The Missing Part of Seed Dispersal Networks: Structure and Robustness of Bat-Fruit Interactions


previous item | next item

Citation
The Missing Part of Seed Dispersal Networks: Structure and Robustness of Bat-Fruit Interactions

Material Information

Title:
The Missing Part of Seed Dispersal Networks: Structure and Robustness of Bat-Fruit Interactions
Series Title:
PLoS ONE
Creator:
Mello, Marco Aurelio Ribeiro
Marquitti, Flávia Maria Darcie
Guimarães Jr., Paulo Roberto
Kalko, Elisabeth Klara Viktoria
Jordano, Pedro
Martinez de Aguiar, Marcus Aloiz
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Species Interactions ( local )
Fruits ( local )
Seeds ( local )
Bats ( local )
Community Ecology ( local )
Plants ( local )
Centrality ( local )
Species Extinction ( local )
Genre:
serial ( sobekcm )

Notes

Abstract:
Mutualistic networks are crucial to the maintenance of ecosystem services. Unfortunately, what we know about seed dispersal networks is based only on bird-fruit interactions. Therefore, we aimed at filling part of this gap by investigating bat-fruit networks. It is known from population studies that: (i) some bat species depend more on fruits than others, and (ii) that some specialized frugivorous bats prefer particular plant genera. We tested whether those preferences affected the structure and robustness of the whole network and the functional roles of species. Nine bat-fruit datasets from the literature were analyzed and all networks showed lower complementary specialization (H2' = 0.37±0.10, mean ± SD) and similar nestedness (NODF = 0.56±0.12) than pollination networks. All networks were modular (M = 0.32±0.07), and had on average four cohesive subgroups (modules) of tightly connected bats and plants. The composition of those modules followed the genus-genus associations observed at population level (Artibeus-Ficus, Carollia-Piper, and Sturnira-Solanum), although a few of those plant genera were dispersed also by other bats. Bat-fruit networks showed high robustness to simulated cumulative removals of both bats (R = 0.55±0.10) and plants (R = 0.68±0.09). Primary frugivores interacted with a larger proportion of the plants available and also occupied more central positions; furthermore, their extinction caused larger changes in network structure. We conclude that bat-fruit networks are highly cohesive and robust mutualistic systems, in which redundancy is high within modules, although modules are complementary to each other. Dietary specialization seems to be an important structuring factor that affects the topology, the guild structure and functional roles in bat-fruit networks.
Original Version:
PLoS ONE, Vol. 6, no. 2 (2011-02-28).

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
This item is licensed with the Creative Commons Attribution License. This license lets others distribute, remix, tweak, and build upon this work, even commercially, as long as they credit the author for the original creation.
Resource Identifier:
K26-05099 ( USFLDC: LOCAL DOI )
k26.5099 ( USFLDC: LOCAL Handle )

USFLDC Membership

Aggregations:
University of South Florida
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

TheMissingPartofSeedDispersalNetworks:Structure andRobustnessofBat-FruitInteractionsMarcoAurelioRibeiroMello1* ,Fla ´ viaMariaDarcieMarquitti2,PauloRobertoGuimara ˜ esJr.3,Elisabeth KlaraViktoriaKalko1,4,PedroJordano5,MarcusAloizioMartinezdeAguiar61 InstituteofExperimentalEcology,UniversityofUlm,Ulm,Germany, 2 ProgramadePo ´ s-graduac ¸a ˜ oemEcologia,UniversidadeEstadualdeCampinas,Campinas,Brazil, 3 DepartamentodeEcologia,UniversidadedeSa ˜ oPaulo,Sa ˜ oPaulo,Brazil, 4 SmithsonianTropicalResearchInstitute,Balboa,Panama, 5 Estacio ´ nBiolo ´ gicadeDon ˜ ana, ConsejoSuperiordeInvestigacionesCient ´ ficas,Sevilla,Spain, 6 InstitutodeF ´ sica‘GlebWataghin’,UniversidadeEstadualdeCampinas,Campinas,BrazilAbstractMutualisticnetworksarecrucialtothemaintenanceofecosystemservices.Unfortunately,whatweknowaboutseed dispersalnetworksisbasedonlyonbird-fruitinteractions.Therefore,weaimedatfillingpartofthisgapbyinvestigating bat-fruitnetworks.Itisknownfrompopulationstudiesthat:(i)somebatspeciesdependmoreonfruitsthanothers,and(ii) thatsomespecializedfrugivorousbatspreferparticularplantgenera.Wetestedwhetherthosepreferencesaffectedthe structureandrobustnessofthewholenetworkandthefunctionalrolesofspecies.Ninebat-fruitdatasetsfromtheliterature wereanalyzedandallnetworksshowedlowercomplementaryspecialization( H2’ =0.37 6 0.10,mean 6 SD)andsimilar nestedness( NODF =0.56 6 0.12)thanpollinationnetworks.Allnetworksweremodular( M =0.32 6 0.07),andhadonaverage fourcohesivesubgroups(modules)oftightlyconnectedbatsandplants.Thecompositionofthosemodulesfollowedthe genus-genusassociationsobservedatpopulationlevel( Artibeus-Ficus,Carollia-Piper, and Sturnira-Solanum ),althoughafew ofthoseplantgeneraweredispersedalsobyotherbats.Bat-fruitnetworksshowedhighrobustnesstosimulated cumulativeremovalsofbothbats( R =0.55 6 0.10)andplants( R =0.68 6 0.09).Primaryfrugivoresinteractedwithalarger proportionoftheplantsavailableandalsooccupiedmorecentralpositions;furthermore,theirextinctioncausedlarger changesinnetworkstructure.Weconcludethatbat-fruitnetworksarehighlycohesiveandrobustmutualisticsystems,in whichredundancyishighwithinmodules,althoughmodulesarecomplementarytoeachother.Dietaryspecialization seemstobeanimportantstructuringfactorthataffectsthetopology,theguildstructureandfunctionalrolesinbat-fruit networks.Citation: MelloMAR,MarquittiFMD,Guimara ˜ esPRJr.,KalkoEKV,JordanoP,etal.(2011)TheMissingPartofSeedDispersalNetworks:StructureandRobustness ofBat-FruitInteractions.PLoSONE6(2):e17395.doi:10.1371/journal.pone.0017395 Editor: AnnaTraveset,InstitutMediterranidEstudisAvanc ¸ats(CSIC/UIB),Spain Received October5,2010; Accepted February2,2011; Published February28,2011 Copyright: 2011Melloetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermits unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. Funding: Fapesp(http://www.fapesp.br)sponsoredM.A.R.Mello(2006/00265-0),F.M.D.Marquitti,M.A.M.Aguiar,andP.R.Guimara ˜ es.M.A.R.Mellowasalso supported(1134644)bytheHumboldtFoundation(http://www.humboldt-foundation.de),E.K.V.KalkobyDFG(http://www.dfg.de),andP.Jordanoby MEC (http://www.educacion.es/).Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanus cript. CompetingInterests: Theauthorshavedeclaredthatnocompetinginterestsexist. *E-mail:marmello@gmail.comIntroductionInthetropics,plant-animalmutualismssuchasseeddispersal arevitalforecosystemfunctioning[1].Ahugebodyofknowledge hasbeenaccumulatedontheecologyofthoseinteractionsat populationlevel[2].However,asthepropertiesofacomplex systemcannotbetotallypredictedbasedonlyonthepropertiesof itselements(inthiscase,populationofanimalsandplants)[3],if wewanttoaddresstheimportanceofseeddispersalasan ecosystemservice,andtobetterunderstanditsroleinmaintaining biodiversity,itisessentialtoanalyzeanimal-fruitinteractionsat thecommunitylevel,i.e.consideringallspeciesatagivenlocality [4].Amongothertools(suchasmultivariateanalysis),network theoryisprovingextremelyhelpfulinthistask,asitprovidesa theoreticalframeworkandusefulanalyticalmethodstoassess patternsofinteractionamongseveralspeciesoffrugivoresand fruits[5].Networktheoryprovidesinnovativetoolsthatcanbe usedassurrogatesforassessingcomplexecologicalconcepts.Itis importanttosay,though,thatnetworkecologydoesnotreplace traditionalcommunityecology,butrathercomplementsit, becausewhiletheformerfocusesmoreontheinteractions,the latterfocusesmoreonthespecies.Therearesomebranchesof communityecology,mainlyguildtheory[6,7],thatdealtwith interactionsforalongtime;networkecologybroughttoolsfrom complexitytheorythatmadeiteasiertoassessthosecomplex systems. Unfortunately,whatweknowaboutseeddispersalnetworksis basedonlyonbird-fruitinteractions[8],althoughotheranimal groupsalsoplayimportantroles[9].Batsrepresentakeydisperser groupthathasbeenneglectedsofarinnetworkstudies,although frugivorousbatsandbirdsarejointlyresponsibleforover80%of theseedraininNeotropicalsites[10].Furthermore,batservices areinmostcaseshighlycomplementarytobirdservices[11].This gapinourknowledgeneedstobeclosedquickly,asevidence indicatesthatdifferentanimalgroupsformspeciessubgroups withinmutualisticnetworksofdifferentkinds,includingpollination[12]andant-plantmutualisms[13].Thereforeourknowledge ofseeddispersalwillbemarkedlybiaseduntilotheranimalgroups arealsostudiedfromanetworkperspective. Moststudiesonbat-fruitinteractionshavebeenlimitedtothe populationlevel,i.e.localinteractionsofsinglebatspecieswith localfruits,mostlyincludingdatafromonlyonesite.ConsequentPLoSONE|www.plosone.org1February2011|Volume6|Issue2|e17395

