Prey diversity as a driver of resource partitioning between river‐dwelling fish species

Ecology and Evolution. 2017;1–11.   | 1 www.ecolevol.org Received:30June2016 | Revised:28December2016 | Accepted:14January2017 DOI:10.1002/ece3.2793 ORIGINAL RESEARCH Prey diversity as a driver of resource partitioning between riverdwelling fish species Javier Sánchez-Hernández1,2 | Heidi-Marie Gabler1 | Per-Arne Amundsen1 ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium, providedtheoriginalworkisproperlycited. ©2017TheAuthors.Ecology and EvolutionpublishedbyJohnWiley&SonsLtd. 1DepartmentofArcticandMarineBiology, FacultyofBiosciences,Fisheriesand Economics,UiTTheArcticUniversityof Norway,Tromsø,Norway 2DepartmentofZoology,Geneticsand PhysicalAnthropology,FacultyofBiology, UniversityofSantiagodeCompostela, SantiagodeCompostela,Spain Correspondence Per-ArneAmundsen,DepartmentofArctic andMarineBiology,FacultyofBiosciences, FisheriesandEconomics,UiTTheArctic UniversityofNorway,Tromsø,Norway. Email:per[email protected] Funding information NorwegianResearchCouncil;Xuntade Galicia,Grant/AwardNumber:PlanI2C Abstract Althoughfoodresourcepartitioningamongsympatricspecieshasoftenbeenexplored inriverinesystems,thepotentialinfluenceofpreydiversityonresourcepartitioningis littleknown.Usingempiricaldata,wemodeledfoodresourcepartitioning(assessedas dietaryoverlap)ofcoexistingjuvenileAtlanticsalmon(Salmo salar)andalpinebullhead (Cottus poecilopus).Explanatoryvariablesincorporatedintothemodelwerefishabundance,benthicpreydiversityandabundance,andseveraldietarymetricstogivea totalofseventeenpotentialexplanatoryvariables.First,aforwardstepwiseprocedure basedontheAkaikeinformationcriterionwasusedtoselectexplanatoryvariables withsignificanteffectsonfoodresourcepartitioning.Then,linearmixed-effectmodels wereconstructedusingtheselectedexplanatoryvariablesandwithsamplingsiteasa randomfactor.Foodresourcepartitioningbetweensalmonandbullheadincreased significantlywithincreasingpreydiversity,andthevariationinfoodresourcepartitioningwasbestdescribedbythemodelthatincludedpreydiversityastheonlyexplanatory variable. This study provides empirical support for the notion that prey diversityisakeydriverofresourcepartitioningamongcompetingspecies. KEYWORDS biodiversity,coexistence,dietaryoverlap,interindividualvariation,mixedmodels,nichetheory 1 | INTRODUCTION Resourcepartitioning,assumedtobeaprincipalmediatorofbiodiversity,hasbeencentralforunderstandinghowacommunityofspeciespersistsovertime.Consumerinteractionshavegenerallybeen viewedfromtheperspectiveofpredatordiversity(e.g.,Griffin,Haye, Hawkins, Thompson, & Jenkins, 2008; Ives, Cardinale, & Snyder, 2005;Northfield,Snyder,Ives,&Snyder,2010),withthediversity ofpreyspeciesrarelybeingtakenintoaccount(butseeDuffyetal., 2007). Biodiversityofthepreycommunitycouldbeimportantinfood resource partitioning, because increased prey diversity should enhance the possibility of interactive segregation in resource utilization(Hillebrand&Matthiessen,2009;Hillebrand&Shurin, 2005).Thus,itispertinenttoaskwhetherthereisadecreasein competitionforfoodamongspecieswhenpreydiversityishigh. Thatis,doespreydiversityinfluencecompetitiveinteractionsand food resource partitioning between sympatric species? If prey diversity is high, sympatric species may be able to segregate in resource use and partitioning may occur, as predicted by niche theory (Schoener, 1974, 1989).Alternatively, if prey diversity is low,sympatricspeciesmayutilizethesameresources,andniche overlapwillbehighorcompetitiveexclusioncouldoccur(Keddy, 2001; Schoener, 1989). The potentially important relationship between prey diversity and dietary overlap between sympatric specieshasrarelybeenexplored,butthefewexamplesthatexist fromfish(Barili,Agostinho,Gomes,&Latin,2011;Targett,1981; Wuellneretal.,2011)andothervertebrates(Jiang,Feng,Sun,& Wang,2008;Martin&Garnett,2013;Zapata,Travaini,Delibes,& Martinez-Peck, 2005)indicatethatincreasedpreydiversitymay 2 | SÁNCHEZHERNÁNDEZ Et al. mitigatecompetitionthroughenhancedresourcepartitioning(but see Wuellner etal., 2011). It should, however, be kept in mind thatvariablesotherthanpreydiversity,suchaspreyabundance, foraging mode, diel patterns, and habitat segregation for feeding,mayalsobemajordeterminantsoffoodresourcepartitioning (e.g.,Crow,Closs,Waters,Booker,&Wallis,2010;Kronfeld-Schor & Dayan, 2003; Nakano, Fausch, & Kitano, 1999; Sánchez- Hernández,Vieira-Lanero, Servia,& Cobo,2011).Consequently, thestudyoffoodresourcepartitioningrequiresaframeworkthat includesthecomplexinterplayamongpreydiversity,preyabundance,fishabundance,anddietvariation. Weexaminedtherelationshipbetweenseveralpossibleexplanatoryvariables(preydiversity,preyabundance,fishabundance,and dietvariationofspecies)andfoodresourcepartitioning(measured asdietaryoverlap)ofcoexistingjuvenileAtlanticsalmon(Salmo salar Linnaeus, 1758; henceforth salmon) and alpine bullhead (Cottus poecilopusHeckel,1836;henceforthbullhead).Weusedthesetwo fishesasmodelspeciesbecausetheirfeedingecologyandcompetitiveinteractionsarewelldocumented(Amundsen&Gabler,2008; Gabler&Amundsen,1999,2010).Bothspeciesfeedonsimilarprey withapreferenceforbenthicinvertebrates(Gabler&Amundsen, 1999),andtheyarepresumedtoberesourcecompetitorsbecause theirdietsandhabitatusearesimilarevenwhenfoodresources arelimited(Amundsen&Gabler,2008;Gabler&Amundsen,1999, 2010).Further,thetwospeciesdonotshowsignificantdielsegregationinfeedinginsubarcticrivers(Gabler&Amundsen,1999). Thisprovidesanopportunitytoexaminepreychoiceandfoodresourcepartitioninginsympatricfishspeciesbycomparingmultiple sitesthatdifferinpreydiversity,preyabundance,andfishdensity. Themainobjectivesweretoexplore(i)whetherfoodresourcepartitioningoccurredbetweenthetwospeciesand(ii)whetherprey diversityoranyotherofthepotentialexplanatoryvariablescould beidentifiedassignificantpredictorsoffoodresourcepartitioning. We hypothesized that food resource partitioningwould increase with increasing prey diversity irrespective of other site-specific characters. 2 | MATERIALS AND METHODS 2.1 | Study area ThestudywascarriedoutinRiverReisa(Figure1),asubarctic,oligotrophicriverinnorthernNorway(latitude69°N).Theriver,approximately140kmlongandaround40mwidealongthestudiedsections, drainsacatchmentareaof2,516km².Theriverdoesnothaveany significantflowregulationstructure,andthemeanannualdischargeis 34m³/swiththewaterflowtypicallypeakingat200–250m³/sinlate June(Gabler&Amundsen,1999).TheReisaNationalParkislocatedin theheadwateroftheReisabasin,andtheparkandsurroundingareas provide grazing for semidomesticated reindeer. The Reisa basin includesamixtureofgrasspaddocksandforest[birch(Betula pubescens Ehrh.)andscattered pine(Pinus sylvestrisL.)],with smallruralareas interspersedinthelowerpart.Thus,agriculture,stockbreeding,and domesticsewageeffluentsaretheprimarybutmodesthumanimpacts onthecatchment.Theclimateistypicallysubarcticwithlong,dark, andcoldwinters,andtheriverisusuallyice-coveredfromNovember untilApril.Geologically,thestudybasinischaracterizedbyanaccumulationofgraniteandgneiss,andboulders,cobble,andgravelconstitutethemainsubstratesoftheriverbottom.Theriparianvegetationis chieflycomposedbydeciduouswoodland(birch)andpineforests.No informationisavailableaboutdriftpatternsormagnitudeofterrestrial subsidiesintotheRiverReisa.Itshouldbenoted,however,thatthe contributionofterrestrialinsectstothedriftinNorwegiansubarctic riversmaybeverynoticeablefromJunetoOctober(Johansen,Elliott, &Klemetsen,2000).Infact,terrestrialinsectsarethelargestgroupin thedriftofanothernorthernNorwegianriver(RiverSaeterelva,latitude68°N)inAugust,butwithverylowdensitiesinMay(Johansen etal.,2000). RiverReisasupportsrecreationalsalmonanglingandtheannualreportedcatchesofsalmonoverthelast20yearshavevariedgreatlyfrom afewhundredkgtonearly12,000kg.Catcheswereparticularlyhigh from2008to2011(approximately 8,000–12,000kg/year)(Svenning, 2011).Thedistributionofthebullheadcoincideswiththatofsalmon, FIGURE1 LocationoftheRiverReisa (inred),northernNorway,showingthe samplingsites(SS,graycircle)labeledfrom theupperpart(SS1)tolowerpart(SS11) | 3 SÁNCHEZHERNÁNDEZ Et al. TABLE1 Foodresourcepartitioning(measuredasdietaryoverlap,%)betweenAtlanticsalmonparrandalpinebullhead,preyavailability(preydiversity—measuredasShannon’sdiversity index,andabundance—estimatedasind./m2),fishabundance(fish/100m2),anddietarymetricsofthetwofishspecies(Levins’index,individualdietaryspecialization,andsurfaceprey contribution)fromthedifferentsamplingsites(SS)intheReisaRiver.1-IS=prevalenceofindividualdietaryspecialization,where1-ISisgivenasmean±SD.Alpinebullhead(bul),Atlanticsalmon (sal),contributionofsurfacepreyinthediet(surface).Overlaptotal=dietaryoverlapcalculatedusingallprey.Overlapaquatic=dietaryoverlapcalculatedwithoutsurfaceprey