Ecology and Evolution. 2017;1–14.  | 1www.ecolevol.org

Received:25May2016  |  Revised:1May2017  |  Accepted:8May2017DOI: 10.1002/ece3.3142

O R I G I N A L R E S E A R C H

Plant diversity has contrasting effects on herbivore and parasitoid abundance in Centaurea jacea flower heads

Norma Nitschke1  | Eric Allan2 | Helmut Zwölfer3 | Lysett Wagner1 |  Sylvia Creutzburg1 | Hannes Baur4,5 | Stefan Schmidt6 | Wolfgang W. Weisser1

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium,providedtheoriginalworkisproperlycited.©2017TheAuthors.Ecology and EvolutionpublishedbyJohnWiley&SonsLtd.

1InstituteofEcology,Friedrich-Schiller-University,Jena,Germany2InstituteofPlantSciences,UniversityofBern,Bern,Switzerland3DepartmentforAnimalEcologyI,UniversityofBayreuth,Bayreuth,Germany4AbteilungWirbelloseTiere,NaturhistorischesMuseumBern,Bern,Switzerland5InstituteofEcologyandEvolution,UniversityofBern,Bern,Switzerland6BavarianStateCollectionofZoology(ZSM),Munich,Germany

CorrespondenceNormaNitschke,InstituteofEcology,Friedrich-Schiller-University,Jena,Germany.Email:N.Nitschke@yahoo.de

Present addressWolfgangW.Weisser,TerrestrialEcologyResearchGroup,DepartmentofEcologyandEcosystemManagement,SchoolofLifeSciencesWeihenstephan,TechnicalUniversityofMunich,Freising,Germany.

Funding informationDeutscheForschungsgemeinschaft,Grant/AwardNumber:WE3081/3-4andWE3081/17-2

AbstractHighbiodiversityisknowntoincreasemanyecosystemfunctions,butstudiesinvesti-gatingbiodiversityeffectshavemorerarelylookedatmulti-trophicinteractions.Westudiedatri-trophicsystemcomposedofCentaurea jacea(brownknapweed),itsflowerhead-infestingtephritidfruitfliesandtheirhymenopteranparasitoids,inagrasslandbiodiversityexperiment.Weaimedtodisentangletheimportanceofdirecteffectsofplantdiversity(throughchangesinapparencyandresourceavailability)fromindirecteffects (mediatedbyhostplantqualityandperformance).Todothis,wecomparedinsectcommunitiesinC. jaceatransplants,whosegrowthwasinfluencedbythesur-roundingplantcommunities (andwheredirectand indirecteffectscanoccur),withpottedC. jaceaplants,whichdonotcompetewiththesurroundingplantcommunity(andwhereonlydirecteffectsarepossible).Tephritidinfestationrateandinsectload,mainlyof thedominant speciesChaetorellia jaceae, decreasedwith increasingplantspecies and functional group richness. These effects were not seen in the pottedplantsandarethereforelikelytobemediatedbychangesinhostplantperformanceand quality. Parasitism rates,mainly of the abundant chalcidwaspsEurytoma com-pressaandPteromalus albipennis,increasedwithplantspeciesorfunctionalgrouprich-ness in both transplants and potted plants, suggesting that direct effects of plantdiversityaremostimportant.Thedifferentialeffectsintransplantsandpottedplantsemphasize the importanceofplant-mediateddirect and indirecteffects for trophicinteractionsatthecommunitylevel.Thefindingsalsoshowhowplant–plantinterac-tions critically affect resultsobtainedusing transplants.Moregenerally, our resultsindicatethatplantbiodiversityaffectstheabundanceofhighertrophiclevelsthroughavarietyofdifferentmechanisms.

K E Y W O R D S

chalcidwasps,grassland,Jena Experiment,Tephritidae,tri-trophicsystem

source: https://doi.org/10.7892/boris.106378 | downloaded: 17.3.2020

2  |     NITSCHKE ET al.

1  | INTRODUCTION

Severalstudieshaveshownthatplantdiversityaffects thediversityand abundance of other trophic levels (e.g., Scherber etal., 2010).However, themechanisms driving these effects remain unclear be-causeimpactsofplantdiversityontrophicinteractions,suchaspara-sitismorpredation,areonlyrarelyinvestigated(Ebeling,Klein,Weisser,&Tscharntke,2012).Pestcontrolbynaturalenemieshasbeenstudiedin agro-ecological research (e.g.,Bianchi,Booij,&Tscharntke, 2006;Menalled,Marino,Gage,&Landis,1999;Thies&Tscharntke,1999)where higher parasitoid efficiency, that is, higher parasitism ratesof insectherbivores, isoften found inmore structurally complexorspecies-richsystems(Andow,1991;Langellotto&Denno,2004;Priceetal.,1980).ThisisoftenreferredtoastheEnemies Hypothesis(Root,1973). However, most of these studies addressed isolated trophiclevelsordidnotconsidereffectsofplantdiversity.The fewstudiesthatdidexamineplantdiversityeffectsonparasitismusedplantcom-munitieswithveryfewspecies,thatis,amaximumdiversityofthreespecies.Morerecently,multi-trophic interactionshavebeenstudiedinexperimentallymanipulatedplantcommunitieswithlongerplantdi-versitygradients,whereplantdiversityeffectswereshowntocascadeupthefoodchain(e.g.,Ebelingetal.,2012,2014;Petermann,Müller,Weigelt,Weisser,&Schmid,2010).However,thesestudieswerenotabletolookindetailatthemechanismsdrivingthesecascadingdiver-sityeffects.

Bottom-upeffectsofbiodiversityontrophicinteractionsmaybecausedbybothdirecteffectsoftheplantcommunityandindirectef-fectsmediatedbychangestohostplantperformance.Plantdiversitycandirectlyaffecthighertrophiclevelsbyincreasingthecomplexityoftheolfactory,optical,andstructuralenvironment.Thismaymasktheodorsorvisualcuesthatherbivoresusetofindtheirhostplants(Coll&Bottrell,1994;Finch&Collier,2000;Randlkofer,Obermaier,Hilker,&Meiners, 2010)making the hosts less apparent to the herbivore.Particular plant speciesmay also cause this effect through associa-tionalresistance(Barbosaetal.,2009):forexample,similarlycoloredneighboringplantsmayattract insectsaway fromthehostplant,orvolatilesemittedbyneighboringplantsmayrepelherbivoresorreducetheirabilitytofindthehostplant.Diversityeffectsmaybedrivenbyanumberofsuchneighboringspeciesandthereforebyageneralin-creaseincomplexityinspecies-richcommunities.Alackofalternativehostsintheplantcommunity,asmayoccurin(diverse)communitiesoftaxonomicallydistantplantspecies,mayalsoreducefoodsupplyforherbivores and therefore herbivory rates on particular target plants(Jactel&Brockerhoff,2007;Root,1973).

As well as reducing herbivore abundance, high plant diversitycouldalsoreducetheefficiencyofparasitoids infindingtheir insecthostsbyincreasingstructuralcomplexityandproducingodorblendsthatmaskthefocalplant(Andow&Prokrym,1990;Bukovinszky,Gols,Hemerik, van Lenteren, &Vet, 2007; Gols etal., 2005; Randlkofer,Obermaier,&Meiners,2007).Alternatively,predatorsandparasitoidscouldbenefitfromplantdiversityifdiverseplantcommunitiesprovideagreater rangeof foodsources,suchasmorefloral resources (nec-tarandpollen)forparasitoids(Araj,Wratten,Lister,&Buckley,2008;

Lavandero,Wratten,Didham,&Gurr,2006)oragreaterdiversityofpreyforgeneralistpredators(Root,1973).Parasitoidscouldalsooper-atemoreefficientlywhenherbivoreabundanceislow(Ebelingetal.,2012),aswouldbeexpectedindiverseplantcommunities.However,lowerherbivoreabundancemayalsodecreaseparasitismrates (e.g.,White&Andow,2005)ifpatchtenuretimeislongerinhigh-densitypatches, as predicted by several patch time allocationmodels (VanAlphen&Bernstein,2008).Analysescontrollingforherbivoreabun-dancearenecessarytotesttheseideas.

In addition to these direct effects of the plant community onhighertrophiclevels,plantdiversitycanalsoindirectlyaffectherbivoreandpredatorcommunities.Indirecteffectsarisewhenplantdiversityinfluencesthegrowthandnutrientlevelsofhostplants(e.g.,Nitschkeetal.,2010;Roscher,Kutsch,&Schulze,2011),whichinturnaffectsinsectherbivores(Awmack&Leather,2002;Mattson,1980).Wecon-sidertheseeffectstobemoreindirectthaneffectsofplantdiversitymediatedbystructural,odor,or resourcediversitybecause theyarecausedbyaneffectoftheplantcommunityonindividualplants,me-diatedbychangesincompetition,whichinturnaffectshighertrophiclevels. In contrast, changes in structural, odor, or resource diversityoccurasadirectconsequenceofchangedplantcommunitydiversity.Areductioninindividualplantperformancewithincreasingdiversityislikelytoreducetheavailabilityofresourcesforherbivoresingeneral,suchasby reducing thenumberof flowerheads for flower feedingherbivores.Theperformanceofindividualplantsmightbereducediftheysuffermorecompetition:forinstance,planttissuenutrientlevelsmaybereducedindiversecommunitiesduetomoreefficientnutrientuseinspecies-richassemblages(vanRuijven&Berendse,2005)and/ordueto increased lightcompetition,whichcausesplants to investmoreinstructural,carbon-richtissues(Hirose&Werger,1995).Thelowerplantqualitycouldreducetheabundanceorperformanceofher-bivoresindiverseplantcommunities.Complexeffectsmayalsooccur;forexample,Kigathi,Weisser,Veit,Gershenzon,andUnsicker(2013)showedthatplantsmaychangetheiremissionofvolatilecompoundswhengrowingincompetitionwithotherplants,whichhaseffectsontheattractionofherbivoresandtheirnaturalenemies.Overall,effectsonthethirdtrophiclevelarelikelytobeweakerastheyareevenmoreindirect(Kagata,Nakamura,&Ohgushi,2005;Scherberetal.,2010).