PAGE 2

ly,littleisknownaboutthecommunitystructureofbat-fruit interactions,withveryfewexceptions[14,15].Tostartwith,itis knownthatwithinthefamilyPhyllostomidae,specializedfrugivory (i.e.completeorstrongdependenceonfruitsforaliving)seemsto haveevolvedonlyonceinthespecies-richlineagecomprisingthe subfamiliesCarolliinae,Rinophyllinae,andStenodermatinae[16]. BatsofthesubfamilyGlossophaginaemayalsofeedonfruits,but thispartoftheirdietrepresentsasecondarychoiceafternectar andpollen;batsoftheLonchophyllinaearealmostexclusively nectarivores,andveryfewreliablerecordsexistonfruitsinthediet oftheanimal-eatingPhyllostominaebats[17].Thisisprobably onereasonwhy,despitealldietarydiversificationamong frugivorousphyllostomids[17],allofthemhaveoneormoreof fivemainplantgeneraasthecoreoftheirdiet: Cecropia , Ficus , Piper , Solanum, and Vismia [17]. Consideringthosefivemainplantgenera,aclosegenus-togenusrelationshipexistsbetweenthemandsomeprimarily frugivorousbatgenera.Whentheirpreferredfruitsareavailable, batsofthegenus Artibeus eatmostlyfruitsof Ficus (Moraceae), whereas Sturnira batsselectprimarily Solanum (Solanaceae),and Carollia batsfeedpreferablyon Piper (Piperaceae)[18].For phyllostomidbatsinfrequentlyusedplantspeciesdonotplaya largeroleinnourishment[17].Whatmostphyllostomidsdoisto shiftbetweenspeciesofthosefivemainplantgenera,depending ontheiravailabilityatdifferentseasons[18,19,20].Someofthem doalsochangetoanectarorinsectdietattimesoffruitscarcity [21].Inwhatconcernsessentialinorganicnutrients,some phyllostomidsobtainthiscomplementfromleaves[22]ormuddy water[23]. SuchclusteringtendencieshavebeenobservedatthecommunitylevelincomprehensivedietarystudiesofbatsinPanama[15]. However,itisimportanttonotethatthereissubstantial geographicvariationinthoseinteractions.InMexico, Artibeus bats feedmainlyon Cecropia [14]andnot Ficus ;thisdifferencemaybe explainedlargelybyresourceavailability,asbatfigsarenotalways abundantinNeotropicallocalities.Moreover,thiskindofspatial variabilityhasbeenexpected,basedonthetheoryofgeographic mosaicsofcoevolution[24]:mostwidespreadspeciesareunder differentselectivepressuresacrosstheirgeographicalrange,having differentsetsofavailablepartnersand,eventually,specializingin differentspecies. Consideringtheevidencefrompopulation-levelstudies,bats withnarrowerdietsseemtofeedonasubsetoftheplants consumedbybatswithbroaderdiets[25];therefore,high nestedness[sensu26]wasexpectedinbat-fruitnetworks.Wealso expectedlowcomplementaryspecialization[27];i.e.batsof differentspeciesshouldfeedonrelativelysimilarsubsetsofplants, consideringthatonlyfiveplantgeneraformthecoreofbatdiets. Moreover,clusteringisprobablylowinbat-fruitnetworks:inbatfruitnetworksthereshouldonlybeafewinterconnectedcohesive subgroupsoffrugivorousspeciesassociatedwithspecificsubsetsof plants,formingguilds[7].Furthermore,genus-genusassociations betweenbatsandplantarelikelytoplayadecisiverolein subgroupformation,althoughthisdoesnotexcludethepossibility thateachplantgenusisdispersedbymorethanonebatgenus. Ultimately,thecombinationofhighnestedness,lowcomplementaryspecialization,andlowclusteringshouldleadtohigh robustnessinbat-fruitnetworks,ashasbeensuggestedforother kindsofmutualisticnetworks[28].Inotherwords,bat-fruit networksshouldberelativelyrobustasregardsremovalofspecies oneitherside(plantsorbats);i.e.whenspeciesarecumulatively removedfromonesideofthenetwork(e.g.plants)mostspecieson theotherside(e.g.bats)shouldstillremain.Furthermore,since primarilyfrugivorousbatsdependonfruitsforaliving,they probablyinteractwithmorefruitspeciesineachnetworkandthus occupymorecentralpositionsandplaymoreimportantrolesin maintainingnetworkstructure.Ifthisistrue,wewouldexpectthe removalofprimaryfrugivorestoresultinlargerdecreasesinthe systemÂ’snestedness. Thusinthepresentstudyweaimedattestingthesehypotheses onthestructureandrobustnessofbat-fruitnetworks,inorderto addanimportantpiecetothepuzzleofmutualisticnetworksand associatedecosystemservices.Asseeddispersalisacrucialservice inthedisruptedandfragmentedlandscapesallovertheNeotropics [2],amorecompleteunderstandingofitsnetworkstructureand fragilityisofgreatimportance.Methods DatasetsWeusedninedatasetsonthedietofNeotropicalfrugivorous batscompiledfromtheliterature.Eightdatasetswereweighted (i.e.containeddataonthefrequencyoftheinteractions)andcame fromfecalanalysisconductedinBrazil,CostaRica,andPeru.We alsousedonelong-termdatasetwithpresence/absenceinformationcompiledbyE.K.V.Kalkoandco-workersonBarro ColoradoIsland(BCI),Panama(fordetailsseereferencelistin AppendixS1).Overall,weincludedonlydatasetsinwhich interactionshadbeensampledatleastforoneyear,andinwhich allfrugivorousbatspeciesinthestudyareawereconsidered,and notonlyapre-definedsubset(forinstance,asinglegenusor species).Asmostdataonplantconsumptionwereobtainedfrom fecalanalysis,somespeciesarelacking,mainlyfruitswithlarge seedsthatarenotswallowedbutdiscardedatthesitewherethe batschewedthefruits[29].ThedatasetfromBCIismore completeasitincludesdatafromobservations,roostinspections, andfruitsthatwerecarriedbythebatsintomistnets. Forouranalysisweconsideredallbatspeciesasseeddispersers, evenifafewofthemmaybeactuallymainlyseedpredators; actually,sofaronlybatsofthegenus Chiroderma areknowntofeed onseeds[30].Weregardtheeffectofseedpredatorsinthiscaseas negligiblesincethesebatsrepresentonlyaverysmallproportionof allfrugivorousspeciesinthearea,andtheyalsousuallydisperseat leastsomeseeds(pers.obs.I.WagnerandE.Kalko).NetworkstructureWeorganizeddatasetsasadjacencymatricesofanimalsand plants, A 6 P ,withbatspeciesas A rowsandplantspeciesas P columns,totestforthenetworkstructureofbat-fruitinteractions (AppendixS2).Intheweighteddatasets,cellvaluesindicatethe numberoffecalsamplesofeachbatspeciesthatcontainedseedsof eachplantspecies.Inthesinglebinarydataset,cellvaluesareonly 0or1,i.e.absenceofpresenceofinteractionbetweeneachbatand plant.GraphsweredrawninPajek2.02[31]andinthepackage bipartiteforR[32]. Theindex NODF inthesoftwareAninhado3.0[33]wasusedto measurethedegreeofnestednessofeachnetwork. NODF isa muchbetternestednessmetricthantheclassicmetric T [34], becauseitismorefine-tunedtotheoriginalconcept,asitisbased onthenestednessofallpairsofcolumnsandrowsinthematrix [26].Inanestednetwork,specieswithfewerinteractionsare connectedtoasubsetofthepartnersofspecieswithmore interactions. NODF valueswerenormalizedinordertovaryfrom0 (notnested)to1(fullynested).Thesignificanceof NODF was estimatedwithaMonteCarloprocedure.First,wegenerated 1,000randommatricesfromtheoriginalmatrix,usingthenull model2[10](nullmodelCeinAninhado),inwhichthe probabilityofinteractionbetweenabatandaplantspeciesisBat-FruitNetworks PLoSONE|www.plosone.org2February2011|Volume6|Issue2|e17395

