Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Do bees matter to cranberry?
The effect of bees, landscape, and local management on cranberry yield
By
Hannah Rodman Gaines Day
A dissertation submitted in partial fulfillment of
the requirements for the degree of
Doctor of Philosophy
(Zoology)
at the
UNIVERSITY OF WISCONSIN‐MADISON
2013
Date of final oral examination: 10/09/13 The dissertation is approved by the following members of the Final Oral Committee: Claudio Gratton, Professor, Entomology Russ Groves, Associate Professor, Entomology Shawn Steffan, Assistant Professor, Entomology Volker Radeloff, Professor, Forest and Wildlife Ecology Neal Williams, Associate Professor, Entomology (University of California, Davis)
i
TABLEOFCONTENTSDEDICATION……………………………………………………………………………………………………………...…iiiACKNOWLEDGEMENTS………………………………………………………………………………………………..ivTHESISABSTRACT…………………………………………………………………………………………………….....viTHESISINTRODUCTION……………………………………………………………………………………………..…1CHAPTER1:Non‐bioticfactorscontributetocranberrypollination Abstract………………………………………………………………………………………………………………7
Introduction………………………………………………………………………………………………………..8
Methods…………………………………………………………………………………………………………….11
Results………………………………………………………………………………………………………………16
Discussion…………………………………………………………………………………………………………18
Literaturecited…………………………………………………………………………………………………..23
Figures………………………………………………………………………………………………………………27
Appendix…………………………………………………………………………………………………………...31
CHAPTER2:Thecontributionofhoneybeestocranberryyieldvarieswiththeamountofwoodlandinthesurroundinglandscape Abstract……………………………………………………………………………………………………………..36
Introduction………………………………………………………………………………………………………38
Methods…………………………………………………………………………………………………………….40
Results………………………………………………………………………………………………………………42
Discussion………………………………………………………………………………………………………….44
Literaturecited…………………………………………………………………………………………………..48
Figures………………………………………………………………………………………………………………51
CHAPTER3:Localandlandscapefactorsinfluencenativebees,yield,andthecontributionofbeestoyield Abstract……………………………………………………………………………………………………………..54
ii
Introduction………………………………………………………………………………………………………55
Methods…………………………………………………………………………………………………………….58
Results………………………………………………………………………………………………………………64
Discussion………………………………………………………………………………………………………….67
Literaturecited…………………………………………………………………………………………………..71
Figures………………………………………………………………………………………………………………76
Appendix…………………………………………………………………………………………………………...84
CHAPTER4:Understandingthebarrierstoimplementationofon‐farmhabitatconservationofnativebeesinWisconsincranberry Abstract……………………………………………………………………………………………………………..89
Introduction………………………………………………………………………………………………………90
Methods…………………………………………………………………………………………………………….93
Results………………………………………………………………………………………………………………95
Discussion………………………………………………………………………………………………………..100
Literaturecited………………………………………………………………………………………………..104
Figures…………………………………………………………………………………………………………….108
Appendix…………………………………………………………………………………………………………117
THESISCONCLUSION………………………………………………………………………………………………...129
iii
Formybuddyoldpal,Becky.Myfirstentomologyteacherandagreatfriend.
Thisoneisforyou.
iv
ACKNOWLEDGEMENTS Therearemanypeoplewhohavehelpedandguidedmethroughoutmygraduate
schoolcareerbutnoneasmuchasmyadvisor,ClaudioGratton.UnderClaudio’sguidanceI
havebecomeabettercommunicatorandscientist.Icannotimagineabetteradvisor,
teacher,androlemodel.Thankyou,Claudio,forbeingawesome.YouaretrulygreatandI
willforeverappreciateyourpatience,dedication,andenthusiasm.Iamalsogratefultothe
cranberrygrowerswhoallowedmetoconductmyresearchontheirmarshes.Iam
humbledbytheirgenerosityandkindness.IamespeciallygratefultoC.J.SearlesCranberry
forallowingmetoconductmyintensivefieldstudyontheirmarshandGaynorCranberry
Inc.forallowingmetodigcranberryplantsoutoftheirmarshformygreenhousestudy.