To separate these direct and indirect effects, we used trans-planted and potted host plants (the common knapweed,Centaurea jacea) placed into experimental plant communities differing in spe-cies richness.We thenanalyzed the responseof its flowerhead in-festingTephritidaeandthehymenopteranparasitoidsattackingthoseTephritidaetothediversityoftheplantcommunity.Hostplantsthatweretransplanted intoexperimentalplantcommunities (transplants)interactedwith the surroundingplant communityover anumberofyears,bothaboveground(mainlylightcompetition)andbelowground(root/nutrientcompetition).Plantdiversitycanhavebothdirectandindirecteffectsontheherbivoreandparasitoidcommunityofthesetransplants.Wecomparedresponsesonthese transplantswithpot-tedplantsthatwereplacedinsidetheexperimentalplantcommunitiesduringthestudyperiodonly(pottedplants)andwhichdidnotinteractbelowgroundwith the surrounding plant community.Aboveground,

     |  3NITSCHKE ET al.

theremayhavebeenlimitedinteraction(shortperiodofpotentiallightcompetition)withotherplants in thecommunity. In this case,plantdiversitycanhaveonlydirecteffectsonhighertrophiclevelsbecauseindirecteffectsmediatedbychangesinplantperformanceorqualityareexcluded.Otheraboveground interactionsbetweenplants, suchascommunicationviavolatileorganiccompounds(VOCs)couldoccurinbothhostplanttypes;however,wedidnot investigatethemasasourceofdiversityeffectsasthiswouldrequireseparateexperiments.

Wehypothesizethatherbivoreabundancedependsbothonindi-rect (host plant quality/performance) and direct effects (communitycharacteristics) of plant diversity.We therefore expect to find thatplant diversity reduces herbivore abundance in both types of hostplants althoughwe might anticipate stronger effects in transplantswherebothdirectandindirecteffectsoperate.Totestforthemech-anismsdrivingpotentialdirectandindirecteffects,wecanmodeltheeffectofotherplantcommunityparametersontheherbivores.Ifre-sourcedensityeffectsareimportant,thenwewouldexpectthatthepresenceofalternativehosts inthecommunity increasedherbivory.Ifapparencydrivestheeffects,thenameasureofapparency,suchastheheightofhostplantsrelativetotherestofthecommunity(whichindicateshoweasyitisforinsectstofindtheirhostusingvisualcues),shouldsignificantlyincreaseherbivory.Structuralheterogeneitymightreduce herbivory, which means that herbivory is lower when LAI,whichcanserveasaproxyforstructuralcomplexity, ishigh.Totestfor indirecteffectsmediatedbyhostplantperformance,wecan in-cludeameasureof resourcedensityperhostplant.Forparasitoids,we hypothesize indirect effects to be ofminor importance and ex-pect positive direct effects of plant diversity.We should thereforeseesimilareffectsofplantdiversityonparasitoidsinbothtransplantsandpottedplants.Parasitoidsarelikelytorespondstronglytostruc-turalcomplexity(strongeffectsofLAIwouldbehypothesizedinthiscase)andtoresourcedensity(quantityoftephritidhosts).Wetestforthese effects using a large grassland diversity experiment, the Jena Experiment(Roscheretal.,2004).TheJena Experimentcontainsexper-imentalplantcommunitiesof20×20m2,whichdifferinplantspeciesrichnessandthenumberofplantfunctionalgroups.Weanalyzeddatafromexperimentalplantcommunitiescontaining1–8speciesand1–4plantfunctionalgroups(i.e.,grasses,legumes,smallherbs,tallherbs).

2  | MATERIALS AND METHODS

2.1 | The study system

Centaurea jacea L. s. l. (brown knapweed; Asteraceae) is native toEurasiaandcommonthroughoutGermany.The long-livedperennialhemicryptophyte (plants with overwintering buds at soil level) re-emergesinspring(Press&Gibbons,1993)producingvegetativesiderosettes, flowers, and fruits between June andOctober (Jongejans,de Kroon, & Berendse, 2006).C. jacea flower heads arewidely at-tackedbyTephritidae,anabundantfamilyofDipterathatmainlyin-habitfruitsorotherseed-bearingorgansoffloweringplants(White,1988). Six species of Tephritidae,with flight periods betweenMayandSeptember,areassociatedwithC. jaceainGermany.Fourofthem

have a narrow host range (i.e., eithermonophagous onC. jacea or usingafewCentaureaspeciesonly(Merz,1994;White,1988):Acinia corniculata (Zetterstedt),Urophora quadrifasciata (Meigen),Urophora jaceana(Hering),andChaetorellia jaceae(Robineau-Desvoidy),thelat-tertwoarelikelytobemonophagousonC. jaceainthestudyarea(HZ,pers.obs.).Twospeciesareassociatedwithmorethan15compositehost plant species of different genera (Merz, 1994):Acanthiophilus helianthi(Rossi),Chaetostomella cylindrica(Robineau-Desvoidy).TheseTephritidaehavedifferent foragingbehaviorsrangingfromdestruc-tivefeedingontheflowerhead(C. jaceae)toinducingcomplexwoodygalls in the capitulum (U. jaceana); however, detailed information isnot available for all potentiallyoccurring species.Parasitoidsof thefamilies Eurytomidae, Pteromalidae, Eulophidae (all Chalcidoidea),Braconidae, and Ichneumonidae attack these flower head phy-tophagesingreatnumbers(Dempster,Atkinson,&Cheesman,1995;Varley,1947;Zwölfer,1988).For tephritidhosts,chalcidwaspsarethemajorparasitoidsandthesehaveabroadhostrange(seeFigureS1 for a potential interaction web based on Tephritidae attacking C. jacea).Sincethestudysiteismowntwiceayear,Tephritidaeandtheirparasitoids,whichcommonlyoverwinterinflowerheadsofthehost plant, recolonize the experimental field site every year fromsourcepopulationsinthesurroundingmeadows.

2.2 | Experimental design

In order to investigate responses of the second (Tephritidae) andthird trophic levels (parasitoids) to plant diversity, the study wasconductedin76uniqueexperimentalplantcommunitiesrepresent-ingagradientofplantspecies(1–16)andplantfunctionalrichnesslevels(1–4).Onplotsof20×20m,setinthefloodplainoftheriverSaale (Jena, Thuringia, Germany), plant communities were estab-lished in 2002 from a pool of 60 grassland species, representingthesurroundingArrhenatherioncommunity.Thegrasslandspecieswereassignedtofourplantfunctionalgroups(legumes,grasses,tallherbs, and small herbs) andmixturesof1,2,4,8, and16 specieswere created by randomly selecting species from the pool of 60.Eachplantspeciesrichnesslevelwasreplicatedon16plots,exceptfor the 16-species communities (14 plots); additionally, four plotsweresownwithall60species(fordetails,seeRoscheretal.,2004;The Jena Experiment and Table S1). The design also manipulatedfunctional group richness tobeasorthogonal aspossible toplantspeciesrichness:thatis,thereare8-or16-speciesplotswithonlyoneortwofunctionalgroupspresent.Experimentalplotswerear-rangedinfourblocks,mostlytoaccountforthechangeinsoilcon-ditionswith increasingdistance to the river (Roscheretal.,2004).The site ismanagedas a typical haymeadowwithvegetation cutandremovedtwiceayear(beginningsofJuneandSeptember).Thediversitygradientismaintainedbyregularweeding.Forthecurrentexperiment, two monocultures with low target plant cover wereomitted. Description of the insect community is based on collec-tionsacrossagradientof1–16plantspecies,whereasanalysesofinsectresponsesarebasedoncollectionsacrossagradientof1–8plantspecies(seebelow).

4  |     NITSCHKE ET al.

WeusedtwotypesofC. jaceahostplantstostudytephritidinfes-tationandparasitismrates.Transplantswereplantedintotheexperi-mentalplotsin2005andastheyinteractedwiththesurroundingplantcommunity,bothdirectandindirecteffectsofplantdiversitycanbeexpected.Additionally,weusedpottedplants thatwereplaced intotheexperimentalplotsinthestudyyear2008.Theseplantsdidnotin-teractatallwiththesurroundingplantcommunitiesbelowgroundandwouldhaveexperiencedonlyminimalabovegroundcompetition.Onlydirect plant diversity effects can be expected in the potted plants.AbovegroundinteractionsbetweenplantsviaVOCscanoccurinbothplanttypes.