PAGE 3

proportionaltotheirtotalnumberofinteractions(i.e.their degree).Second,wedefinedtheP-valueastheproportionof randommatricesthathada NODF valueequalorhigherthanthe valueobtainedfortherealmatrix.Whennorandommatriceshad higher NODF thantherealmatrix,wedefinedP , 0.001. Inthisstudy,weworkedwithconceptsofspecializationfrom ecologicaltheory(theoreticalvariables)andfromnetworktheory (operationalvariables).Insummary,weaimedattestingpredictions basedonecologicaltheorywithnetworksurrogates.Weused conceptsthatdealwitheachnetworkasawhole(networklevel),and conceptsthatdealwitheachspecies(specieslevel).Atthenetwork level,weusedtheoperationalconceptof‘‘complementary specialization’’totestforinteractionspecializationinthecommunityasawhole.Thisconceptdoesnottakeintoaccountdietary preferencesorcoevolutionaryassociations,asisusualinecological theory.Itonlyconsidersthenumberofinteractionsestablishedbya specieswithinanetwork(i.e.itsdegree)andhowthoseinteractions differamongspecies.Weusedthe H2’ index[35],whichvariesfrom 0(allspeciesinteractingwiththesamepartners)to1(eachspecies interactswithaparticularsubsetofpartners)toassesscomplementaryspecialization.Thisindexhastheadditionaladvantageof reducingsamplingbiases,asitconsidersaspeciesasspecialized, onlywhenitinteractsveryfrequentlywithanotherspeciesthathasa fewotherpartnersinthenetwork.Thesignificanceof H2’ was estimatedwithaMonteCarloprocedure.First,wegenerated 10,000randommatricesusingthenullmodelPatefield[35],in whichtheinteractionfrequencybetweentwospeciesisproportional totheirtotalsumofinteractions.Second,wedefinedtheP-valueas theproportionofrandommatricesthathada H2’ valueequalor higherthanthevalueobtainedfortherealmatrix.Whenno randommatriceshadhigher NODF thantherealmatrix,wedefined P , 0.001.Allanalysesoncomplementaryspecializationweremade inRwiththepackagebipartite. Inordertotestwhetherfeedingpreferencesofparticularbat generaforparticularfruitgenera[17]producedguildsinthe community[7],weusedasasurrogateofguildtheconceptof module,assessedwithafunctionalcartographyalgorithmfor modularity[36].Modularityisameasureofhowmuchthe networkisstructuredincohesivesubgroupsofvertices(modules), inwhichthedensityofinteractionsishigherwithinthanamong subgroups.Modularitywascalculatedwiththeindex M (from0, nosubgroups,to1,totallyseparatedsubgroups)withasimulated annealingalgorithminthesoftwareNetcarto(kindlyprovidedby R.Guimera `uponrequest);itssignificancewasestimatedwitha MonteCarloprocedure:100randommatricesweregenerated withthenullmodelCe(nullmodel2of[10]),inwhichthe probabilityofinteractionbetweenabatandaplantspeciesis proportionaltotheirtotalproportionofinteractions(i.e.their degree).Second,wedefinedtheP-valueasthenumberofrandom matricesthathadan M valueequalorhigherthanthevalue obtainedfortherealmatrix.Whennorandommatriceshad higher M thantherealmatrix,wedefinedP , 0.001.Weusedthe originalbipartitenetworksinthisanalysis,followingotherstudies onmutualisticnetworks[e.g.12],becauseunipartiteprojections changethemeaningoflinksfromseeddispersaltonicheoverlap, andwewantedtoassesstheguildstructureofthenetworks.Asthe softwareNetcartowasmadeforunipartitenetworks(inwhich plant-plantandanimal-animalconnectionsareallowed),we createdacostume-madeprocedureforthisanalysis,combininga MatLabcode(forgeneratingrandommatrices)withaFortran code(forautomatingthecalculationandcompilationof M -values). Weassessedtheconsistencyofthegenus-genusassociations betweenbatsandplantsatnetworklevelinthefollowingway. Inallnetworks,batsofthesamegenuswerefoundtogetherinthe samemodules.Sowecountedinhowmanynetworkseachbat genuswasfoundinamodulethatcontainedplantsofits supposedlypreferredgenus.Thenwepooledtogetherdatafrom differentnetworks,andbuiltone2 6 2tableforeachgenus:rows containedthecountofmodulesthatfollowedourpredictionand thecountofmodulesthatdidnot;columnscontainedobserved andexpectedvalues(equalproportions).Differenceswereassessed withGtests(withYatescorrection).LeveloffrugivoryTherearemanyecologicalconceptsofdietaryspecializationat thespecieslevel,andwedecidedtoworkwiththeconceptof‘levelof frugivory’:thedependenceonfruitsforliving,whenconsideringall kindsoffoodeatenbytheanimalspecies.Wefollowedaconcept developedforfrugivorousbirds[37],andbasedourclassificationon theconsensusthatonlybatsofthefamilyPhyllostomidaefeedon fruitsintheNeotropics,andthatphyllostomidsofthesubfamilies CarolliinaeandStenodermatinaedependstronglyonfruitsfor living(category‘‘primary’’).SomemembersoftheGlossophaginae takefruitsassecondaryfood(‘‘secondary’’),whereasothermembers ofthissubfamilyandofthesubfamiliesPhyllonycterinaeand Phyllostominaeseldomfeedonfruits(‘‘occasional’’). Table1. Parametersmeasuredintheninestudiedbat-fruitnetworks:speciesrichness,numberofplantspecies,numberofbat species,degreeofnestedness( NODF ),complementaryspecializationintheinteractions( H2’ ),modularity( M ),numberofmodules found( Modules ),androbustnesstotheextinctionofbatsorplants( R ). H2’ couldnotbecalculatedforthenetworkKalkoBCIasit hasonlybinarydata.NetworkRichnessPlantsBatsNODFH2’MModulesR(bats)R(plants) Faria1996231580.550.360.3340.590.69 Garciaetal.2000201460.410.390.4450.410.58 Gorchovetal.19953726110.670.300.2440.690.84 Hayashi1996191270.530.510.3240.550.63 KalkoBCI6947220.39n/a0.3660.650.74 Lopezetal.20065036140.480.340.3640.650.78 Passosetal.2003292270.580.390.3340.500.68 Pedro1992181170.640.480.2940.420.60 Silveira200612660.750.180.2030.530.59 doi:10.1371/journal.pone.0017395.t001 Bat-FruitNetworks PLoSONE|www.plosone.org3February2011|Volume6|Issue2|e17395