Sampling sites SS1 SS2 SS3 SS4 SS5 SS6 SS7 SS8 SS9 SS10 SS11 Preyresources Diversity 0.74 0.94 0.78 0.50 0.70 0.98 0.67 0.75 0.77 0.79 0.90 Abundance 393.2 304 82.2 95.3 63.6 83.3 120.1 123.1 63.6 144.6 77.9 Fishabundance Alpinebullhead 36.6 35.1 1.9 20 28.6 22.6 0.9 19.4 2.8 18.4 2 Atlanticsalmon 5.3 2.5 8.9 2.9 0.001 0.001 16.3 3.9 8.9 4.4 3.9 Browntrout 0 0 1.14 0.70 0 0 5.70 4.96 4.61 1.90 3.82 Arcticcharr 0 1.53 0 0 0 0 0.79 5.86 0.82 4.43 1.27 Total 41.90 39.13 11.94 23.60 28.60 22.60 23.73 34.10 17.14 29.13 10.99 Diet Levins(bul) 3.0 5.3 5.9 4.2 4.8 6.9 3.9 5.8 6.8 5.3 6.2 Levins(sal) 5.8 2.2 9.0 5.6 4.1 6.6 4.3 5.4 2.8 1.9 3.6 1IS(bul) 0.54±0.17 0.64±0.18 0.59±0.11 0.62±0.17 0.61±0.16 0.80±0.10 0.49±0.16 0.69±0.14 0.76±0.12 0.66±0.13 0.64±0.13 1IS(sal) 0.47±0.04 0.42±0.22 0.50±0.08 0.64±0.23 0.54±0.13 0.75±0.09 0.58±0.17 0.61±0.13 0.46±0.16 0.45±0.21 0.63±0.06 Surface(bul) 0 0 0 4.6 9.5 1.8 0 0 0.9 0.2 0 Surface(sal) 0 65 11.7 6 42.5 0 0 0 0 20 30 Overlaptotal 54.9 11.5 31.7 62.5 31.6 33.5 34.4 44.6 33.2 41.3 25.3 Overlapaquatic 54.9 44.0 37.5 65.4 57.4 34.4 35.5 44.6 33.6 51.4 40.3 Samplingsize Alpinebullhead(n)37 69 5 57 32 32 9 43 17 29 11 Atlanticsalmon(n) 9 8 16 12 13 961 20 10 7 14 4 | SÁNCHEZHERNÁNDEZ Et al. andthesefisharethedominantspeciesinthefishcommunityofthe river.Otherfishspecies,suchasArcticcharrSalvelinus alpinus(Linnaeus, 1758),browntroutSalmo truttaLinnaeus,1758,andthree-spinesticklebackGasterosteus aculeatusLinnaeus1758,arealsopresentintheriver basin (Gabler &Amundsen, 2010). A natural waterfall located about 90kmfromthesearepresentstheupstreamlimittomigratingfish. 2.2 | Sampling Samplingprotocolsusedinthisstudyconformtotheethicallawsof the country. Based on previous knowledge of the study area (e.g., Amundsen & Gabler, 2008; Gabler & Amundsen, 1999, 2010), the samplingdesignwasperformedtomatchwiththedistributionofthe modelspeciesaswellastoensurevariationsinbioticconditions(fish andbenthicinvertebrates)amongsamplingsitesalongtheRiverReisa. Samplingoffishandbenthicinvertebrateswasconductedat11sites alongthemaincourseoftheriverinAugust2004(Figure1).Augustis thetimewhentheaquaticfoodresourcesupplyislowestrelativeto theenergeticrequirementsofsalmonandbullheadsandthustheperiodwhencompetitiveinteractionsshouldbestrongest(Amundsen, Bergersen,Huru, & Heggberget,1999; Amundsen & Gabler, 2008; Amundsen, Gabler, Herfindal, & Riise, 2000; Gabler & Amundsen, 2010).Someofthestudysectionswererelativelyclosetoeachother; theminimumdistanceapartwasbetweenSS11andSS10andwas FIGURE2 Proportionofdifferent preygroupsinthestomachcontentsof Atlanticsalmonparr(whitebars)andalpine bullhead(blackbars)(thecategory“others” includeschydorids,watermites,and unidentifiedpreytaxa).Dataarepresented foreachsamplingsiterankedfromthe highesttothelowestfoodresource partitioning(dietaryoverlapvalue).The presenteddietaryoverlapvaluesare calculatedwithallpreytypesincluded(i.e., withthehighesttaxonomicalresolutionas inTableS2) | 5 SÁNCHEZHERNÁNDEZ Et al. approximately3km,whereasthemaximumdistanceapartwasabout 10kmbetweenSS5andSS4(Figure1).Weassumedfishmovement betweensamplingsiteswouldbenegligible,andthestudysections weredeemedindependent.Datafromapublishedstudy(Gabler& Amundsen,1999)withmonthlysamplingduringtheice-freeseason wereincludedintheanalysestoexamineforpossibleseasonalvariations in the prey diversity food resource partitioning relationship. Thesedatawerecollectedfollowingthesamesamplingprotocolasin thepresentstudy,allowingdirectcomparisonbetweenstudies. Fishandbenthicinvertebratesampleswerecollectedfromriffles withcobbleandgravelasthemainsubstrate.Priortosampling,siteselectionwasvisuallyperformedtoensurehabitatsimilarityamongsamplingsitestodiminishanypossibleerrorintheresultsrelatedtofield samplingtechniquessuchasbiasinfishremovalrateamongsampling sites. Thus, habitat conditions among sampling sites were deemed similar,butnospecifichabitatmeasurementsweretaken. Benthic invertebrates were collected at each site to study the preyavailability.Sampleswerecollectedimmediatelyafterfishsampling near to where electrofishing was conducted. Three parallel samplesweretakenusingthekickingmethod(Williams&Feltmate, 1992),standardizedbykickingfor3mininsideametalframedefining1.5×1.5mofthebottom.Benthicinvertebratesweresortedand