Transplants were grown from seed (supplier: Rieger-HofmannGmbH, Blaufelden-Raboldshausen, Germany) and planted into ex-perimental plots 3years after establishment of the experiment (i.e.,in 2005). Five plantswere planted in each plot, arranged in a rowand spaced 25cm apart. Transplants were regularly mown, alongwith the rest of the plant community.We found that their survivalstronglydecreasedwithincreasingplantspeciesrichness(andpartlywithfunctionalgrouprichness),totheextentthatnoneofthetrans-plants survived in60-speciesplots andonly about30% survived inthe16-speciesplots (seeNitschkeetal.,2010andFigureS4).Sinceplant species richness also decreases transplant biomass and thenumberof flowerheadsonplants (Nitschkeetal.,2010,FigureS4),theproportionofsurvivingplantswithflowerheadswasevenlower(<20% in2007,<3% in2008,TableS2) in the16-speciesplots.Wethereforeexcludedthesecommunitiesfromtheanalysisandsampledtransplantsonlyinplotswith1–8plantspecies.HarvesttookplaceinAugust(week35)2007and2008,shortlybeforemowingofthefieldsite(week36).Werecordedthetotalnumberofflowerheads(flow-eringandbuds)perplanttogetherwithitsheight.In2007,theflowerheads(includingthosenotyetflowering)werecutandalltheflowerheadsofagivenplantwerestoredtogetheratroomtemperatureuntilinsectsemerged;atthispoint,allemerginginsectswereidentified.Wethendissected10%ofthefloweringflowerheadsperplant(i.e.,thoseat themost advancedphenological stage), butwe always dissectedat leastfiveflowerheads,whichcouldbemorethan10%onplantswithfewflowerheads.Both,unemergedinsectsandemptypupaeofemergedinsectsaredetectedbydissectionandgiveaclearmeasureof the number of infestations that occurred per flower head (insect load).Moreover, dissectionof the single flowerheadsallowed for aprecisedeterminationoftheproportionofflowerheadsthatwerein-fested(tephritid infestation rate).Thesetwotephritidresponsesarenoteasilyderivedfrompureemergencedataontheplantlevel.However,theparasitoidcommunityiswellrepresentedbytheemergedinsects,andasweassumethattherearenodifferencesinemergencesuccessbetweenhostsandparasitoids,wedefinedtheparasitism rate tobetheproportionofemergedinsectsthatwereparasitoids.

In2008,westreamlinedandstandardizedourdatacollectionbyonly collecting 10% (and at least five) of the flower heads of eachtransplant (those at themost advanced phenological stage) and bystoring flower heads individually at room temperature until insectsemergedandcouldbeidentified.Wethendissectedallflowerheadsandadditionallyrecordedtheinsectswhichdidnotemerge.

Pottedplantscamefromamonocultureplot(3.5×3.5m)estab-lishedin2002onthesamefieldsite.Plantsweredugupinmid-May2008andtransferredto3-literpots(threeplantsperpot)filledwithsubsoil from the field site.After fertilizing once (COMPOBlaukorn,14%N,7%P2O5,17%K2O2,0.02%B,9%S,2%MgO),barkmulchwasappliedtoreducewaterlossfromevaporation.Inmid-June(week25),allpottedplantswerecut3cmaboveground,atthesametimeastheexperimentalplotsweremown.PottedplantswereplacedintheexperimentalplantcommunitiesatthebeginningofJuly2008(week28)andwerewateredwhenrequired.Threepotswereplacedoneachplotandweattemptedtoinsurethesamesizedistributionofplantsoneachplot.Plantsweredividedintothreesizeclassesbasedontheirnumberofflowerheads(0–2,3–5,and≥6flowerheadsperpot)andweplacedoneplant fromeach size class into eachplot. Potswerearranged in a rowand spaced25cmapart.We chose to add threepotsperplotinordertomatchthenumberoftransplantsstillsurviv-ingontheplots:theinitialfiveplantsperplotin2005haddecreasedtoaboutthreeonaverageby2008.Priortoset-upintheplots,anyflowerheadsthathadregrownsincecuttingwereremovedfrompot-tedplantstoexcludeanypreviousflowerhead-infestation(seeTableS2).Pottedplantshadthereforegrownunaffectedbythecommuni-ties intowhich theywereplaced.Pottedplants remained in theex-perimental plant communities for 7weeks, before being harvested.Aboveground interactions (e.g., shading by neighboring plants) overthe courseofonegrowing seasonareexpected tobeofminor im-portanceforpottedplantperformance,comparedtotheimpacttheseinteractionshadonthetransplantsoveranumberofyears(Nitschkeetal.,2010).Atharvest,werecordedthetotalnumberofflowerheads(floweringandbud)perpot, and theirmaximumheight in the field.Allflowerheadsofapot(uptoamaximumof20)werecollectedandstored individually at room temperature until insects emerged.Wethendissectedtheflowerheadstoidentifytephritidsandparasitoidsthathadnotemerged.

Weuseddatafromseparatelystoredflowerheadstoderivetro-phic relationships between the different species ofTephritidae andparasitoids(fordetailsonidentificationandassignmentofparasitoidstohostspecies,seesupplement,p.1).

2.3 | Response variables for the higher trophic levels

Wecalculatedaseriesofvariablesfromourdissectionandemergencedata.Wecalculatedtephritid infestation rateastheproportionofdis-sectedflowerheadsthatwereinfested,andinsect loadasthenumberofinfestationsperdissectedflowerhead.Thetotalnumberofinfes-tationswasthesumof tephritidandparasitoid individuals (becauseeachparasitoidmusthaveemergedfromatephritid),aswellaspupae.In2007,parasitism ratewasdefinedastheproportionofallemergedinsectsthatwereparasitoids,while in2008,parasitism ratewasde-fined as the proportion of hosts thatwere found to be parasitizedfollowingflowerheaddissection.Exceptforthreecases,parasitoidswereallsolitaryandcouldbeunequivocallyrelatedtotephritidhosts.Inorder to compare thepottedplantswith the transplants,weex-cludedthepottedplantdatafromthe16-speciesplotsandanalyzed

     |  5NITSCHKE ET al.

all responses across the 1–8 plant species gradient for all datasets(transplants2007and2008,potted2008).

2.4 | Variables mediating diversity and plant functional group effects on higher trophic levels

Inordertoidentifypotentialmediatorsofdiversityeffects,weusedaseriesofvariablesascovariatesinourstatisticalmodels.LAI(LeafAreaIndex)isameasureofabovegroundspaceuseandlightpenetration(Welles&Norman,1991)andcommonlyincreaseswithplantspeciesrichness in experimental plant communities (Spehn, Joshi, Schmid,Diemer,&Korner, 2000).LAI is derived frommeasurements of “alllightblockingobjects”inacommunityandreflectsstructuralcomplex-ity(Rutten,Ensslin,Hemp,&Fischer,2015);therefore,weusedthisparameterasproxyforhabitatcomplexity.Sincehostplantapparency(theprobabilitythatahostplantisfoundbyitsherbivore(Endara&Coley,2011))canaffectinsectherbivory(e.g.,Castagneyrol,Giffard,Pere,&Jactel,2013),we includedthevariablerelative heightofthehost plant as an apparencymeasure. Relative height indicates howeasyitisforinsectstofindtheirhostplantusingvisualcues,asaplantwhichismuchtallerthanitsneighborswillbeeasytofind,whereasonethatisshorterwillbehardtolocate.ThereissomeevidencethatTephritidae,especiallythosewithanarrowhostrange,usevisualcues(especiallyshapes,butalsosizeandtoa lesserextentcolor) tofindtheirhosts.Intests,femaletephritidstendedtopreferovipositionsitemodelssubstantiallylargerthannaturalsites(e.g.,fruits).Specifically,femalesofonespeciesofUrophoraandChaetorellia(twoofthegenerainoursamples)wereshowntobemostattractedtowardsophisticatedvisualmimicsoffloralbudsoftheirhostplant(Díaz-Fleischer,Papaj,Prokopy,Norrbom,&Aluja,1999).Here,wefocusonthevisualap-parencyof thehostplants aswewereable tomeasure this.Othertypesofapparency,suchaschemicalodorapparencycouldalsoplayarolebutcouldnotbequantifiedhere.PlantcommunityLAIandheight weremeasuredonallplotstwiceayearduringpeakstandingbiomass.LAIwas recordedwithanLAI-2000plantcanopyanalyzer (LI-COR)usinghighresolutionandaviewcapmasking45°oftheazimuthto-wardstheoperator.Alonga10mtransect,10measuresweretakenin the plant community (at 5cm above ground) and combined intoonemeanvalueperplot(Weigeltetal.,2010).Floralandvegetativeheightsweremeasured everymeter along a10m transect and av-eraged (Weigeltetal.,2010).Here,weusedmeasures taken in lateAugust2007and2008.Thevariablerelative heightperfocalplantin-dividual (orpergroupofplants inapot)wascalculatedbydividing C. jaceamaximumheight in the fieldby the largervalueof the twoheightmeasures (vegetativeor floral)of the respectiveplot.Toac-countforresourceconcentrationeffects,weincludedtephritid infes-tation rate as ameasureof resourcedensity for theparasitoids.AssomeoftheTephritidaefeedonseveralAsteraceaespecies,wealsoincludedthepresence of Asteraceaeinthecommunitiesinourmodelstoaccountforpotentialspilloversfromrelatedhosts.Toaccountforpotentialindirecteffectsofdiversity,mediatedbyhostplantperfor-mance,we included thenumber of flower heads per host plant as ameasureofresourcedensityforTephritidae.

2.5 | Statistical analysis

All analyses were conducted using the R statistical software (RDevelopmentCoreTeam2010,Version2.12.1).Wetestedforeffectsofplantdiversityandplantfunctionalcompositionon(i)thesecond(tephritids)and(ii)thethirdtrophiclevel(parasitoids)usingmixedef-fectsmodelsfittedwiththe“lme4”package (Bates&Sarkar,2007).Responsevariableswere(i)tephritid infestation rateandinsect loadand(ii)parasitism rate. Inordertodisentanglethedirectandindirectef-fectsofplantdiversity,weanalyzedeachresponsevariableforbothtransplants(directandindirecteffects)andpottedplants(directonly).Duetolowtransplantsurvivalinthe16-speciesplots(Nitschkeetal.,2010),weexcludedthemfromtheanalysisofthetransplantsand,forcomparability,fromtheanalysisofthepottedplants.Weusedthreedifferentdatasets:2007 transplants (1–8),2008 transplants (1–8),and2008 potted plants (1–8)(seeTableS3forthenumberofplotsineachdataset).