PAGE 4

Bat-FruitNetworks PLoSONE|www.plosone.org4February2011|Volume6|Issue2|e17395

PAGE 5

Acomprehensivedatasetonbat-fruitinteractionsacrossthe Neotropics(365papersand4,100records)wasusedtorefineour classification,asitallowedustoseehowmanyrecordsof frugivoryeachspecieshas.Tobuildthisdatasetwestartedfrom informationpublishedinbat-plantdatabases[38],andcomplementeditwithliteraturemostlyfromSouthAmerica.Intotal, ourdatabasecomprises365papersand4,100recordsof interactions.Weconsideredonlyrecordswithtaxonomic resolutiontothespecieslevelforbothbatsandplants.This wayweseparatedsecondaryfromoccasionalfrugivoresinthe nectarivoroussubfamilies,basedonhowofteneachspecieshas beenrecordedfeedingonfruits. Finally,wetestedhowtheleveloffrugivoryofeachspecies explaineditsfunctionalroleinthenetwork.Weusedtwonetwork surrogatesforfunctionalrole,whichareexplainedinthenext section.Therelationshipsbetweenleveloffrugivoryand proportionofinteractions,andbetweenleveloffrugivoryand betweennesscentrality,weretestedwithKruskal-Wallistests.FunctionalrolesEachspeciesinamutualisticnetworkhasadifferentpatternof interaction,andthereforeplaysadifferentroleinthefunctioning andmaintenanceofitscommunity;thisisthespeciesÂ’functional roleorEltonianniche[39].Weassessedthefunctionalroleofeach speciesintheseeddispersalnetworkwithtwonetworksurrogates (seedetailsonthecalculationsin[40,41]).Thefirstsurrogatefor functionalrolewasthespeciesÂ’proportionofinteractions,i.e.to howmanyotherspeciesitisconnectedinthenetworkinrelation tothetotalnumberofpossiblepartnersavailable(i.e.relativeor normalizeddegreekr).Second,weassessedeachspeciesÂ’ betweennesscentrality( bc ),i.e.theproportionofshortestpaths (geodesics)thatcontainthetargetspeciesinrelationtoallexisting shortestpathsbetweenallspeciespairsinthenetwork.Apath betweentwospeciesinanetworkisdefinedasthenumberoflinks fromonetotheother[41];aspeciesthatisincludedinahigh proportionofgeodesicshasacentralpositioninthenetwork. ProportionofinteractionsreflectsthespeciesÂ’localnichebreadth,Figure1.Guildsandfunctionalrolesinthenetworks. Thestudiednetworkshaveamodularstructure,withprimaryfrugivorespositionedin thecenterofmostmodules,thusplayingimportantfunctionalrolesineachguild.Inthosegraphs,verticesrepresentspecies(circles=bats, triangles=plants),andspecieswithmorelinksorwhicharemorecentralwererepresentedclosertothecenterofthegraph.Thesizeofeachvertex isproportionaltohowcentralitisinthenetwork(betweennesscentrality),i.e.howimportantitsfunctionalroleis.Linksrepresentinteraction sof frugivoryandseeddispersal(lines).Colorsrepresentmodulesfoundinouranalysis.SpeciesnamesfollowthesamenumbersusedinAppendixS1. doi:10.1371/journal.pone.0017395.g001 Figure2.Differencesamongspeciesinnetworkproperties. Interestingdifferenceswerefoundregardingthefunctionalrolesofdifferent species.(a)Batsandplantsinteractedwithasimilarproportionofpartnersinthenetworks(similarproportionofinteractions),whereas(b)bats occupiedmorecentralpositions(higherbetweennesscentrality).Batspeciesconsideredasprimaryfrugivores(c)interactedwithahigherpropor tion ofplantsand(d)occupiedmorecentralpositionsthanbatspeciesconsideredassecondaryoropportunisticfrugivores.Themainhorizontalline showsthemedian,boxesrepresentquartiles,andwhiskersdepict95%intervals. doi:10.1371/journal.pone.0017395.g002 Bat-FruitNetworks PLoSONE|www.plosone.org5February2011|Volume6|Issue2|e17395