identifiedtothelowesttaxonpossible,andpreyabundancewascalculatedasnumberofindividualsperm2.PlecopteraandEphemeroptera nymphsandTrichopteralarvaewereidentifiedtospecieslevel,and othertaxatothegenusorfamilylevel.Preydiversity(Hʹ)wascalculatedasShannon’sdiversityindex(Shannon&Weaver,1949): where pi is the proportion of species i in the benthic invertebrate samples. Fishwere collected using portable backpack electrofishing gear withpulseddirectcurrent(GeOmegabackpackmodel;700–1,400V, 5Amaximumintensity,40–80Hz)andasingleanodeof30cmdiameter.Three-passremovalelectrofishingwasconductedateachsampling sitewith30minbetweenpassesfollowingthestandardizedprocedures described for the EU Water Framework Directive (European Commission,2000)bytheCENdirectiveonfishingwithelectricityin wadeablerivers(CEN,2003).However,duetolargeriverwidthsand depths,nonetswereusedtoblocktheupstreamanddownstream boundaries. Fish sampling was conducted in an upstream direction fromtheriverbanktoawaterdepthofabout70cmoverastream sectionof100m.Eachfishwasidentified,measured(forklength,mm), andpreservedin96%ethanolforlaterdissectionanddietaryanalysis.Althoughthedepletionmethodwithonlythreepassesmaybe inadequatetoestimatefishabundance,particularlyinsamplingevents withlowcaptureprobabilities(e.g.,Dorazio,Jelks,&Jordan,2005),the abundanceofeachfishspecies(hereoverallfishdensityregardlessof fishlength)ateachsitewasestimatedasnumberoffishper100m2 usingZippinmultiple-passdepletionmethod(Zippin,1956).Although thisfishdensityestimationmightberough,itisassumedtoprovide representativeestimatesoftherelativefishabundanceamongsamplingsites.FishabundancesaregiveninTable1. Toavoidbiasresultingfrompossibledifferencesinfeedingbehaviorofdifferentsizeclassesoffish(e.g.,Dineen,Harrison,&Giller,2007; Hesthagen,Saksgård,Hegge,Dervo,&Skurdal,2004),onlyindividuals <100mmwereusedfordietanalysis.Intotal,179salmonparrand 341bullheadswerecaught,ofwhich142salmonand341bullheads wereusedforstomachcontentanalyses(SCA).Weattemptedtocollectatleasttenindividualsofeachfishspeciesfromeachsamplingsite. Althoughthisgoalwasnotalwaysachievedforsalmon(successfulin sevenofelevensamplingsites)andbullhead(nineofelevensampling sites;seesamplingsizesofeachlocalityinTable1),weassumethat thecapturedindividualsarerepresentativeoftheentirepopulation. Additionally,todismissanypossibleimpactoftheunequalsampling sizes,wegenerated1,000bootstrapsamples(seeSection2.3below). Thestomachswereopened,andthepercentageoftotalfullness wasvisuallydetermined,rangingfromempty(0%)tofull(100%)(see subjectivemethodsinHyslop,1980).Eachpreyitemwasthenidentifiedtothesametaxonomiclevelasforthebenthicinvertebratesamples.Therelativecontributionofeachpreytothetotalstomachfullness was estimated according toAmundsen, Gabler, and Staldvik (1996). Thatis,thesumofallpreycategoriesofastomachmeetsthevisually determinedtotalfullness.Inmathematicalterms,thecontributionof eachpreytothedietisdescribedaspercentpreyabundance(Ai): where Siisstomachfullnessofpreytypei,Stisthetotalfullnessofall preycategories,andnisthenumberoffishwithpreyiinthestomach.For thegraphicalrepresentation,preytypicallycaughtatthewatersurfaceincludingbitingmidges(Culicoidesspp.),aerialstagesofaquaticinsects,spiders,andunidentifiedterrestrialinsectswerecombinedanddesignated as“surfaceprey.”Similarly,theaquatictaxaweregroupedintosevenprey categories (Ephemeroptera, Plecoptera,Trichoptera, Diptera, Mollusca, Coleoptera,andothers)fortheplottingofthedietgraphs(Figure2). Dietaryoverlap(Pjk)wascalculatedaspercentageoverlap(Krebs, 1989)usingthelowesttaxonomicresolutionsofprey: where Pjkisthepercentageoverlapbetweenspeciesjandk,andAij andAikarethepercentpreyabundanceofresourceiusedbyspecies jandk,nisthetotalnumberofresourcecategories.Preydiversity analyseswererestrictedtobenthicinvertebrates,sodietaryoverlaps werealsocalculatedandanalyzedwithoutinclusionofsurfaceprey. Weaddressedthetrophicnicheatthepopulationlevelbyestimatingnichebreadth(B)usingLevins’index(Levins,1968): where Piistheproportionofeachpreytypeiinthedietexpressedas fractionratherthanpercentage(Amundsen,Knudsen,&Bryhni,2010). To study individual dietary specialization, the proportional similarity (PSi) index was calculated (Bolnick, Yang, Fordyce, Davis, & Svanbäck,2002): (1) H �=− ∑ pilog10p i (2) A i=100 n ∑ i=1 Si ( n ∑ i=1 St )−1 (3) P jk = [n ∑ 1 (minimum Aij,Aik) ] (4) B =1∕ ∑ P 2 i 6 | SÁNCHEZHERNÁNDEZ Et al. where Pijistheproportionofresourcecategoryjinthedietofindividuali,andQjtheproportionofresourcecategoryjinthedietofthe population.Thisindexcompareseachindividual’sdiettothatofthe population,withvaluesrangingbetween0and1.Forindividualsthat specializeonasingleorfewpreytypes,PSivaluesarelow,whereas forindividualsthatconsumeresourcesinasimilarproportiontothe populationasawhole,PSivaluesapproach1(Bolnicketal.,2002). Theoverallprevalenceofindividualspecializationwascalculatedas theinverseoftheaverageindividualPSivalues(Quevedo,Svanbäck, &Eklöv,2009). 