Foreachof the responsevariablesanddatasets,we fitteda fullmodelwithplotasarandomfactor(toaccountforthenesteddesignofourstudy,withseveralmeasuresfromeachplot).Wethencarriedoutatwo-stepanalysis.Inthefirststep,wetestedforbottom-upef-fectsofplantdiversityandfunctionalcompositionandincludedfixedeffectsforplant species richness(logtransformed),functional group rich-ness,andthepresence of particular functional groups(legumes,grasses,andsmallandtallherbs).Asitwasnotpossibletofitthepresenceofallfourfunctionalgroups,togetherwiththenumberoffunctionalgroups,wedeterminedwhichvariablesshouldbeincludedforeachresponse.Foreachresponse,wefittedfivemodelswitheachofthefivefunc-tionalgroupvariablesandexcludedthevariablewhosemodelhadthelargestAICvalue(indicatedinTable1).Toaccountforspatialvariationintephritidandparasitoidcommunities,wealso includedblockasafixedeffect(wetreateditasfixedbecauseithasonlyfourlevelsandestimatingvarianceforvariableswithfewlevelsisunreliable).ThefullmodelinRsyntaxisasfollows(shownhereforthecasewheremodelswithgrasspresencehadthehighestAIC):

In a second step, we included several covariates to determinewhetherthesemediatedtheeffectofanyofthedesignvariables.Aspotentialmediatorsofdiversityeffects(speciesrichnessorfunctionaldiversity),weincludedcommunity LAI(aproxyforstructuralcomplex-ity),therelative maximum heightoftransplants/pottedplants(asimpli-fiedmeasureofC. jaceaapparency;i.e.,thelikelihoodofahostplantbeingfoundbyitsherbivore)andthenumber of flower headsperhostplant in theanalysisof tephritid infestation, and tephritid infestation rateintheanalysisofparasitismrate(asmeasureofresourcedensity).Asmediatorsoffunctionalgrouppresence(tallandsmallherbs),wein-cludedthepresence of Asteraceaeinthecommunitytoaccountforpo-tentialspilloversfromrelatedplantspecies.BoththetephritidspeciesandtheirassociatedparasitoidscanalsouseotherhostplantswithintheAsteraceae(withalmostnoassociationsoccurringwithplantsof

y ∼ block + species richness + number of functional groups

+ legumes + tall herbs + small herbs + (1|plot ID)

6  |     NITSCHKE ET al.

TABLE 1 EffectsofcommunityandC

enta

urea

jace

acharacteristicsontephritidandparasitoidresponses

     |  7NITSCHKE ET al.

otherfamilies).Thefullmodelfromthesecondstepisasfollows(againassuminggrasspresencewasdropped):

In all cases, fullmodelswere simplified by progressively remov-ingnon-significantterms,andcomparingmodelswithlikelihoodratioteststoproduceaminimaladequatemodel(finalmodel).Significanceoftermsinthefinalmodelwasassessedbyseparatelyremovingtermsfromthatmodel;however,blockwasalwaysretainedinthemodels.For infestation rate andparasitism rate (binary responses),weusedageneralized linear mixed model with binomial error distributions.Other responsevariableswere transformed if necessary inorder tomeettheassumptionofthemodel(indicatedinTable1).

FigureswerecreatedinRusingthepackages“effects”(Fox,2003),“plotrix”(Lemon,2006),and“gplots”(Warnes,2010).Dataforfigureswerederived from the statisticalmodel using thepackage “effects”(version3.1-2).Values for responsesthatweresignificantlyaffectedbythepredictortobeshowninthefigurecamefromthefinalstep-2modelandwerethereforecorrectedfortherandomeffectsandanyother significant fixed effects.Where the responsewas not signifi-cantlyaffectedbythepredictortobeshowninthefigure,wecouldnot use values from aminimal adequatemodel and therefore tooktheminsteadfromasimplifiedmodelofthetype:response ~ block + predictor + (1|plot ID). In thesecases,valueswereonlycorrectedforblock and randomeffects.Graphical illustration of the trophic rela-tionshipswasproducedinR(package“bipartite,”Dormann,Gruber,&Fründ,2008).Forananalysisofspeciesco-occurrences,seesupple-ment.Significantresultsarereportedasmean±SEderivedfromthefinalstep-2model.

3  | RESULTS

3.1 | Description of the insect community in flower heads

Wefound855tephritidindividualsintotalandidentifiedfourteph-ritidspeciesattackingC. jacea.ThestenophagousChaetorellia jaceae (Robineau-Desvoidy, 1830) accounted for 93.5% and 84.4% of in-dividuals in2007and2008, respectively.Urophora jaceana (Hering,1935),agall-inducingstenophagoustephritid,madeupbetween1%and3%ofindividuals,whereasUrophora quadrifasciata(Meigen,1826)(Figure1)hadsimilarlylowabundancein2007,butwasmorecommonin2008,with12.8%of individuals.TheeuryphagousAcanthiophilus helianthi(Rossi,1794)wasveryrare(<2%ofindividuals)inbothyears(Figure2,TableS4).

Wecollected510parasitoidsin2007and290in2008.Weiden-tifiedimaginesoffiveendoparasiticHymenopteraspecies:fourchal-cidwasps (Aprostocetus forsteri (Walker, 1847), Eurytoma compressa (Fabricius, 1794), Pronotalia trypetae Gradwell, 1957, Pteromalus

albipennis Walker, 1835), and one braconid (Bracon subrugosus Szepligeti, 1901). E. compressa andP. albipennis dominated the par-asitoidcommunity, togetheraccounting foralmost90%of imaginescollected in either year. However, their relative proportions varied: E. compressarepresented54.5%and24.1%in2007and2008,respec-tively,whereasP. albipenniswaslessnumerousin2007(34.9%)thanin2008(64.1%)(Figure2,TableS4).Thetwogregariousparasitoidspe-cies(A. forsteriandP. trypetae)wererareandasingleindividualoftheectoparasiticchalcidwaspTorymus cf.chloromeruswas found in the2008collection.

Thetwomostabundantparasitoidspecies,E. compressaandP. al-bipennis,attackedallofthefourtephritidhostspeciesandweremost

y ∼ block + LAI + relative height + number of flower heads

+ Asteraceae presence + species richness+

number of functional groups + legumes+

tall herbs + small herbs + (1|plot ID)

F IGURE  1 Urophora quadrifasciataonbrownknapweed(Centaurea jacea).Picture:SebastianKönig

F IGURE  2 CompositionofthehostandparasitoidcomplexfoundinCentaurea jaceaflowerheads.Insectimaginesabundancesin2007and2008arelistedbelowthespeciesnamesinthelegendontheright.Combineddatafromtransplants(1–8speciesplots)andpottedplants(1–16speciesplots)

2007 2008 2007 2008

Per

cent

age

of to

tal a

bund

ance

0

20

40

60

80

100

Chaetorellia jaceae

Urophora quadrifasciata

Urophora jaceana

Acanthiophilus helianthi

Eurytoma compressa

Pteromalus albipennis

Pronotalia trypetae

Bracon subrugosus

Aprostocetus forsteri

Torymus spec

Hosts Parasitoids

387

14

12

1

278

178

41

13

0

0

368

56

6

6

70

186

17

12

4

1

2007 2008

8  |     NITSCHKE ET al.

abundantonthecommonesttephritidChaetorellia jaceae. E. compressa wasonlyoncefoundparasitizingA. helianthi or U. quadrifasciataandP. albipennis was also only once found onU. jaceana. The braconid, B. subrugosus, mainly parasitizedC. jaceae and only one interactionwithU. quadrifasciata was recorded. The gregarious P. trypetae wasonlyeverfoundinC. jaceae.Theobservedlinksbetweenhostandpar-asitoidspeciesaresummarizedinFigure3.Besidesthetephritidhostsandtheirparasitoids,anumberofothertaxaassociatedwithCentaurea jacea occurred in the samples (Chloropidae (Diptera, 67 imagines),Cecidomyiidaelarvae(Diptera,uncounted),Cynipidae(Hymenoptera,2imagines),Lepidopteralarvae(8),Cortinicara gibbosa(Herbst1793),(Coleoptera,11imagines)).

3.2 | Responses of the second trophic level (Tephritidae) to plant diversity

Only plant functional group presence affected the herbivorousTephritidaeintheanalysisofpottedplants.Insect load(i.e.,thenum-berofinfestationsperdissectedflowerhead)waslowerinthepres-enceofgrasses(1.5±0.1individuals)thanintheirabsence(1.7±0.1individuals) (χ2=3.83,p = .050, Figure4), both in the first and sec-ond stepof the analysis (i.e.,without andwith covariates,Table1).Tephritid infestation rate(i.e.,theproportionofdissectedflowerheadsthatwere infested)wasnotaffectedbyanyvariables in thepottedplants.Plantdiversityandfunctionalcompositionhavethereforefewdirecteffectsontheherbivorecommunity.

Largereffectsofplantdiversityonherbivoreswere foundwhenthetransplantdatafrom2007and2008 (wherebothdirectand in-direct effects can be expected)were analyzed. In 2007, insect load decreasedwithincreasingplantspeciesrichnessfrom1.9±0.1indi-vidualsperheadinmonoculturesto1.4±0.1individualsperheadin8-speciesmixtures(χ2=6.40,p = .011,Figure4);however,nosignifi-canteffectwasdetectedin2008.Thecovariatesdidnothavesignifi-canteffectsoninsect loadineitheryear(Table1).Tephritid infestation rate respondedmore to functional diversity and decreasedwith in-creasingplantfunctionalgrouprichnessbyca.70%in2008(χ2=7.64,p = .006)butnotin2007(Table1,Figure5).Inthesecondstepoftheanalysis,thenumber of flower headswassignificantinboth2007and2008(2007:χ2=14.73,p < .001;2008:χ2=9.75,p = .002).Tephritid infestation rate increased with increasing resource density in bothyears(Table1,FigureS2).Theseresultssuggestthatplantspeciesandfunctionalgrouprichnesseffectsarestrongerwhenbothdirectandindirecteffectsareoperating.