PAGE 6

Bat-FruitNetworks PLoSONE|www.plosone.org6February2011|Volume6|Issue2|e17395

PAGE 7

whereasbetweennesscentralityreflectshowimportantitsnicheis inthewholeinteractionsystem.Specieswithahighproportionof interactionscomparedtootherspeciesinthesamenetworkare called‘hubs’,whereasspecieswithahighbetweennesscentrality arecalled‘connectors’.RobustnesstoextinctionsTotestfortherobustnessofbat-fruitnetworkstospecies extinctions,weusedanetworksurrogateobtainedfroma procedurebasedoncumulativeremovalsofspeciesfromthe networkatrandom[42].First,weremovedonespeciesfromone sideofthenetwork(e.g.plants);whenanotherspeciesfromthe otherside(e.g.bats)wasconnectedonlytotheremovedspecies,it wasalsoremovedfromthenetwork(i.e.asecondaryloss).Then anotherspecieswasalsoremoved,withoutputtingthefirstone back,untilallspeciesfromthechosesidewereremoved.Inthis wayanextinctioncurvewasgeneratedbyplottingthenumberof remainingspeciesontheonesideagainstthecumulativenumber ofspeciesremovedfromtheotherside.Thesameprocedurewas carriedoutforbothsidesofeachnetwork,resultinginonecurve forplantsandanotherforanimals.Theareabeloweachcurve( R ) wascalculatedasameasureoftherobustnessofthewholesystem. R =1correspondstoaveryslowdecreaseinthecurve,andthus representsasysteminwhichmostplantsremainaftertheremoval ofmostanimals,orvice-versa; R =0correspondstoaveryfast decreaseinthecurve,andthusrepresentsasystemthatalready collapsesafterthefirstfewspecieshavebeenremoved.Weran100 randomizationsforeachnetwork.Thisanalysiswascarriedoutin thepackagebipartiteforR.Ontheonehand,itisknownthat someofthestudiedbatspecies,forinstance,doalsofeedonother itemsuchaspollen,nectarandinsects[20].Ontheotherhand, someofthestudiedplantspeciesaredispersedalsobyotherkinds ofanimals,suchasbirdsorprimates[43].Therefore,itis importanttosaythatspeciesremovalsinoursimulationsdonot necessarilyrepresentrealextinctionsinnature,asweweredealing withseeddispersalsystems,andnotwholeecologicalcommunities. Removalsinourstudyrepresentexclusionofparticularspecies fromtheseeddispersalservice. Atspecieslevel,wewantedtotesthowtheremovalofparticular speciesaffectedthewholenetworkstructure,dependingonitslevel offrugivory.Todothiswealsosimulatedsingle-speciesremovals ineachnetworkinAtaque1.0(M.A.M.deAguiarandF.M.D. Marquitti,designedforthisstudy)usingajackknifeprocedure.We excludedonespeciesanditsinteractionsinanetworkatatime, andobservedthepercentageofchangethatthisremovalcausedin thedegreeofnestedness( NODFr).Thisvaluewascalculatedas: NODFr=( NODFobs– NODFori)/ NODFori,where NODFobsisthe observedvalueof NODF afteraspecieswasremoved,and NODForiistheoriginalvalueof NODF ofthecompletematrix.The relationshipbetweenleveloffrugivoryand NODFrwasassessed withSpearmancorrelations.Asnestednessishypothesizedas improvingnetworkrobustness[28],weassumedthatadecreasein nestednesshasanegativeeffectonthenetworkbydecreasingits robustness.ResultsTotalnetworksize(numberofbatandplantspecies)varied from12to69(average 6 SD:31 6 18).Corroboratingourfirst prediction,allninenetworkswerenested( NODF =0.56 6 0.12,all P , 0.01)(Table1).Furthermore,complementaryspecialization wasalsosignificant( H2’ =0.37 6 0.10,allP , 0.001). Modularitywaslowinthebat-fruitnetworks( M =0.32 6 0.07, allP , 0.01)withanaverageof4 6 1modulesineachnetwork (varyingfrom3to6).Ismostcases,speciesofthethreemain generaoffrugivorousphyllostomidswerefoundtogetherinthe samemodulewiththeplantgeneraassumedtobetheirpreferred ( Artibeus =0.67,G=7.94,P=0.004; Carollia =0.78,G=16.81, P , 0.001; Sturnira =0.67,G=7.94,P=0.004),althoughsome plantgeneraweredispersedbymorethanonebatgenus(Fig.1). Eachspecies,batorplant,interactedonaveragewithabout one-thirdofallpartnersavailableineachnetwork.Theproportion ofinteractionswassimilarbetweenbats( kr=0.29 6 0.24)and plants( kr=0.28 6 0.18)(df=94,t=0.29,P=0.77).Betweenness centralitywashighlyvariableamongspeciesineachnetwork.On averagebats( bc =0.10 6 0.15)hadhighervaluesthanplants ( bc =0.03 6 0.05)(df=94,t=4.48,P , 0.001)(Fig.2).Primary frugivoresshowedhighervaluesthansecondaryandoccasional frugivores,bothforproportionofinteractions(N=87,df=2, K=16.76,P , 0.001)andforbetweennesscentrality(N=87, df=2,K=9.91,P=0.007)(Fig.2). Therobustnessofbat-fruitnetworkstocumulativeextinctions wasrelativelyhigh,bothforbats( R =0.55 6 0.10,range0.41–0.69) andplants( R =0.68 6 0.09,range0.58–0.84)(Fig.3).Therewas alsoahighrobustnesstotheremovalofsinglespecies. Proportionalchangeinnestedness( NODFr)variedfrom0to 3.7%,andwaslowerthan1%inmostcases.Furthermore,there werehardlyanysecondarylosses( SLr=0.0/0.00:median/ quartiles,varyingfrom0to3.7%,mostcases=0).Removalof specieswhichinteractedwithahigherproportionofavailable partnerscausedlargerchangesinnestednessinbothbats(N=87, r= 2 0.46,P , 0.001)andplants(N=198,r= 2 0.44,P , 0.001) (Fig.4).Theremovalofprimaryfrugivorescausedlargerdecreases innestednessthantheremovalofsecondaryoroccasional frugivores(N=87,df=2,K=6.87,P=0.03)(Fig.4).DiscussionInourstudyweaddanimportantpiecetothepuzzleofseed dispersalnetworksbydescribingthestructureandrobustnessof bat-fruitnetworks,andshowinghowtheyareinfluencedby dietaryspecializationatboththenetworkandspecieslevel.Batfruitnetworksshowedtoberobustsystemswithlowcomplementaryspecialization,butalsoamodularstructure,inwhichprimary frugivoreshavethemostimportantfunctionalroles. Thefirstpointtoexamineiswhatbat-fruitnetworkshavein commonwithothernetworksoffacultativemutualism.The pervasivenestedtopologyobservedinbat-fruitnetworks,aswellas inmanyothermutualisticsystems,suggeststhatsimilarprocesses maybestructuringverydifferentfacultativemutualisms,ranging frompollination[44]tomarinecleaningsymbiosis[45].Nestednessisassumedtoincreaseresilienceandbiodiversity[28],since specieswithfewinteractionstendtobemorefragilethanspecies withmanyinteractions,whichformacoreofhighly-connected andresistantspecies.Thecoresofthestudiedbat-fruitnetworks werecomposedbythethreemainfrugivorousbatgenera( Artibeus , Carollia and Sturnira )andtheirfivemainfood-plantgenera( Cecropia , Ficus , Piper , Solanum and Vismia ).Someauthorshavestatedthata nestedtopologymayalsoemergefromrandomnetworks[27]. However,thisisstillacontroversyamongnetworkecologists.TheFigure3.Robustnesstocumulativespeciesremoval. Thesimulationsofcumulativeremovalsofspeciesshowedthatbat-fruitnetworksare veryrobustbothtoremovalsofbatsandplants,asextinctioncurvesdeclinedslowlyonaverage. doi:10.1371/journal.pone.0017395.g003 Bat-FruitNetworks PLoSONE|www.plosone.org7February2011|Volume6|Issue2|e17395