2.3 | Statistical analyses Therelationshipsbetweendietaryoverlap(henceforthfoodresource partitioning)andthebioticvariables(explanatoryvariables)wereinvestigatedwithlinearmixed-effectmodelsusingsamplingsiteasa randomfactor.Thedatawerehierarchicallystructuredwithexplanatoryvariables being nestedwithin samplingsites,and mixed-effect models were used to account for potential random effects among sampling sites. Thus, the random part contains components that allowforheterogeneityofvariablesamongthestudiedsamplingsites. Seventeenpotentialexplanatoryvariablesoffoodresourcepartitioningwereconsidered(Table2).First,weselectedfixedterms(i.e.,explanatoryvariablesthataredeterministic)thatdescribetheresponse variableY(herefoodresourcepartitioning)asafunctionoftheexplanatory variables. The optimal fixed component was established basedonastepwiseforwardselectionmethod(stepfunction).This procedureenabledustoselectwhichexplanatoryvariablesaresignificant,andwhicharenot.ThisselectionwasmadeaccordingtoAkaike informationcriteria(AIC)(Akaike,1974).Nineexplanatoryvariables wereselectedformodelsimulations(seesignificantexplanatoryvariablesinTable2).Next,webuiltmodelsbasedonthenineselected explanatoryvariablesusingtherestrictedmaximumlikelihood(REML) estimation for linear regression models. REML aims to correct the estimatorforthevariance,andassuggestedbyZuur,Ieno,Walker, Saveliev, and Smith (2009), this procedure should be used to fit modelswithmanyfixedterms(heren = 9).Modelselectionwasalso establishedusingAIC.Whensamplesizeissmallorthenumberofparametersislarge,AICc(AICcorrectedforsmall-samplebias)orQAICc (AICcforoverdisperseddata)shouldbeusedinsteadofAIC(Anderson & Burnham, 2002). In the present study AICc was used for model selection,withthebestmodelbeingtheonewith the lowestAICc (5) PS i=1−0.5 ||| Pij −Qj ||| = ∑ (Pij,Qj ) Explanatory variables Definition Correlation Preydiversity* Macrozoobenthosdiversitycalculatedas Shannon’sdiversityindex R=−.73,p = .011 Preyabundance* Macrozoobenthosabundanceestimatedasind./m2R=.07,p = .831 Atlanticsalmon abundance* Density(fish/100m2)ofAtlanticsalmonparr R=−.01,p = .992 Alpinebullhead abundance* Density(fish/100m2)ofalpinebullhead R=.14,p = .684 Browntroutabundance Density(fish/100m2)ofbrowntrout R=−.04,p = .895 Arcticcharrabundance* Density(fish/100m2)ofArcticcharr R=.03,p = .934 Totalfishabundance Totalfishcommunitydensity(fish/100m2) R=.18,p = .601 Surfaceprey(Atlantic salmon)* Contributionofsurfacepreyinthedietof Atlanticsalmonparr R=−.67,p = .024 Surfaceprey(alpine bullhead) Contributionofsurfacepreyinthedietof alpinebullhead R=.15,p = .666 Nichebreadth(Atlantic salmon) Levins’indexofAtlanticsalmonparr R=.32,p = .330 Nichebreadth(alpine bullhead)* Levins’indexofalpinebullhead R=−.50,p = .120 Individualspecialization (Atlanticsalmon) IndividualdietaryspecializationofAtlantic salmonparr R=.22,p = .508 Individualspecialization (alpinebullhead) Individualdietaryspecializationofalpine bullhead R=−.16,p = .628 Stomachfullness (Atlanticsalmon) Stomachfullness(%)ofAtlanticsalmonparr R=.54,p = .085 Stomachfullness(alpine bullhead) Stomachfullness(%)ofalpinebullhead R=.08,p = .818 Size(Atlanticsalmon)* Forklength(mm)ofAtlanticsalmonparr R=−.50,p = .114 Size(alpinebullhead)* Forklength(mm)ofalpinebullhead R=−.24,p = .484 TABLE2 Fulllistofexplanatory variablesusedtoexploretheirpossible influenceofonfoodresourcepartitioning (measuredasdietaryoverlap)between juvenileAtlanticsalmon(Salmo salar)and alpinebullhead(Cottus poecilopus). Significantexplanatoryvariablesafter stepwisevariableselection(*).Pearson’s rankcorrelationbetweeneachexplanatory variableandfoodresourcepartitioningis shown(significantonesmarkedinbold) | 7 SÁNCHEZHERNÁNDEZ Et al. values.Thestrengthofassociationbetweenfoodresourcepartitioningandexplanatoryvariablesfromthebestmodelswastestedusing Pearson’srankcorrelation.Finally,weransensitivityanalysestotest whetherlinear mixed-effectmodelswere the sameafter excluding surfacepreyfromthedietaryanalyses.Asignificancelevelofp = .05 wasusedinallanalyses.ModelswereperformedusingR3.2.2(RCore Team2015)using“nlme”(Pinheiro,Bates,DebRoy,&Sarkar,2016) and“MuMIn”(Bartoń,2016)packages.Thebootstrappingtechnique wasperformedusingthe“boot”package(Canty&Ripley,2016)employingtechniquesoutlinedinZuuretal.