Iwouldalsoliketothankmycommitteemembers,VolkerRadeloff,NealWilliams,
RussGroves,andShawnSteffanfortheircontinuedsupportandfeedbackthroughoutmy
PhD.IalsothankPaulMitchellforhisenthusiastichelpwithmyhoneybeechapter.I
wouldliketothankthemembersoftheGrattonLabfortheirhelpwithdatacollection,
statisticalanalysis,andgeneralfeedback,especiallyRachelMallinger,TaniaKim,Brian
Spiesman,andDavidHoekman.IwouldalsoliketothankDavidDuncanforhelpwith
statisticalanalysis.Thecranberryresearchgroup,especiallyEricZeldinandBeth
Workmaster,providedvaluablefeedbackonfieldandgreenhouseexperiments.John
Ascher,MikeArduser,andJasonGibbsspentmanyhoursidentifyingandverifyingmy
nearly10,000beespecimens.DanCariveauwasasoundingboardthroughoutmy
dissertationandprovidedvaluablefeedbackonearlydrafts.IwouldliketothankBritt
SearlesandtheSandhiltonforprovidingaplaceformetostayduringmyfieldwork.Jayne
v
SojkaandDanMahrhelpedwithmyinitialfieldsiteselectionandprovidedcontinued
supportthroughoutmyPhD.
Iamalsogratefultothefieldandlabassistantswhohelpedmecollectandprocess
mydataespeciallyCollinSchwantesandAmandaRudiewhostuckwithmethrough18
hourfielddays,blisteringheat,numbingcold,wind,rain,anddarknesstogetthedata
collected.Theirdedicationandeffortwasremarkable.IalsothankEmilyFricke,Adam
Higgins,CarlKaiser,LauraSmith,MerrittSingleton,AllisonDehnel,BrianHedberg,Chase
Fritz,MollyWaytes,GavinJones,FedericaLacasella,JuliaTiede,AliNelson,Alex
Wenninger,CaitlinBergstrom,EricWiesman,ChristopherWatson,SachaHorn,andScott
Leefortheirhelpinthefieldandlab.
Icouldn’thavefinishedmydissertationwithouttheweeklyaccountabilitymeetings
withAlexLyon,MarySaunders,andDavidDuncan.Iamalsogratefultothefolksatthe
WritingCenterforprovidingstructureandwritingadvicethroughoutthewritingprocess.
IamgratefultoKatharinaUllmannandMiaParkforbeinggreatfriendsandgiving
meadviceonfieldworkandresearch.
Andfinally,Iamgratefultomyfamily,friends,andmyhusband,Steve,for
supportingme,lovingme,andenrichingmylife.Icouldn’thavedoneitwithoutyou.
vi
THESISABSTRACT
Nativebeesprovidethemajorityofpollinationservicestowildandcultivated
floweringplants.Unfortunately,astheareaofmodernagricultureexpands,fragmenting
anddestroyingnaturalhabitat,thepersistenceofnativebeesandthepollinationservices
theyprovideisthreatened.Withthecurrentdeclineinmanagedbees,understandinghow
localmanagementdecisionsandlandscapefactorsinfluencenativebeesandtheir
contributiontocroppollinationbecomesmoreurgent.UsingWisconsincranberryasa
modelsystem,Iexaminedthepollinationrequirementsofcranberry,thecontributionof
beestopollination,andtheinfluenceoflocalandlandscapefactorsonnativebees.I
furtherinvestigatedthebarrierstofarmeradoptionofon‐farmconservationprogramsfor
nativebees.Contrarytopreviousstudies,Ifoundthatnon‐bioticfactorscontribute
significantlytocranberrypollination(chapter1).Myresearchdemonstratesthat,evenin
theabsenceofbees,cranberryisabletoproducefruit.Fruitproductionwas,however,
increasedwhenbeeswerepresentsupportingthepracticeofusinghoneybeesfor
pollination.Atafarmscale,cranberryyieldwaspositivelycorrelatedwithanincreasing
stockingdensityofhoneybeehives,butonlyatfarmslocatedinlow‐woodlandlandscapes
(chapter2).Whenhoneybeeswereabsent,allfarmshadsignificantyield,butthoseinhigh
woodlandlandscapeshadamarginallyhigheryieldthanfarmsinlowwoodland
landscapes.Farmsinhighwoodlandlandscapesalsohadahigherabundanceandrichness
ofnativebeesalthoughtherewasonlyaweakrelationshipbetweennativebeesandyield
(chapter3).Thecontributionofbeestocranberrypollinationincreasedwithincreasing
floweringuprightdensity,suggestingthatlocalmanagementcanenhancethecontribution
vii
bybees.Additionalresultsindicatethatcranberrygrowersareinterestedincreating
pollinatorhabitatontheirfarmsbutarelimitedbyalackoftechnicalsupportand
perceivedtimeandfinancialcommitments(chapter4).Thisdissertationcontributesto
ourunderstandingofcranberrypollinationbiologyandhowlocalandlandscapefactors
influencebeesandtheircontributiontoyield.Iprovidepracticalguidelinesforgrowerson
pollinationmanagementforcranberryandforconservationprofessionalsonincreasing
cranberrygrowerparticipationinfederalpollinatorconservationprograms.