PAGE 8

mainreasonforfindingasignificantlynestedstructureinrandom matriceswasaconceptualerrorintheoriginalnestednessformula, whichwasderivedfromthe T metric[34].Thisflawwasshown recentlyandanewnestednessmetrichasbeenproposed: NODF , whichismuchbettertunedwiththeoriginalconceptofnestedness andfitsbettertostudiesonmutualism[26].Thisnewmetric, whichweusedinouranalysis,wasnotevaluatedbycriticsof nestednessanalysis,andcontraryto T ,itgivesconsistentlylow valuesforrandommatrices. Thelowlevelofcomplementaryspecializationobservedinbatfruitnetworkswassimilartovaluesfrombird-fruitnetworks (mediancloseto0.30)[46].Therefore,incontrasttopollination networks,lowcomplementaryspecializationdoesindeedseemto beacommonpropertyofseeddispersalnetworks.Itisassumed thatseeddispersalisamorediffuseinteractioncomparedto pollination,becauseitismoredifficultforplantstodevelop mechanismsthatrestrictaccesstofruits.Furthermore,tobea legitimateseeddisperser,ananimalhasjusttoavoidkillingthe seedsandthentransportthemawayfromthemother-plant[47]. However,tobealegitimatepollinatoramuchfinermatchis required,ultimatelyaimingatcarryingpollenfromflowersofone individualtootherindividualplantofthesamespecies[48]. Becausespecializationdependsnotonlyonthetypeofinteraction, butalsoonthegroupsoforganisms,itwillbeinterestinginthe futuretostudybat-flowernetworksinordertotestifcomplementaryspecializationishigherthaninbat-fruitnetworks. Itisinterestingtonotethattheecologicalandnetworkconcepts ofspecializationusedinourstudyrevealedcontrastingrelationships.Ontheonehand,accordingtotheecologicalconceptof dietaryspecializationusedhere,aphyllostomidbatspeciesthat hasahighleveloffrugivorysuchas Sturniralilium [49],andso dependsonfruitsforliving,maybeconsideredasaspecialist, comparedtootherphyllostomidsviewedasgeneralistsforfeeding equallyonmanykindsoffood(e.g.fruits,nectar,insects),suchas Phyllostomushastatus [50].Ontheotherhand,thosespeciesthatare ecologicallymorespecializedturnedouttobeverygeneralistic accordingtothenetworkconcept,astheyinteractedonmanyfruit specieswithintheirnetworks.Therearedifferencesevenbetween speciesofthesamegenus,asforinstanceall Carollia batsare primaryfrugivores,but C.perspicillata feedsonamuchlarger varietyofplantsthan C.castanea ,atleastinBarroColorado, Panama.Therefore,wehavetobecarefulwheninterpreting specializationinanetworkcontext,andweneedtostateclearly whichecologicalconceptisbeingoperationalizedwithwhich networkconcept.Manyspeciesidentifiedasgeneralistsinthe networksenseareprobablyinfactspecialistsaccordingtoabroad ecologicalconceptsuchasdietaryspecialization.Inthisstudy, ecologicallyspecializedfrugivores(i.e.primaryfrugivores)were showntobemoreimportantformaintainingthewholenetwork structure.Inthisway,comparedwithotherstudieswehavegone onestepfurtherintheassessmentofthefunctionalroleofdifferent speciesinmutualisticnetworks,becausewehavedirectlylinked networkimportancetofunctionalrole. Intermediatenestedness,lowcomplementaryspecializationand lowmodularityseemtoleadtoacohesivestructurewithabalance betweenredundancywithinmodulesandcomplementarityamong modules,becausesomekeybatgeneraareresponsiblemainlyfor dispersingtheirpreferredplant genera,andsoeachnetworkis composedofmoduleswithaphylogeneticsignal.Itisinterestingto noticethatthegenus-genusassociationsuncoveredinpopulation studiesofbatsandplants[18]seemtoinfluencethestructureof moduleswithinbat-fruitnetworks.Therelationshipsbetween Artibeus Ficus , Carollia Piper ,and Sturnira Solanum wereconsistentamong differentlocalnetworksdespitesomegeographicalvariations.Infact, Figure4.Changescausedbytheremovalofsinglespecies. With simulationsofsingle-speciesremovals,weobservedthatspeciesof both(a)batsand(b)plantsthatinteractwithahigherproportionof mutualisticpartnersaremoreimportantformaintainingthewhole networkstructure,astheirremovalcauseslargerdecreasesin nestedness.Furthermore,(c)theremovalofbatspeciesconsideredas primaryfrugivorescausedalsolargerdecreasesinnestednessinthe wholenetwork. doi:10.1371/journal.pone.0017395.g004 Bat-FruitNetworks PLoSONE|www.plosone.org8February2011|Volume6|Issue2|e17395