(2009)foranadditionaltest ofthemodel.Weappliedaparametricbootstrap(n = 1,000)onthe bestlinearmixed-effectsmodelexplainingvariationoffoodresource partitioningbetweenAtlanticsalmonparrandalpinebullhead.The modelwasappliedonthebootstrappeddatafollowingthesamemodelingproceduresasdescribedabove.Residualsofthefinalselected model(originaldata,bootstrappeddata,andsensitivityanalyses)were visuallyinspectedfordeviationsfromnormalityandheteroscedasticity,withoutfindinganyevidenceforviolationofmodelassumptions (seeFig.S1). 3 | RESULTS 3.1 | Prey resources Preydiversityvariedwidelyamonglocalities,withtheShannonindex rangingfrom0.50to0.98,andpreyabundancesvariedamongsamplingsites,rangingfrom63.1to393.2ind./m2(Table1).Chironomidae wasusually the mostabundant taxon, butinsome localitiesBaetis spp.,Ephemerella aurivillii(Bengtsson),andCapniasp.werenumerically dominant(taxarecordedinbenthicinvertebratesamplesaregivenin TableS1). 3.2 | Food resource partitioning Both salmon and bullhead fed mainly on benthic invertebrates (Figure2),butdifferenceswerefoundbetweenthespeciesandamong localitiesinthecontributionsofthedifferentpreytaxatothediet.In general,EphemeropteranymphsandlarvalDipteraandTrichoptera dominatedthedietofbothfishspecies,withabundancevaluesrangingbetween21.3%and94.7%.Surfacepreywasanimportantdietary componentforsalmoninsomelocalities(Figure2)(detailsofstomach contentanalysesaregiveninTableS2). Meandietaryoverlapbetweensalmonandbullheadwas36.8%, butoverlapvariedquitewidelyamongsamplingsites,rangingfrom 11.5%to62.5%(Table1).Amodelthatincludedpreydiversityasthe onlyexplanatoryvariablewasthebestone,havingthelowestAICc value(TableS3),andparametersofthismodelaregiveninTable3. Dietaryoverlapexhibitedasignificantnegativecorrelationwithprey diversity(Figure3a;R=−.726,p = .011),andinclusionofdatafrom theseasonalstudiesgaveasimilarrelationship(Figure3b;R=−.899, p =<.001).Thus,atsamplingsiteswithrelativelyhighpreydiversity, thesalmonandbullheadsegregatedinresourceuseandfoodresourcepartitioningwashigh,whereaswhenpreydiversitywaslow, foodresourcepartitioningwasalsolow.Oursensitivityanalysesdid notaltertheresults,andthebestmodelwasalsothemodelincluding only prey diversity as explanatory variable (AICc=80.1;Table S4).Additionally,themodelremainsthesameusingbootstrapped data, corroborating a significant negative correlation between dietaryoverlapandpreydiversity(R=−.991,p < .001). Variablesotherthanpreydiversitycouldinfluencefoodresource partitioning(seeTable2),andmodelsimulationsafterforwardvariable TABLE3 Summaryofthebestlinearmixed-effectsmodel explainingvariationoffoodresourcepartitioningbetweenAtlantic salmonparrandalpinebullhead.Standarderror=SE Value SE t value p value Intercept 95.02 18.64 5.096 <.001 Preydiversity −75.21 23.75 −3.166 .011 FIGURE3 Relationshipbetweenpreydiversityandfoodresource partitioning(measuredasdietaryoverlap)betweenAtlanticsalmon parrandalpinebullheadat(a)elevensitesinRiverReisa,(b)withdata onseasonalvariationincluded(filledcircles),(c)betweenabundance ofsurfacepreyinthedietofAtlanticsalmonparrandfoodresource partitioning.Bothfoodresourcepartitioningandpreydiversityhave beenestimatedwiththehighesttaxonomicalresolutionoftheprey. Significantlineartrendswith95%confidencelimitsareshown 8 | SÁNCHEZHERNÁNDEZ Et al. selection(seeTableS3)suggestthattheabundanceofsurfacepreyin thedietofsalmonandfishabundance(salmon,bullhead,andArctic charrabundance)mayhavehadsomeinfluenceonfoodresourcepartitioning.Theabundanceofsurfacepreyinthedietofsalmongave a significant negative correlation with dietary overlap between the salmonandbullhead(Figure3c;R=−.671,p = .024). 