.24,261 265. Hoogland,R.,Morris,D.,Tinbergen,N.,1956.Thespinesofsticklebacks( Gasterosteus and Pygosteus )asmeansofdefenseagainstpredators( Persa and Esox ).Behaviour 10,205 236. Hyslop,E.J.,1980.Stomachcontentsanalysis areviewofmethodsandtheirapplication.J.Fish.Biol.17,411 429. L'AbŽe-Lund,J.H.,Aass,P.,Sgrov,H.,1996.Preyorientationinpiscivorousbrown trout.J.Fish.Biol.48,871 877. Layman,C.A.,Quattrochi,J.P.,Peyer,C.M.,Allgeier,J.E.,2007.Nichewidthcollapseina resilienttoppredatorfollowingecosystemfragmentation.Ecol.Lett.10,937 944. Layman,C.A.,Araujo,R.,Boucek,R.E.,Hammerschlag-Peyer,C.,Harrison,E.,Jud,Z., Matich,P.,Rosenblatt,A.,Vaudo,J.,Yeager,L.,Post,D.,Bearhop,S.,inpress.Applyingstableisotopestoexaminefoodwebstructure:anoverviewofanalyticaltools. Biol.Rev. http://dx.doi.org/10.1111/j.1469-185X.2011.00208 Light,W.,Adler,H.,Arnold,D.E.,1983.Evaluationofgastriclavageforstomachanalyses.N.Am.J.Fish.Manage.3,81 85. Matich,P.,Heithaus,M.R.,Layman,C.A.,2011.Contrastingpatternsofindividualspecializationandtrophiccouplingintwomarineapexpredators.J.Anim.Ecol.80,294 305. McMicheal,R.H.,Peters,K.M.,Parsons,G.R.,1989.Earlylifehistoryofthesnook, Centropomusundecimalis ,inTampaBay,Florida.NortheastGulfSci.10,113 125. Miranda,E.L.,1986.Removalofstomachcontentsfromlivelargemouthbassusinghydrogenperoxide.N.Am.J.Fish.Manage.6,286. Miranda,L.,Collazo,J.A.,1997.Foodhabitsof4speciesofwadingbirs(Ardeidae)ina tropicalmangroveswamp.ColonialWaterbirds20,413 418. Munday,P.L.,Wilson,S.K.,1997.Comparativeef cacyofcloveoilandotherchemicals inanaesthetizationof Pomacentrusamboinensis ,acoralreef sh.J.FishBiol.51, 931 938. Murphy,B.R.,Willis,D.W.,1996.FisheriesTechniques,Seconded.AmericanFisheries Society,Bethesda,MD. Olson,R.J.,Boggs,C.H.,1986.ApexpredationbyYellow nTuna( Thunnusalbacares ): independentestimatesfromgastricevacuationandstomachcontents,bioenergetics,andcesiumconcentrations.Can.J.Fish.Aquat.Sci.43,1760 1775. Post,D.M.,2002.Usingstableisotopestoestimatetrophicposition:models,methods, andassumptions.Ecology83,703 718. Pradel,R.,Hines,J.E.,Lebreton,J.D.,Nichols,J.D.,1997.Capture recapturesurvival modelstakingaccountoftransients.Biometrics53,60 72. Reimchen,T.E.,1991.Evolutionaryattributesofhead rstpreymanipulationand swallowinginpiscivores.Can.J.Zool.69,2912 2916. Richoux,N.B.,Jaquemet,S.,Bonnevie,B.T.,Cherel,Y.,McQuaid,C.D.,2010.Trophicecologyofgrey-headedalbatrossesfromMarionIsland,SouthernOcean:insightsfrom stomachcontentsanddiettracers.Mar.Biol.157,1755 1766. Rosenblatt, A.E.,Heithaus,M.R.,2011.Doesvariationinmovementtacticsandtrophic interactionsamongAmericanalligatorscreatehabitatlinkages?J.Anim.Ecol.80, 786 798. Schmidt,B.R.,Schaub,M.,Anholt,B.R.,2002.Whyyoushouldusecapture recapture methodswhenestimatingsurvivalandbreedingprobabilities:onbias,temporary emigration,overdispersion,andcommontoads.Amphibia-Reptilia23,375 388. Sha and,P.L.,Foote,K.J.,1983.Alowerlethaltemperaturefor ngerlingsnook ( Centropomusundecimalis ).NortheastGulfSci6,175 177. Shuman,B.D.A.,Peters,E.J.,2007.Evaluationofpulsedgastriclavageonthesurvivalof captiveshovelnosesturgeon.J.AppliedIchthyol.23,521 524. Stevens,P.W.,Blewett,D.A.,Poulakis,G.R.,2007.Variablehabitatusebyjuvenilecommonsnook, Centropomusundecimalis (Pisces:Centropomidae):applyingalifehistorymodelinasouthwestFloridaestuary.Bull.Mar.Sci.80,93 108. Taylor,R.G.,Whittington,J.A.,Grier,H.J.,Crabtree,R.E.,2000.Age,growth,maturation, andprotandricsexreversalincommonsnook, Centropomusundecimalis ,fromthe eastandwestcoastsofsouthFlorida.Fish.Bull.98,612 624. Wainwright,P.C.,Huskey,S.H.,Turingan,R.G.,Carroll,A.M.,2006.Ontogenyofsuction feedingcapacityinsnook, Centropomusundecimalis .J.Exp.Zool.305,246 252. Wanner,G.A.,2006.Evaluationofagastriclavagemethodonjuvenilepallidsturgeon. N.Am.J.Fish.Manage.26,587 591. White,G.C.,Burnham,K.P.,1999.ProgramMARK:survivalestimationfrompopulations ofmarkedanimals.BirdStudy46,120 138. Wip i,M.S.,Baxter,C.V.,2010.Linkingecosystems,foodwebs,and shproduction: subsidiesinsalmonwatersheds.Fisheries35,373 387. 113 A.B.Barbouretal./JournalofExperimentalMarineBiologyandEcology422-423(2012)107 113