PAGE 9

thosegeographicvariationscanbeexplainedbydifferencesinthe localavailabilityofplantspeciesuponwhichfrugivorousbatsfeed, whichiscausedmainlybydifferencesinthegeographicdistribution ofbatsandplantsthatinteractwitheachother.Thispattern reinforcestheassumedbackgroundofacoevolutionaryhistory betweenthosebatandplantgenera[51].Amongecologicallysimilar species,suchasthoseofthegenera Carollia and Sturnira ,itisverylikely thatacoupleoffactors,inparticularfruitsecondarymetabolites,play majorroles[52,53]inpermittingresourcepartitioning,and ultimatelytheircoexistence[54]. Specializationandredundancymightexplainthehighrobustnessobservedinbat-fruitnetworks.Highrobustnesstorandom removalofnodes(‘error’inthenetworkjargon)isacommon featureofmanycomplexnetworks[55],includingmutualistic networks[56].Itistypicalofnetworkswithascale-freeorbroadscaletopology,whereonlyafewspecieshaveadisproportionally highnumberofinteractions,andmostspecieshavefew interactions[55].Inthecaseofbat-fruitsystems,thistolerance isprobablyenhancedbythehighredundancy.Itseemsthatwithin eachmodule,batsofthesamegenusplayredundantrolesinthe dispersalofplantspeciesoftheirpreferredgenera.Inturn, modulesarecomplementarytoeachother,asspeciesineach moduleareresponsibleforaparticularpartofthewholedispersal service.Finally,despitethosegenus-genusassociations,manyplant generaaredispersedalsobyotherbatgeneraandnotonlytheir mainpartners.Theresultisarobustsystem,inwhichthereare back-upsbothwithinandoutsideeachmodule,ensuringthatmost plantspeciescontinuetobeservicedevenintheabsenceoftheir mainmutualists.Thisfindingisofgreatrelevancetoconservation, asbatsarepredominantlyinvolvedinseeddispersalservicesat pioneerstages,thereforebeingthemaingroupresponsiblefor forestregeneration[10,11]. Inconclusion,dietaryspecialization(hereassessedaslevelof frugivory)seemstobeanimportantstructuringfactorinbat-fruit networks.Itwouldbeinterestingforfuturestudiestogoonestep furtherandstudyhowphysiologicaldifferencesamongbatspecies (e.g.theabilitytocopewithparticularsecondarymetabolites)may explainnichesegregationatnetworklevel.Aftermorestudiesare conducted,wemaybeabletousenetworkpropertiestohelp defineconservationprioritiesandevenrestoredegradedareasina moreefficientway.Forexample,plantspeciespointedoutashubs inseeddispersalnetworkscanbegoodcandidatesforreforestation programs,astheyarelikelytoattractmoredisperserspeciesand accelerateregeneration.Andspeciesidentifiedasconnectorsmay bealsoimportant,astheywillhelptoincreasethesystem’s cohesivenessand,ultimately,robustness.SupportingInformationAppendixS1Datasetsonbat-fruitinteractionsusedinour analysis. (PDF)AppendixS2Matricesanalyzedinthisstudy. (PDF)AcknowledgmentsManycolleaguesfromUniversita ¨tUlm,UniversidadeFederaldeSa ˜o Carlos,Estacio ´nBiolo ´gicadeDon ˜ana,UniversidadeEstadualde Campinas,UniversidadedeSa ˜oPauloandUniversidadeEstadualPaulista contributedtothisworkindifferentways.CarstenDormann,Ma ´rio Almeida-Neto,NicoBlu ¨thgen,RogerGuimera `andVladimirBatagelj helpedususetheirnetworksoftware.DalvaMatosprovideduswithlab infrastructureatUFSCar.AuthorContributionsConceivedanddesignedtheexperiments:MARMPRG.Performedthe experiments:MARMFMDM.Analyzedthedata:MARMFMDM. Contributedreagents/materials/analysistools:FMDMMAMdA.Wrote thepaper:MARMEKVKPRGPJ.Designedthesoftwareusedinanalysis: FMDMPRGMAMdA.References1.WrightSJ(2002)Plantdiversityintropicalforests:areviewofmechanismsof speciescoexistence.Oecologia130:1–14. 2.DennisAJ,GreenRJ,SchuppEW,WestcottDA(2007)Seeddispersal:theory anditsapplicationinachangingworld.Wallingford:CABIPublishing.720p. 3.JohnsonS(2002)Emergence:theconnectedlivesofants,brains,citiesand softwareNewYork:PenguinBooks.pp288. 4.WalkerBH(1992)Biodiversityandecologicalredundancy.Conservation Biology6:18–23. 5.BascompteJ,JordanoP(2007)Plant-animalmutualisticnetworks:the architectureofbiodiversity.Ann ualReviewofEcologyEvolutionand Systematics38:567–593. 6.KalkoEKV,HandleyCOJ,HandleyD(1996)Organization,diversityandlongtermdynamicsofaneotropicalbatcommunity.In:CodyM,SmallwoodJ,eds. Long-termstudiesinvertebratecommunities.LosAngeles:AcademicPress.pp 503–553. 7.RootRB(1967)Thenicheexploitationpatternoftheblue-graygnatcatcher. EcologicalMonographs37:317–350. 8.MelloMAR,MarquittiFMD,Guimara ˜esJr.PR,JordanoP,KalkoEKV(2008) Smalldifferencesmayleadtobigconsequences:batandbirdseeddispersal networks(Abstract,LuisFBarcardiAward2008).Biotropica40:777. 9.TerborghJ,PitmanN,SilmanM,SchichterH,Nu ´n ˜esVP(2002)Maintenance oftreediversityintropicalforests.In:LeveyDJ,SilvaWR,GalettiM,eds.Seed dispersalandfrugivory:ecology,evolution,andconservation.NewYork:CABI Publishing.pp1–17. 10.Galindo-Gonza ´lesJ,GuevaraS,SosaVJ(2000)Batandbird-generatedseed rainsatisolatedtreesinpasturesinatropicalrainforest.ConservationBiology 14:1693–1703. 11.MuscarellaR,FlemingTH(2007)Theroleoffrugivorousbatsintropicalforest succession.BiologicalReviews82:573–590. 12.OlesenJM,BascompteJ,DupontYL,JordanoP(2007)Themodularityof pollinationnetworks.ProceedingsoftheNationalAcademyofSciencesofthe UnitedStatesofAmerica104:19891–19896. 13.GuimaraesPR,Rico-GrayV,OliveiraPS,IzzoTJ,dosReisSF,etal.(2007) Interactionintimacyaffectsstructureandcoevolutionarydynamicsinmutualisticnetworks.CurrentBiology17:1797–1803. 14.daSilvaAG,GaonaO,MedellinRA(2009)Dietandtrophicstructureina communityoffruit-eatingbatsinLacandonForest,Mexico.Journalof Mammalogy89:43–49. 15.GianniniNP,KalkoEKV(2004)Trophicstructureinalargeassemblageof phyllostomidbatsinPanama.Oikos105:209–220. 16.DatzmannT,vonHelversenO,MayerF(2010)Evolutionofnectarivoryin phyllostomidbats(PhyllostomidaeGray,1825,Chiroptera:Mammalia).BMC EvolutionaryBiology10:165. 17.LobovaTA,GeiselmanCK,MoriSA(2009)Seeddispersalbybatsinthe Neotropics.NewYork:NewYorkBotanicalGardenPress.465p. 18.FlemingTH(1986)Opportunismversusspecialization:theevolutionoffeeding strategiesinfrugivorousbats.In:EstradaA,FlemingTH,eds.Frugivoresand SeedDispersal.Dordrecht:Dr.W.JunkPublishers.pp105–118. 19.FlemingTH(1982)Foragingstrategiesofplant-visitingbats.In:KunzTH,ed. Ecologyofbats.NewYork:PlenumPress.pp287–326. 20.MelloMAR,SchittiniGM,SeligP,BergalloHG(2004)Seasonalvariationin thedietofthebat Carolliaperspicillata (Chiroptera:Phyllostomidae)inanAtlantic forestareainsoutheasternBrazil.Mammalia68:49–55. 21.TschapkaM(2005)Reproductionofthebat Glossophagacommissarisi (Phyllostomidae:Glossophaginae)intheCostaRicanrainforestduringfrugivorousand nectarivorousperiods.Biotropica37:408–414. 22.Zorte ´aM(2003)Reproductivepatternsandfeedinghabitsofthreenectarivorous bats(Phyllostomidae:Glossophaginae)fromtheBrazilianCerrado.Brazilian JournalofBiology63:159–168. 23.BravoA,HarmsKE,EmmonsLH(2010)Puddlescreatedbygeophagous mammalsarepotentialmineralsourcesforfrugivorousbats(Stenodermatinae)in thePeruvianAmazon.JournalofTropicalEcology26:173–184. 24.ThompsonJN(2005)Thegeographicmosaicofcoevolution.London:The UniversityofChicagoPress.443p.Bat-FruitNetworks PLoSONE|www.plosone.org9February2011|Volume6|Issue2|e17395