4 | DISCUSSION Therewasanegativecorrelationbetweenpreydiversityanddietary overlapofsalmonandbullhead,supportingthehypothesisthathigh prey diversity may enhance food resource partitioning between sympatricspeciesandtherebyfacilitatetheircoexistence.Inpreviousstudies,highdietaryoverlapwasobservedbetweensalmonand bullheadatahomogeneousriversitewithalowdiversityofzoobenthos(Gabler&Amundsen,1999),whereasstrongdietarysegregation wasobservedbetweensalmonandEuropeanbullhead(Cottus gobio Linnaeus, 1758) in a more heterogeneous river that had relatively highdiversityofzoobenthos(Gabler,Amundsen,&Herfindal,2001). Thus,resourcepartitioningbetweensalmonandbullheadspeciesmay berelatedtobetween-riverdifferencesinpreydiversityandhabitat characteristics(Gabler&Amundsen,1999;Gableretal.,2001).The presentstudyrevealsthatfoodresourcepartitioningbetweensalmon andbullheadcanvarywithinariversystem,betweensitesatrelatively shortdistancesfromeachother,andbetweenseasonsatagivensite, withresourcepartitioningbeingstronglydependentonpreydiversity atdifferentsites.Studiesonotherspecieshavegivenindicationthat highpreydiversitymayenhancefoodresourcepartitioning(see,e.g., Hillebrand&Shurin,2005;Jiangetal.,2008;Martin&Garnett,2013; Zapataetal.,2005)andthatcompetitionforfoodishighwhenprey diversityislow(Barilietal.,2011;Hillebrand&Shurin,2005;Targett, 1981).Ourstudycorroboratesthesefindingsandsupportsthenotion thathighpreydiversitymaypromoteconsumercoexistencethrough foodresourcepartitioning. Itishypothesizedthatpreydiversityasitrelatestocompetition forfoodresourcesandpartitioningcouldhaveaninfluenceonsegregation and species coexistence in consumerswith similar trophic nicherequirements.Theoreticalconsiderationsthataddressrelationships between dietary overlap, competition, and coexistence posit thatcompetitionforcessympatricspeciestodivergeandsegregate inresourceuse(Schoener,1974,1989),theweakerspeciesmaybe excluded(e.g.,Eloranta,Knudsen,&Amundsen,2013;Nakanoetal., 1999;Schoener,1989),orecologicallysimilarsympatricspeciesmay converge and overlap in resource use (e.g., Cucherousset, Aymes, Santoul, & Céréghino, 2007; Keddy, 2001; Paterson etal., 2014; Wiens, 1993). These are seemingly contradictory standpoints. The first consideration encapsulates the competitive exclusion principle (Gause,1934;Hardin,1960)thathasbeenwidelyacceptedbymany inthescientificcommunity,andthesecond,althoughbeingmorecontroversial,hasalsoreceivedsomesupport(see,e.g.,Bengtsson,1991; Grant,1972;terHorst,Miller,&Powell,2010).Wesuggestthatthe apparentcontradictionscanberesolvedifpreydiversityistakeninto account.Ourreasoningisasfollows:Competitionforfoodmayresult FIGURE4 Schematicillustrationof thepotentialinfluenceofpreydiversity onresourcepartitioningbetweentwo stream-dwellingfishspeciesinsympatry (hereAtlanticsalmonparrandalpine bullhead).Forexample,ifpreydiversityis low,itisprobablethattherewillbestrong competitionbecausepreydiversityis insufficienttoallowsympatricconsumers tospecializeandsegregateinpreyuse | 9 SÁNCHEZHERNÁNDEZ Et al. ineitherhighandlowdietaryoverlapbetweensympatricconsumers dependingonpreydiversity.Ifpreydiversityishigh,thecompeting speciesmaysegregateinresourceusebyspecialization,forexample, exploitationofGlossosoma intermedium(Klapalek1892)bybullhead anduseofsurfacepreyandApatania stigmatella(Zetterstedt1840)by salmon.Underthesecircumstances,competitionresultsinresource segregationthroughalowdegreeofdietaryoverlap,aspredictedby classicnichetheory(e.g.,Jiangetal.,2008;Martin&Garnett,2013; Schoener,1989;Targett,1981).Thestrengthofcompetitionmaybe lowbecausethecompetingspecieshavethepossibilitytominimize negativeeffectsbysegregatingtheiruseoffoodresourcesviaspecialization.Ontheotherhand,ifpreydiversityislow,itisprobablethat therewillbestrongcompetitionbecausepreydiversityisinsufficient toallowsympatricconsumerstospecializeandsegregateinpreyuse (Figure4).Thiscomplieswithasituationthatleadstoatheoretical predictionthatcompetitionwillresultinhighnicheoverlapifthespeciesaresymmetricalintheircompetitiveabilities(Ågren&Fagerstrøm, 1984;Gilbert,2012;Keddy,2001),orifcompetitionisverystrong (Martin & Genner, 2009; Schoener, 1989; Wiens, 1993).Thus, the seeminglycontrastingconsiderationsabouthowcompetitionaffects dietaryoverlapmaynotbetrulycontradictory,butbothmaybevalid dependinguponthescaleofpreydiversity. Inadditiontopreydiversity,preyabundancemightbeafactorthat influencesthestrengthofcompetition(see,e.g.,Triplet,Stillman,& Goss-Custard,1999).Theintuitiveexpectationisthatcompetitionfor