3.3 | Response of the third trophic level (parasitoids) to plant diversity

Plant species richness and functional grouppresencebothaffectedparasitoidcommunities in thepottedplants. In the first stepof theanalysis (without covariates), the presence of legumes reduced the

F IGURE  4  Influenceofplantspeciesrichnessandgrasspresenceinexperimentalplantcommunitiesoninsect loadperflowerheadinthethreedatasets.Significanceinfinalstep-2modelsisabbreviated:*p ≤ .05,n.s.p > .05. Means±SEderivedfromfinalstep-2orsimplifiedmodels(seeMethods)

1.2

1.4

1.6

1.8

2.0

2.2

Plant species richness1 2 4 8

Inse

ct lo

ad

Transplants 2007 Transplants 2008 Potted plants 2008

* n.s. n.s.

1.2

1.4

1.6

1.8

2.0

2.2

– Grasses + Grasses

n.s. n.s. *

F IGURE  3 Host–parasitoidinteractionwebfromcombineddataof2007and2008.Speciesarerepresentedbyarectangle,widthofrectangletherebyreflectingtheirabundances.Thelowerrectangleslisthostspecies,andtheupperlineparasitoidspecies.Thesizeoflinesconnectingthetwotrophiclevelsrepresentstheinteractionstrength.Connectionsinblackrefertointeractionsbasedonsingleobservations

B. s

ubru

gosu

s

P. a

lbip

enni

s

P. t

rype

tae

E. c

ompr

essa

U. q

uadr

ifasc

iata

A. h

elia

nthi

C. j

acea

e

U. j

acea

na

Parasitoids

Hosts

     |  9NITSCHKE ET al.

parasitism rate (χ2=10.86, p=.001, Table1, Figure6: 35.8±5.0%withlegumesvs.65.0±5.0%without),andplantspeciesrichnesshadamarginally significanteffect (χ2=3.73,p=.053).Whencovariateswereincludedinthestep-2model,plantspeciesrichnesssignificantlyincreasedparasitism(χ2=4.42,p=.036,Table1,Figure6).Thenum-ber of flower headsalsoreducedparasitism rateinthismodel(χ2=4.19,p=.041,Table1, FigureS3).We therefore findevidence for directeffectsofplantspeciesrichnessonparasitoidsbutnotonherbivores.

Plantspeciesandfunctionalgrouprichnessbothaffectedpara-sitism rateinthetransplants,wheredirectandindirectplantdiversityeffects could occur. Plant functional richness increasedparasitism rates in 2007 from 54.2±3.2% in plots with a single functionalgroupto90.4±6.1%inplotswithfourfunctionalgroups(χ2=3.83,p=.050, Table1, Figure6). However, in 2008 plant species rich-ness, rather than functional group richness, increased parasitism rate(from14.4±3.4%inmonoculturesto48.8±6.8%in8-species

F IGURE  5  Influenceofthenumberofplantspeciesandfunctionalgroupspresentinexperimentalcommunitiesontephritidinfestation rateinthethreedatasets.Significanceinfinalstep-2modelsisabbreviated:**p ≤ .005,n.s.p > .05. Means±SEderivedfromfinalstep-2orsimplifiedmodels(seeMethods)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Plant species richness1 2 4 8

Infe

stat

ion

rate

Transplants 2007 Transplants 2008 Potted plants 2008

n.s. n.s. n.s.

1 2 3 4

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Number of functional groups

n.s. ** n.s.

F IGURE  6  Influenceofthenumberofplantspecies,functionalgroups,legume,andtallherbpresenceinexperimentalcommunitiesonparasitism rateinthethreedatasets.Significanceinfinalstep-2modelsisabbreviated:***p ≤ .001,**p ≤ .005,*p ≤ .05,n.s.p > .05.Means±SEderivedfromfinalstep-2orsimplifiedmodels (seeMethods)

0.0

0.2

0.4

0.6

0.8

1.0

Plant species richness1 2 4 8

Par

asiti

sm ra

te

Transplants 2007 Transplants 2008 Potted plants 2008

n.s.****

1 2 3 4

0.0

0.2

0.4

0.6

0.8

1.0

Number of functional groups

**n.s.n.s.

0.0

0.2

0.4

0.6

0.8

1.0

– Legumes + Legumes

n.s.n.s.***

Par

asiti

sm ra

te

0.0

0.2

0.4

0.6

0.8

1.0

– Tall herbs + Tall herbs

**n.s.n.s.

10  |     NITSCHKE ET al.

mixtures,χ2=8.63,p=.003,Table1,Figure6).Thepresenceoftallherbsalsoincreasedparasitismratesbyca.24%(χ2=4.62,p=.032,Figure6).Whenaddingcovariatestothemodelsinthesecondstepof the analyses, an additional LAI effect was seen in transplantsofbothyears (Table1,Figure7). IncreasingLAIvalues intheplantcommunities reduced parasitism rates (2007: χ2=4.25, p=.039;2008: χ2=5.88, p=.015). Including LAI in themodel did not re-movetheothereffectsoftheplantcommunity(Table1),suggestingthat these are notmediated by changes in structural complexity.ThelackofasignificantLAIeffectinpottedplants(thoughpointingintothesamedirection,seeFigure7)mayresultfromthegenerallyslightlyelevatedpositionofpottedplantsthatcausedaproportionofthemtobehigherthanthesurroundingplantcommunity.Theco-variatesAsteraceae presence,relative height,andtephritid infestation rateneversignificantlyaffectedparasitoids.Plantdiversityeffectsonparasitoidswerethereforefoundwhendirecteffectsaloneandwhenbothdirectand indirecteffectscouldoperate.However,wewerenotabletoexplainthemechanismsdrivingtheseeffectswithourcovariates.

4  | DISCUSSION

4.1 | Herbivores and parasitoids show opposing responses to plant diversity

PlantdiversityhadopposingeffectsontheherbivoreandparasitoidcommunitiesofCentaurea jaceainalargegrasslanddiversityexperi-ment.Herbivoreabundance(infestation rateandinsect load)declinedwithincreasingplantdiversity,whichagreeswiththeresultsofseveralpreviousstudies(Balvaneraetal.,2006;Haddadetal.,2009;Unsickeretal.,2006).However,theseeffectswerevariableanddidnotoccurforallmeasuresoftheherbivorecommunityorinallyears.Herbivorepopulationsarehighlytemporallyvariable(Solbreck&Sillén-Tullberg,1986;Walker,Hartley,&Jones,2008)andplantdiversityeffectsmayonlybedetectedinparticularyears.Plantdiversitymostlytendedto

havenegative(orinonecaseneutral)effects,soalthoughthereisvari-ationinthestrengthandsignificanceofeffects,thedirectionislargelyconsistent.However,thevariationinstrengthoftheeffectindicatesthat, although diversitymay act in a similarway in different years,other drivers of herbivore abundance (such as climate or dispersal)may frequentlymask theeffectsofplantdiversity. In contrast, andinlinewiththeEnemies Hypothesis(Root,1973)andarangeofmorerecentstudies(Albrecht,Duelli,Schmid,&Muller,2007;Bianchietal.,2006;Haddadetal.,2009;Vanbergen,Hails,Watt,&Jones,2006),parasitoidabundance increasedwithplantdiversity.We found thatbothplantspeciesrichnessandfunctionalgrouprichnesswereimpor-tantcomponentsofdiversitythataffectedtheabundanceofhighertrophiclevels.Wecomparedresponsesonpottedplants,whereonlydirecteffectsofplantdiversity,mediatedbychangesinresourceden-sity or apparency, can occur,with responses on transplants,whereindirecteffects,mediatedbychangingresourcelevels,couldoccurinadditiontothedirecteffects.Withthiscomparison,wewereabletoshowthatherbivorecommunitiesrespondedmorestronglytoindirecteffectsofplantdiversity,perhapsmediatedbychanges in resourcelevels,whereasparasitoidcommunities respondedmorestrongly todirecteffectsofplantdiversity.

The two experimental variables plant species richness and plant functional group richnessrepresentdifferentaspectsofplantdiversity:Functional group richnesseffectssuggest thatplantspecies fromthesame functional group are redundant and have similar effects (i.e.,similarplantcharacteristics),whereaseffectsofplant species richness implythatthereisvariationbetweenspecieswithinfunctionalgroups(Roscheretal.,2004).Itisnotclearwhythedifferentmeasuresofdi-versityshoulddrivedifferentmeasuresofherbivoreabundanceandhave effects in differentyears.However, these results suggest thatbothaspectsofdiversitymaybeimportantundercertainconditionsandthattheymayhavesimilareffects(likelyonplantqualityandre-sourceavailability).Plant speciesandfunctional group richnessalsobothreduced C. jacea performance traits (Nitschke etal., 2010), furthersupportingthisidea.

F IGURE  7  Influenceofcommunity“leafareaindex”(LAI)onparasitism rateinthethreedatasets.Significanceinfinalstep-2modelsisabbreviated:*p ≤ .05,n.s.p > .05.Effects(greyline)and95%confidenceintervals(CI,dottedlines)derivedfromfinalstep-2orsimplifiedmodels(seeMethods)

1.0 2.0 3.0 4.0

0.0

0.2

0.4

0.6

0.8Transplants 2007 (*)

LAI

Par

asiti

sm ra

te

0.5 1.0 1.5 2.0 2.5

0.0

0.2

0.4

0.6

0.8Transplants 2008 (*)

LAI0.5 1.0 1.5 2.0 2.5 3.0

0.0

0.2

0.4

0.6

0.8Potted plants 2008 (n.s.)