PAGE 10

25.ThiesW,KalkoEKV(2004)Phenologyofneotropicalpepperplants (Piperaceae)andtheirassociationwiththeirmaindispersers,twoshort-tailed fruitbats,CarolliaperspicillataandC.castanea(Phyllostomidae).Oikos104: 362–376. 26.Almeida-NetoM,Guimara ˜esPR,Guimara ˜esJr.PR,LoyolaRD,UlrichW (2008)Aconsistentmetricfornestednessanalysisinecologicalsystems: reconcilingconceptandmeasurement.Oikos117:1227–1239. 27.Blu ¨thgenN,FrundJ,VazquezDP,MenzelF(2008)Whatdointeraction networkmetricstellusaboutspecializationandbiologicaltraits?Ecology89: 3387–3399. 28.BastollaU,FortunaMA,Pascual-GarciaA,FerreraA,LuqueB,etal.(2009) Thearchitectureofmutualisticnetworksminimizescompetitionandincreases biodiversity.Nature458:1018-U1091. 29.MeloFPL,Rodriguez-HerreraB,ChazdonRL,MedellinRA,CeballosGG (2009)SmallTent-RoostingBatsPromoteDispersalofLarge-SeededPlantsina NeotropicalForest.Biotropicainpress. 30.NogueiraMR,PeracchiAL(2003)Fig-seedpredationby2speciesof Chiroderma : discoveryofanewfeedingstrategyinbats.JournalofMammalogy84:225–233. 31.BatageljV,MrvarA(2008)Pajek–programforlargenetworkanalysis. Ljubljana. 32.DormannCF,GruberB,Fru ¨ndJ(2008)Introducingthebipartitepackage: analyzingecologicalnetworks.RNews8:8–11. 33.Guimara ˜esPR,Guimara ˜esP(2006)Improvingtheanalysesofnestednessfor largesetsofmatrices.EnvironmentalModellingandSoftware21:1512–1513. 34.AtmarW,PattersonBD(1993)Themeasureoforderanddisorderinthe distributionofspeciesinfragmentedhabitat.Oecologia96:373–382. 35.Blu ¨thgenN,MenzelF,Blu ¨thgenN(2006)Measuringspecializationinspecies interactionnetworks.BMCEcology6:1–12. 36.Guimera `R,AmaralLAN(2005)Functionalcartographyofcomplexmetabolic networks.Nature433:895–900. 37.KisslingWD,GaeseKB,JetzW(2009)Theglobaldistributionoffrugivoryin birds.GlobalEcologyandBiogeography18:150–162. 38.GeiselmanCK,MoriSA,BlanchardF(2002)Databaseofneotropicalbat/plant interactions. 39.EltonC(2001)AnimalEcology.Chicago:TheUniversityofChicagoPress.296p. 40.CostaLD,RodriguesFA,TraviesoG,BoasPRV(2007)Characterizationof complexnetworks:asurveyofmeasurements.AdvancesinPhysics56:167–242. 41.NooyW,MrvarA,BatageljV(2005)Exploratorysocialnetworkanalysiswith Pajek.NewYork:CambridgeUniversityPress.334p. 42.BurgosE,CevaH,PerazzoRPJ,DevotoM,MedanD,etal.(2007)Why nestednessinmutualisticnetworks?JournalofTheoreticalBiology249: 307–313. 43.ShanahanM,SoS,ComptonSG,CorlettR(2001)Fig-eatingbyvertebrate frugivores:aglobalreview.BiologicalReviews76:529–572. 44.BascompteJ,JordanoP,MelianCJ,OlesenJM(2003)Thenestedassemblyof plant-animalmutualisticnetworks.ProceedingsoftheNationalAcademyof SciencesoftheUnitedStatesofAmerica100:9383–9387. 45.OllertonJ,McCollinD,FautinDG,AllenGR(2007)FindingNEMO: nestednessengenderedbymutualisticorganizationinanemonefishandtheir hosts.ProceedingsoftheRoyalSocietyB-BiologicalSciences274:591–598. 46.Blu ¨thgenN,MenzelF,HovestadtT,FialaB,BluthgenN(2007)Specialization, constraints,andconflictinginterestsinmutualisticnetworks.CurrentBiology17: 341–346. 47.HoweHF(1984)Constraintsontheevolutionofmutualisms.TheAmerican Naturalist123:764–777. 48.HoweHF,SmallwoodJ(1982)Ecologyofseeddispersal.AnnualReviewof EcologyandSystematics13:201–228. 49.MelloMAR,KalkoEKV,SilvaWR(2008)Dietandabundanceofthebat Sturniralilium (Chiroptera)inaBrazilianmontaneAtlanticforest.Journalof Mammalogy89:485–492. 50.CampbellP,AkbarZ,AdnanAM,KunzTH(2006)Resourcedistributionand socialstructureinharem-formingOldWorldfruitbats:variationsona polygynoustheme.AnimalBehaviour72:687–698. 51.HeithausER(1982)Coevolutionbetweenbatsandplants.In:KunzTH,ed. Ecologyofbats.NewYork:PlenumPress.pp327–367. 52.BianconiGV,MikichSB,TeixeiraSD,MaiaBHLNS(2007)Attractionoffruiteatingbatswithessentialoilsoffruits:apotentialtoolforforestrestoration. Biotropica39:136–140. 53.CipolliniML,LeveyDJ(1997)Whyaresomefruitstoxic?Glykoalkaloidsin Solanum andfruitchoicebyvertebrates.Ecology78:782–798. 54.Marinho-FilhoJS(1991)Thecoexistenceoftwofrugivorousbatspeciesandthe phenologyoftheirfoodplantsinBrazil.JournalofTropicalEcology7:59–67. 55.AlbertR,JeongH,BarabasiAL(2000)Errorandattacktoleranceofcomplex networks.Nature406:378–382. 56.MemmottJ,WaserNM,PriceMV(2004)Toleranceofpollinationnetworksto speciesextinctions.ProceedingsoftheRoyalSocietyLondonB271:2605–2611.Bat-FruitNetworks PLoSONE|www.plosone.org10February2011|Volume6|Issue2|e17395


printinsert_linkshareget_appmore_horiz

Download Options

close


  • info Info

    There are only PDFs associated with this resource.

  • link PDF(s)



Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.