foodshouldbehigherwhenresourcesarescarcethanwhentheyare abundant.Althoughpreyabundancewasasignificantpredictorvariableintheforwardstepwiseprocedure,otherexplanatoryvariables were more influential (see Table2). In this regard, fish abundance couldplay aroleingoverningfoodresourcepartitioning.Thisisin agreementwithpreviousworks,demonstratingthatfishabundance canexacerbatecompetitionforfoodinfishassemblages(Elliott,1994; Engelhardetal.,2013).Itisimportanttonotethatthecorrelationbetweendietaryoverlapandfishabundancewasnotsignificant,sofish abundancemayoperatesynergisticallywithpreydiversitytoenhance food resource partitioning (see models including fish abundance in TableS3).Forexample,Barilietal.(2011)reportedthathighfishabundanceanddiversitycanpromotetrophicspecializationinsympatric species,therebyenhancingfoodresourcepartitioning,anddensity- dependentforagingbehaviormayoccurwhenresourcesarelimited (Sánchez-Hernández&Cobo,2013). Noteworthy, ourinterferences regarding the influence of density dependence on the competition forfoodshouldbetakenwithsomecautionbecauseouranalysesincludedtheoverallfishdensityregardlessoffishsize.Nevertheless,it isreasonabletopositthatfishabundancemayimpactonthemechanismsinvolvedinfoodresourcepartitioningasdiscussedearlier. Our study revealed that surface prey were strongly representedinthedietofsalmonatsiteswheredietaryoverlapbetween salmonandbullheadwaslowest,suggestingthatsurfacefeeding maybeacontributingfactorthatdrivesfoodresourcepartitioningbetweenstream-dwellingfishspecies(see,e.g.,Dineenetal., 2007;Sánchez-Hernández,Gabler, &Amundsen, 2016;Sánchez- Hernández,Servia,Vieira-Lanero,&Cobo,2013).Althoughsalmon and bullhead usually feed primarily on benthic invertebrates (Amundsen&Gabler,2008;Gabler&Amundsen,1999,2010),our studyclearlydemonstratesthatbullheadfeedlessonsurfaceprey thandosalmon,withtheproportionofsurface-driftforagersbeing substantiallyhigherinsalmonthaninbullhead(Sánchez-Hernández etal.,2016).Thus,thisstudycorroboratestheflexibilityofsalmon adoptingitsforagingmodesinrelationtobullhead.Theinference isthatbullheadhasapreferenceforforagingclosetothebottom, whereassalmonmayforagethroughoutthewatercolumnandcan adoptdifferentforagingmodestoovercomecompetitionwiththe co-occurringspecies,butthelackofdriftsamplingdidnotallowus toassesswhetherornotthisfeedingbehavioradoptedbysalmon ismotivatedbydriftavailabilityorfoodcompetitionwithbullhead. However,thisflexibilityislikelytobeinfluencedbypreyavailabilities(e.g.,Nakanoetal.,1999;Sánchez-Hernández&Cobo,2013), leadingtorelationshipsthatareinfluencedbybenthicinvertebrate diversityandtheavailabilityofsurfaceprey.Dielpatternsoffeeding andhabitatutilizationhavethepotentialtoinfluencefoodresource partitioning between sympatric species (e.g., Crow etal., 2010; Kronfeld-Schor&Dayan,2003;Sánchez-Hernándezetal.,2011). However,previousstudiesinRiverReisarevealednostrongsegregationinforagingtime(dielfeedingrhythms)andspace(habitat) betweensalmonandbullhead(Amundsen&Gabler,2008;Gabler& Amundsen,1999,2010),whichsupportsourmainconclusionthat preydiversityisthemaindriverofresourcepartitioninginthese twospecies.Still,thecapacitytoforageatthewatersurface(surface feeding) by salmon needs to be acknowledged as a spatial segregationinfeedingcontributingtotheobservedresourcepartitioningbetweenthetwomodelspecies. Preydiversityemergedasthestrongestpredictorofresourcepartitioningbetweensalmonandbullhead,althoughresourcepartitioning wasalsoinfluencedtosomeextentbysurfacepreyuse,andfishand preyabundances.Preydiversityandsurfacepreymayhaveoperated synergisticallytoenhancefoodresourcepartitioningbetweensalmon andbullhead.Additionalworkwillbeneededtoexploreandenhance ourunderstandingofhowtheinterfacebetweenaquaticandterrestrial ecosystems influences ecological processes, such as resource partitioning. ACKNOWLEDGEMENTS ThanksareduetoR.GuttormsenandS.Sandringforassistanceduring field work. We appreciate constructive comments from Dr. M. Jobling,whichconsiderablyimprovedthequalityofthemanuscript. We also thank two anonymous reviewers for valuable comments thathelpedimprovethemanuscript.Financialsupportwasprovided bytheNorwegianResearchCouncil.J.Sánchez-HernándezwassupportedbyapostdoctoralgrantfromtheGalicianPlanforResearch, Innovation,andGrowth2011–2015(PlanI2C,XuntadeGalicia). CONFLICT OF INTEREST Nonedeclared.