LAI

     |  11NITSCHKE ET al.

4.2 | Plant diversity indirectly affects the herbivorous Tephritidae

Plant diversity effects on Tephritidaewere few andwere only de-tected in the transplants where both direct and indirect effects ofplantdiversitycanoccur.Thefactthatnosucheffectwasobservedinpottedplants (directeffectsonly) impliesthattheseeffectswerelargely mediated by changes in plant quality or performance traitsalong the plant diversity gradient. Accordingly, we found little evi-dence that apparency or the availability of alternative hosts in theplant community affected theherbivores.Thismaybebecause theherbivores (dominatedby themonophagousChaetorellia jaceae) arevery efficient at finding their hosts and can locate them regardlessof their surroundings. This is supported by the lack of an effect ofstructural complexity. We hypothesized that increasing structuralcomplexityinthecommunitieswouldimpairthetephritids’hostfind-ingabilities,butourproxyforcomplexity(LAI)wasneverretainedinthefinalherbivoremodels.Wedidfindthatgrasspresencereducedtephritidabundance,whichmightbeexplainedbyassociationalresist-anceeffects(Barbosaetal.,2009);however,thereisnoevidencethatthiscanexplaintheeffectofplantdiversityonherbivoreabundance.Instead,variation inhostplantperformanceandquality seemmorelikely to explain theplant diversity effects. Transplant performance(i.e., biomass, thenumberof flowerheads) declineswith increasingplantdiversity (Nitschkeetal.,2010)and this is likely to result inadecrease in foodquality for the herbivores. In agreementwith thisidea,anincreaseinthenumberofflowerheadsdidincreaseinfesta-tionratesinthetransplants.ThisindicatesthattheTephritidaearere-sourcelimited,whichagreeswithfindingsbyDempsteretal.(1995).However, contrary to our expectations, resource density per hostplantdidnotexplaintheeffectofplantfunctionalgrouprichnessontheherbivorecommunity(i.e.,functionalrichnessremainedsignificantwhenfittedalongsidethenumberofflowerheads).Thissuggeststhatotherhostplantcharacteristicsmustaccountforthenegativeeffectoffunctionaldiversity.

Inadditiontohostplantsize,nutritionalqualityisalsolikelytoaf-fectherbivorecommunities.Nutritionalqualityisexpectedtodeclinewith increasing plant diversity as a result of increased light compe-titionand/or increasednutrientuseefficiency indiverseplantcom-munities, and thus tonegativelyaffectherbivoreabundance (Abbasetal.,2013;butseeEbelingetal.,2014).Theoppositepatterncouldoccur ifshadingreducesplantchemicaldefensesandincreasesspe-cific leaf area andhencepalatability (Crone&Jones, 1999;Guerra,Becerra, & Gianoli, 2010;Mraja, Unsicker, Reichelt, Gershenzon, &Roscher,2011).However,thenegativeresponseoftephritidspeciestodiversitysuggestsadeclineinplantqualityinthiscase.Adeclineinplantqualitycouldaffectoviposition,asmanytephritidspeciescanas-sesshostplantqualityandadjustclutchsizeinresponse(Burkhardt&Zwölfer,2002;Freese&Zwölfer,1996;Pittara&Katsoyannos,1992;Rieder,Evans,&Durham,2001).Forinstance,BurkhardtandZwölfer(2002) found thatovipositing femalesof thegall formingU. jaceana preferredhigh-qualityplantsand flowerheadswhich resulted in in-creasedlarvalgrowthandfecundity.Theproductionofagalliscostly

andtime-consuming,whichmeansthatthereis likelytobeastrongadvantageforU. jaceanainassessinghostplantqualityandnotinvest-inginpoorqualityplants.OthermonophagousspecieslikeChaetorellia jaceae,themostabundantspeciesinourstudy,mayalsohaveevolvedmechanismstoassesshostplantquality.Wemightexpectverydiffer-ent patterns formore generalist herbivores that could benefit fromdietmixingindiverseplantcommunities(Pfisterer,Diemer,&Schmid,2003).Thesenegativeindirecteffectsofplantdiversityonherbivorecommunities,mediatedby changes in host plant quality, havebeenlargelyoverlookedbutourresultssuggestthattheymaybeimportant,particularlyformonophagousspecies.

4.3 | Plant diversity directly affects parasitoid communities

In accordance with predictions from the Enemies Hypothesis (Root,1973),wefoundthatparasitismratesincreasedwithincreasingplantfunctionalgrouporspeciesrichness.Thisagreeswithanobservationalstudy,whereplantdiversitywasaffectedbygrazingintensity,whichfoundthattheparasitismratebyPteromalus elevatusonthetephritidTephritis conura increased with plant species richness (Vanbergenetal.,2006).Ourstudyallowsusto lookatthemechanismsdrivingtheseeffects inmoredetail.Asthediversityeffect inpottedplantscannotbeattributedtochangesinhostplantperformance,itmustbecausedbychangesintheplantcommunity,whichsuggeststhatdirecteffectsofdiversityare themost important.Oneof themaindirecteffectsofplantdiversitymaybetoincreasetheavailabilityoffloralresources,whichisexpectedtobenefitparasitoidsthroughincreasednectar provision (e.g., Araj etal., 2008; Lavandero etal., 2006). Inour study system, floweravailability increaseswith increasingplantspeciesrichness(Ebeling,Klein,Schumacher,Weisser,&Tscharntke,2008), which means diverse communities would provide more re-sourcesforparasitoids.ThecovariateAsteraceae presenceintheplantcommunitydidnotaffectparasitoidabundance,whichsuggeststhattheparasitoidsmaybequitegeneralistandabletofeedonarangeoffloweringplants.Thisissupportedbythefactthatthepresenceoftallherbs inthecommunitydidsignificantly increaseparasitismrates intransplants. Incontrast, thepresenceof legumesdecreasedparasit-ismandthismightbeexplainediflegumepresenceresultsinasmallerrelativeabundanceofotherherbs.Thiswouldreduceparasitoidabun-danceifparasitoidsdonotfeedonlegumeflowersandpreferopenflowerssuchasthoseoftallherbs intheAsteraceae,Apiaceae,andDipsacaceaefamilies.Indeed,studiesontheattractivenessofvariousplantspeciestonectarfeedingparasitoidssuggestthatChalcidoideapreferentiallyvisitopenandeasilyaccessible inflorescencessuchasthoseofApiaceaeandAsteraceae(Sivinski,Wahl,Holler,AlDobai,&Sivinski,2011)andavoidlegumes(Jervis,Kidd,Fitton,Huddleston,&Dawah,1993).Measuringthecoverofdifferentflowertypes intheplotswouldbeneeded toexplicitly test this idea.Moreover, extra-floral nectaries (EFNs) in plantsmay serve as an alternative nectarsourceforparasitoids,andifdiversecommunitiescontainmoreplantswithEFNsthenthiscouldexplainourresults.Only3ofour60specieshaveEFNsaccordingtothe“WorldListofPlantswithEFNs”(Weber,

12  |     NITSCHKE ET al.

Porturas,&Keeler,2015).However,thelist isnotfullycomprehen-siveandonlycontainsthosespeciescurrentlyreportedtohaveEFNs.FurthermeasuresofEFNsonthefieldsitewouldbeneededtotestforsuchaneffect.Inadditiontothesepotentialeffectsoffloralresourceavailability, parasitism ratemaybe affected by host resource avail-ability.Tephritid infestation rate and insect loadbothdecreasedwithincreasingplantdiversity(functionalgrouporspeciesrichness),whichmeanstheincreaseinparasitismcouldalsobedrivenbydecreasinghostdensity.Thisresponsewasreported inastudyontrapnestingbeesonthesamefieldsite (Ebelingetal.,2012).However,suchaneffectseemsless likelyasanexplanationofourresultsbecausewefoundthattephritid infestation ratedidnotaffectparasitismrate.Thisis in accordancewith results from a study byWalker etal. (2008),whichimpliesthathostdensitywaseithernotamaindriverofparasit-ismratesorthatparasitoidscannoteasilydeterminehostdensity.Thelatternotionissupportedbyareductioninparasitism rateswhenthenumber of flower headswashighinpottedplants.Thismightindicatedecreasingparasitoidefficiencywithanincreasingnumberofpoten-tialhostlocations,thatis,a“dilutioneffect.”TheparasitoidsthatwefoundcanattackherbivoresassociatedwithsomeotherAsteraceaespeciesfoundonourfieldsite,withalmostnootherassociationsre-corded(Noyes,2016).However,thevariableAsteraceae presence did notaffectparasitoidparameters,whichimpliesthatparasitoidswerenotusingalternativehostswithinthehighdiversitycommunities.Wedetectedanotherdirectcommunityeffectonparasitoidabundance:ourproxy for structural complexity (LAI) significantly reducedpara-sitism rate suggesting that host findingwas impaired by increasingcomplexity of the vegetation. Although LAI is positively related toplantdiversity (Spehnetal.,2000), ithasopposingeffectsonpara-sitoidabundance.Thisimpliesthat,althoughanincreaseinstructuralcomplexitymightreducetheabilityofparasitoidstofindtheirhostsin diverse plant communities, other benefits of plant diversity are sufficienttooverridethiseffects.

Ourresultssuggestthatdiverseplantcommunitiesharboramoreefficientparasitoidcommunity,likelybecauseofagreaterprovisionoffloralresources.

4.4 | General conclusion

Ourstudyinamodeltri-trophicsystem—Centaurea jacea,Tephritidae,parasitoids—suggests opposing responses of herbivores and parasi-toids toplantdiversitywithclearereffects seen inparasitoids thaninherbivores.Thenegativeherbivore responseseems tobemostlydrivenbychangesinhostplantquality.Thissuggeststhatsomeofthenegativeeffectsofplantdiversityonherbivoreabundancefound inpreviousstudiescouldbeexplainedbythesemoreindirecteffectsonplantquality.Futurestudiesshouldthereforeconsidercontrollingforchangesinqualityandpottedplantsplacedintocommunitiesareause-fulwaytodothis.Incontrasttotheherbivores,parasitoidabundanceincreasedwithdiversity.Thisseemstobepartlydrivenbyincreasedresourceavailability(i.e.,nectarandpollen)althoughdirectmeasuresoftheresourcesavailabletoparasitoidswouldbeneededtoconfirmthis.Theincreaseinparasitismrate(anddecreaseinherbivory)also

argue for thevalueofdiverseplantcommunities inprovidingmoreefficientpestcontrol.Our results showthatplantdiversity isakeydriveroftheabundanceofhighertrophiclevelsandthatawidevari-etyofmechanismscanoperatetoexplaintheseeffects.

ACKNOWLEDGMENTS

Wethankanumberof enthusiastic students and studenthelpers fortheirengagementinrecordingsofplanttraitsin2007and2008.KeesvanAchterberg(Naturalis,NL)andGérardDelvare(Cirad,F)aregrate-fullyacknowledgedforkindlyidentifyingBraconidaeandasubsampleofEurytomidae,respectively.WethankMateuszJochymforinvaluableRsupportandSebastianT.MeyerandAnneEbelingforhelpfuldiscussions.KristinaHahn(Blume 2000)arrangedalargesupplyofplantersforthepot-tedplantexperiment.WethanktheDeutscheForschungsgemeinschaft(DFG)forfunding(WE3081/3-4andWE3081/17-2).WealsothanktheBurgergemeindeBernforsupportingtheworkofHB.Twoanonymousreviewers’commentsonformerversionsofthemanuscriptsignificantlyimprovedthepresentationofourresults.

CONFLICT OF INTERESTS

Nonedeclared.

DATA ACCESSIBILITY

Datawillbemadepubliclyavailableviathedatapublisher“Pangaea”(https://pangaea.de).

REFERENCES

Abbas,M.,Ebeling,A.,Oelmann,Y.,Ptacnik,R.,Roscher,C.,Weigelt,A.,…Hillebrand,H.(2013).Biodiversityeffectsonplantstoichiometry.PLoS One,8(3),e58179.

Albrecht,M.,Duelli,P.,Schmid,B.,&Muller,C.B.(2007).Interactiondiver-sitywithinquantifiedinsectfoodwebsinrestoredandadjacentinten-sivelymanagedmeadows.Journal of Animal Ecology,76,1015–1025.

Andow, D. A. (1991). Vegetational diversity and arthropod population response.Annual Review of Entomology,36,561–586.

Andow,D.A., & Prokrym,D. R. (1990). Plant structural complexity andhost-findingbyaparasitoid.Oecologia,82,162–165.

Araj,S.E.,Wratten,S.,Lister,A.,&Buckley,H.(2008).Floraldiversity,para-sitoidsandhyperparasitoids-Alaboratoryapproach.Basic and Applied Ecology,9,588–597.

Awmack,C.S.,&Leather,S.R.(2002).Hostplantqualityandfecundityinherbivorousinsects.Annual Review of Entomology,47,817–844.

Balvanera, P., Pfisterer, A. B., Buchmann, N., He, J. S., Nakashizuka, T.,Raffaelli,D.,&Schmid,B.(2006).Quantifyingtheevidenceforbiodi-versityeffectsonecosystemfunctioningandservices.Ecology Letters,9,1146–1156.

Barbosa,P.,Hines,J.,Kaplan,I.,Martinson,H.,Szczepaniec,A.,&Szendrei,Z. (2009). Associational resistance and associational susceptibility:Havingrightorwrongneighbors.Annual Review of Ecology, Evolution, and Systematics,40,1–20.

Bates,D.,&Sarkar,D. (2007).Lme4: linear mixed-effects models using S4 classes,0.99875-9[Online].Retrievedfromhttp://www.r-project.org/[Accessed2011].

Bianchi, F. J. J. A., Booij, C. J. H., & Tscharntke, T. (2006). Sustainablepest regulation in agricultural landscapes: A review on landscape

     |  13NITSCHKE ET al.

composition,biodiversityandnaturalpest control.Proceedings of the Royal Society B: Biological Sciences,273,1715–1727.

Bukovinszky,T.,Gols,R.,Hemerik,L.,vanLenteren,J.C.,&Vet,L.E.M.(2007).Timeallocationofaparasitoidforaginginheterogeneousveg-etation:Implicationsforhost-parasitoidinteractions.Journal of Animal Ecology,76,845–853.

Burkhardt,B.,&Zwölfer,H.(2002).Macro-evolutionarytrade-offs inthetephritidgenusUrophora:Benefitsandcostsofanimprovedplantgall.Evolutionary Ecology Research,4,61–77.

Castagneyrol,B.,Giffard,B.,Pere,C.,&Jactel,H.(2013).Plantapparency,an overlooked driver of associational resistance to insect herbivory.Journal of Ecology,101,418–429.

Coll,M.,&Bottrell,D.G. (1994).Effectsofnonhostplantsonan insectherbivoreindiversehabitats.Ecology,75,723–731.

Crone,E.E.,&Jones,C.G.(1999).Thedynamicsofcarbon-nutrientbalance:Effects of cottonwood acclimation to short-and long-term shade onbeetlefeedingpreferences.Journal of Chemical Ecology,25,635–656.

Dempster, J. P.,Atkinson, D.A., & Cheesman,O. D. (1995). The spatialpopulation-dynamicsofinsectsexploitingapatchyfoodresource–1.Populationextinctionsandregulation.Oecologia,104,340–353.

Diaz-Fleischer,F.,Papaj,D.R.,Prokopy,R.J.,Norrbom,A.L.,&Aluja,M.(1999).Evolutionoffruit flyovipositionbehavior. InM.Aluja&A.L.Norrbom(Eds.),Fruit flies (Tephritidae): Phylogeny and evolution of behav-ior(pp.811–842).BocaRanton,FL:CRCPress.

Dormann, C. F.,Gruber, B., & Fründ, J. (2008). Introducing the bipartitepackage:Analysingecologicalnetworks.R- News,8,8–11.

Ebeling,A.,Klein,A.M.,Schumacher,J.,Weisser,W.W.,&Tscharntke,T.(2008).Howdoesplantrichnessaffectpollinatorrichnessandtempo-ralstabilityofflowervisits?Oikos,117,1808–1815.

Ebeling,A.,Klein,A.M.,Weisser,W.W.,&Tscharntke,T.(2012).Multitrophiceffects of experimental changes in plant diversity on cavity-nestingbees,wasps,andtheirparasitoids.Oecologia,169,453–465.

Ebeling,A.,Meyer,S.T.,Abbas,M.,Eisenhauer,N.,Hillebrand,H.,Lange,M.,…Weisser,W.W.(2014).Plantdiversityimpactsdecompositionandher-bivoryviachangesinabovegroundarthropods.PLoS One,9(9),e106529.

Endara,M.J.,&Coley,P.D. (2011).The resourceavailabilityhypothesisrevisited:Ameta-analysis.Functional Ecology,25,389–398.

Finch,S.,&Collier,R.H.(2000).Host-plantselectionbyinsects-atheorybasedon‘appropriate/inappropriatelandings’bypestinsectsofcruci-ferousplants.Entomologia Experimentalis Et Applicata,96,91–102.

Fox,J.(2003).EffectdisplaysinRforgeneralisedlinearmodels.Journal of Statistical Software,8,1–27.

Freese,G.,&Zwölfer,H. (1996).Theproblemofoptimalclutchsize inatritrophicsystem:TheovipositionstrategyofthethistlegallflyUrophora cardui(Diptera,Tephritidae).Oecologia,108,293–302.

Gols,R.,Bukovinszky,T.,Hemerik,L.,Harvey,J.A.,vanLenteren,J.C.,&Vet,L.E.M.(2005).Reducedforagingefficiencyofaparasitoidunderhabitatcomplexity:implicationsforpopulationstabilityandspeciesco-existence.Journal of Animal Ecology,74,1059–1068.

Guerra,P.C.,Becerra,J.,&Gianoli,E. (2010).Explainingdifferentialher-bivory in sun and shade: The case of Aristotelia chilensis saplings.Arthropod- Plant Interactions,4,229–235.

Haddad,N.M.,Crutsinger,G.M.,Gross,K.,Haarstad,J.,Knops,J.M.H.,&Tilman,D.(2009).Plantspecieslossdecreasesarthropoddiversityandshiftstrophicstructure.Ecology Letters,12,1029–1039.

Hirose,T.,&Werger,M.J.A. (1995).Canopy structure andphoton fluxpartitioningamongspeciesinaherbaceousplantcommunity.Ecology,76,466–474.

Jactel,H.,&Brockerhoff,E.G.(2007).Treediversityreducesherbivorybyforestinsects.Ecology Letters,10,835–848.

Jervis,M.A.,Kidd,N.A.C.,Fitton,M.G.,Huddleston,T.,&Dawah,H.A.(1993).Flower-visitingbyhymenopteranparasitoids.Journal of Natural History,27,67–105.

Jongejans, E., de Kroon, H., & Berendse, F. (2006). The interplay be-tween shifts in biomass allocation and costs of reproduction in four

grasslandperennialsundersimulatedsuccessionalchange.Oecologia,147,369–378.

Kagata,H.,Nakamura,M.,&Ohgushi,T. (2005).Bottom-upcascade inatri-trophic system: Different impacts of host-plant regeneration onperformanceofawillow leafbeetleand itsnaturalenemy.Ecological Entomology,30,58–62.

Kigathi,R.N.,Weisser,W.W.,Veit,D.,Gershenzon,J.,&Unsicker,S.B.(2013).Plantssuppresstheiremissionofvolatileswhengrowingwithconspecifics.Journal of Chemical Ecology,39,537–545.

Langellotto,G.A.,&Denno,R.F.(2004).Responsesofinvertebratenaturalenemies tocomplex-structuredhabitats:Ameta-analytical synthesis.Oecologia,139,1–10.

Lavandero,B.,Wratten,S.D.,Didham,R.K.,&Gurr,G.M.(2006).Increasingfloraldiversityforselectiveenhancementofbiologicalcontrolagents:Adouble-edgedsward?Basic and Applied Ecology,7,236–243.

Lemon,J. (2006).Plotrix:ApackageintheredlightdistrictofR.R- News,6,8–12.

Mattson, W. J. (1980). Herbivory in relation to plant nitrogen-content.Annual Review of Ecology and Systematics,11,119–161.

Menalled,F.D.,Marino,P.C.,Gage,S.H.,&Landis,D.A.(1999).Doesagri-culturallandscapestructureaffectparasitismandparasitoiddiversity?Ecological Applications,9,634–641.

Merz,B. (1994).Diptera:Tephritidae. Insecta Helvetica Fauna. 10,1–198.Geneva:HGEpress.

Mraja,A.,Unsicker,S.B.,Reichelt,M.,Gershenzon,J.,&Roscher,C.(2011).Plantcommunitydiversityinfluencesallocationtodirectchemicalde-fenceinPlantago lanceolata. PLoS One,6(12),e28055.

Nitschke, N., Ebeling, A., Rottstock, T., Scherber, C., Middelhoff, C.,Creutzburg,S.,…Weisser,W.W.(2010).TimecourseofplantdiversityeffectsonCentaurea jaceaestablishmentandtheroleofcompetitionandherbivory.Journal of Plant Ecology- UK,3,109–121.

Noyes, J. S. (2016). Universal ChalcidoideaDatabase.WorldWideWebelectronicpublication.http://www.nhm.ac.uk/chalcidoids

Petermann,J.S.,Müller,C.B.,Weigelt,A.,Weisser,W.W.,&Schmid,B.(2010).Effectofplantspecies lossonaphid-parasitoidcommunities.Journal of Animal Ecology,79,709–720.

Pfisterer,A.B.,Diemer,M.,&Schmid,B.(2003).Dietaryshiftandloweredbiomass gain of a generalist herbivore in species-poor experimentalplantcommunities.Oecologia,135,234–241.

Pittara, I. S.,&Katsoyannos,B. I. (1992).Effectof shape, sizeandcoloronselectionofovipositionsitesbyChaetorellia australis. Entomologia Experimentalis Et Applicata,63,105–113.

Press,B.,&Gibbons,B.(1993).Wild flowers of Britain and Europe.London:NewHolland.

Price,P.W.,Bouton,C.E.,Gross,P.,McPheron,B.A.,Thompson,J.N.,&Weis,A.E. (1980). Interactionsamong3trophic levels- influenceofplantsoninteractionsbetweeninsectherbivoresandnaturalenemies.Annual Review of Ecology and Systematics,11,41–65.

RCoreTeam(2010).R: A language and environment for statistical computing. Vienna,Austria: R Foundation for StatisticalComputing. https://ww-w.R-project.org

Randlkofer,B.,Obermaier,E.,Hilker,M.,&Meiners,T.(2010).Vegetationcomplexity -The influenceof plant species diversity andplant struc-turesonplantchemicalcomplexityandarthropods.Basic and Applied Ecology,11,383–395.

Randlkofer,B.,Obermaier,E.,&Meiners,T. (2007).Mother’schoiceof theovipositionsite:Balancingriskofeggparasitismandneedoffoodsupplyfor theprogenywithan infochemical shelter?Chemoecology,17,177– 186.

Rieder,J.P.,Evans,E.W.,&Durham,S. L. (2001).Distributionof insectattacksinbiologicalcontrolofweeds:infestationofCentaurea virgata flowerheadsbyagallfly.Biological Control,20,254–260.

Root,R.B. (1973).Organizationofaplant-arthropodassociation in sim-ple and diverse habitats - The fauna of collards (Brassica oleracea).Ecological Monographs,43,95–120.

14  |     NITSCHKE ET al.

Roscher,C.,Kutsch,W.L.,&Schulze,E.D.(2011).Lightandnitrogencom-petition limitLolium perenne inexperimentalgrasslandsof increasingplantdiversity.Plant Biology,13,134–144.

Roscher,C., Schumacher, J., Baade, J.,Wilcke,W.,Gleixner,G.,Weisser, W.W.,…Schulze, E.D. (2004).The role of biodiversity for element cyclingandtrophicinteractions:Anexperimentalapproachinagrass-landcommunity.Basic and Applied Ecology,5,107–121.

van Ruijven, J., & Berendse, F. (2005). Diversity-productivity relation-ships: initialeffects, long-termpatterns,andunderlyingmechanisms.Proceedings of the National Academy of Sciences of the United States of America,102,695–700.

Rutten,G.,Ensslin,A.,Hemp,A.,&Fischer,M.(2015).Verticalandhorizon-talvegetationstructureacrossnaturalandmodifiedhabitat typesatMountKilimanjaro.PLoS One,10(9),e0138822.

Scherber,C.,Eisenhauer,N.,Weisser,W.W.,Schmid,B.,Voigt,W.,Fischer,M.,…Tscharntke,T.(2010).Bottom-upeffectsofplantdiversityonmul-titrophicinteractionsinabiodiversityexperiment.Nature,468,553–556.

Sivinski,J.,Wahl,D.,Holler,T.,AlDobai,S.,&Sivinski,R.(2011).Conservingnaturalenemieswithfloweringplants:EstimatingfloralattractivenesstoparasiticHymenoptera andattraction’s relationship to flower andplantmorphology.Biological Control,58,208–214.

Solbreck,C.,&Sillén-Tullberg,B. (1986). Seedproductionand seedpre-dationinapatchyandtime-varyingenvironment.Dynamicsofamilk-weed-tephritidflysystem.Oecologia,71,51–58.

Spehn,E.M.,Joshi,J.,Schmid,B.,Diemer,M.,&Korner,C.(2000).Above-ground resourceuse increaseswithplant species richness inexperi-mentalgrasslandecosystems.Functional Ecology,14,326–337.

Thies,C.,&Tscharntke,T.(1999).Landscapestructureandbiologicalcon-trolinagroecosystems.Science,285,893–895.

Unsicker, S. B., Baer, N., Kahmen, A., Wagner, M., Buchmann, N., &Weisser,W.W. (2006). Invertebrate herbivory along a gradient ofplantspeciesdiversityinextensivelymanagedgrasslands.Oecologia,150,233–246.

VanAlphen,J.J.M.,&Bernstein,C.(2008).Informationacquisition,infor-mationprocessing,andpatch timeallocation in insectparasitoids. InÉ.Wajnberg,C.Bernstein&vanAlphenJ.J.M.(Eds.),Behavioral ecol-ogy of insect parasitoids - from theoretical approaches to field applications (pp.172–192).BlackwellPublishingLtd.

Vanbergen,A.J.,Hails,R.S.,Watt,A.D.,&Jones,T.H.(2006).Consequencesforhost-parasitoid interactionsofgrazing-dependenthabitathetero-geneity.Journal of Animal Ecology,75,789–801.

Varley,G.C.(1947).Thenaturalcontrolofpopulationbalanceintheknap-weedgall-fly(Urophora jaceana).Journal of Animal Ecology,16,139–187.

Walker,M.,Hartley,S.E.,&Jones,T.H.(2008).Therelativeimportanceofre-sourcesandnaturalenemiesindeterminingherbivoreabundance:Thistles,tephritidsandparasitoids.Journal of Animal Ecology,77,1063–1071.

Warnes,G.R.(2010).gplots:VariousRprogrammingtoolsforplottingdata(includesRsourcecodeand/ordocumentationcontributedby:BolkerB,BonebakkerL,GentlementR,HuberW,LiawA,LumleyT,MaechlerM,MagnussonA,Moeller S, SchwartzM, andVenables B) [Online].http://www.r-project.org/

Weber,M.G.,Porturas,L.D.,&Keeler,K.H. (2015).Worldlistofplantswithextrafloralnectaries.Retrievedfromwww.extrafloralnectaries.org[accessed:07January2017].

Weigelt,A.,Marquard,E.,Temperton,V.M.,Roscher,C.,Scherber,C.,Mwangi,P.N.,…Weisser,W.W.(2010).TheJenaExperiment:Sixyearsofdatafromagrasslandbiodiversityexperiment.Ecology,91(3),930–931.

Welles,J.M.,&Norman,J.M.(1991).Instrumentforindirectmeasurementofcanopyarchitecture.Agronomy Journal,83,818–825.

White, I. M. (1988). Tephritid Flies Diptera: Tephritidae. Handbook for the identification of British insects,Vol.10,Part5a.London:R.Entomol.Soc.London.134pp.

White,J.A.,&Andow,D.A.(2005).Host-parasitoidinteractionsinatrans-geniclandscape:Spatialproximityeffectsofhostdensity.Environmental Entomology,34,1493–1500.

Zwölfer,H.(1988).Evolutionaryandecologicalrelationshipsoftheinsectfaunaofthistles.Annual Review of Entomology,33,103–122.

SUPPORTING INFORMATION

Additional Supporting Informationmay be found online in the sup-portinginformationtabforthisarticle.

How to cite this article:NitschkeN,AllanE,ZwölferH,etal.PlantdiversityhascontrastingeffectsonherbivoreandparasitoidabundanceinCentaurea jaceaflowerheads.Ecol Evol.2017;00:1–14.https://doi.org/10.1002/ece3.3142