1
THESISINTRODUCTION
Theexpansionofmodernagricultureandhumanactivityhasledtothealteration
andfragmentationofthelandscape(Vitouseketal.1997,Foleyetal.2005).Theresulting
lossofhabitathasprofoundeffectsontheabilityoforganismstopersist(reviewedby
Fahrig2003)andtheprovisioningofecosystemservices(Lucketal.2003).Theresponse
ofspeciesthatdirectlyinfluenceecosystemservicessuchaspollinationisofparticular
importance.Insect‐derivedecosystemservicesaloneareworthanestimated$8billionper
yearanddirectlyimpacthumanwellbeing(LoseyandVaughan2006,Eilersetal.2011).
Pollinationisaparticularlyimportantserviceinagro‐ecosystemsasitisrequiredby
orbenefitstwo‐thirdsofglobalcrops(Kleinetal.2007).Ashoneybeescontinuetodecline
(Ellisetal.2010),farmersmayneedtoseekalternativewaystopollinatetheircrops.
Nativebeesprovidevaluablecroppollinationservices(LoseyandVaughan2006,Winfree
etal.2008)butareatriskofdeclineduetohabitatlossandfragmentation,intensified
agriculture,andagri‐chemicalexposure(Pottsetal.2010).Sincebeesarecentralplace
foragersandhavelimitedflightdistances(Greenleafetal.2007,Zurbuchenetal.2010),
theyrequirereadilyaccessiblenestingandforagingresourcesthroughouttheirentireflight
season.Becauseofagriculturalpracticesthatdisturbthesoil(e.g.,tilling),limitthe
abundanceoffloweringweedyspecies(e.g.,useofherbicides),andcultivatecropsas
monocultures,modernagriculturallandscapesprovidepoorhabitatforbees.
Onewayforfarmerstolessentheeffectsofmodernagriculturalpracticesonbeesis
throughtheadoptionofon‐farmhabitatconservationprograms.Providinghabitatwithin
agriculturallandscapescouldgreatlyenhancelocalresourceavailabilityandtheabilityof
1
2
beestopersistandprovidepollinationserviceswithincropfields.Cost‐shareconservation
programsthroughtheUnitedStatesDepartmentofAgricultureNaturalResources
ConservationService(USDANRCS)providegrowerswithfinancialincentivestoinstall
pollinatorhabitat.Inaddition,forfarmersofpollinator‐dependentcrops,suchas
cranberries,theadoptionoftheseprogramscouldprovideadirecteconomicreturninthe
formofincreasedpollinationservicesfromnativebees.Inorderforthisstrategytobe
successful,however,growersmustadoptthesepractices.Becauseoftheirdependenceon
pollinators,demonstratedcommitmenttoenvironmentalstewardshipthroughthe
adoptionofIntegratedPestManagement(IPM)programs,andtheavailabilityofnon‐crop
habitatontheirfarms,cranberrygrowersshouldbeidealcandidatesforadoptionof
federalon‐farmconservationprograms(ColquhounandJohnson2010).
Beforewecandeterminethesignificanceofpollinatordeclinetocropproduction,
however,weneedtounderstandthepollinationrequirementsofthecropanddetermine
howmuchbeescontributetoitspollination.Therefore,theobjectiveofthisstudywasto
gainabetterunderstandingofthefactorsthataffectpollinationandnativebeesinthe
Wisconsincranberrysystem.
Cranberryisgenerallyconsideredapollinator‐dependentcrop,althoughestimates
ofyieldlossintheabsenceofbeesvarygreatly(DelaplaneandMayer2000),suggesting
thatthepollinationrequirementsarenotclearlyunderstood.Therefore,inChapter1,I
designedcomplementaryfieldandgreenhouseexperimentstotestthecontributionofboth
bioticandnon‐bioticfactorstocranberrypollination.Iincludedtreatmentsforbiotic(i.e.
beeorbyhand),non‐biotic(i.e.agitation),andnopollination.Fromtheresultsofthese
2
3
studies,wewillgainabetterunderstandingofhowmuchpollinationcanbeattributedto
beesasopposedtonon‐bioticfactors.
ThemajorityofWisconsincranberrygrowersusehoneybeestomeettheir
pollinationrequirements,althoughtheoptimalhivestockingdensityisnotknown.While
therecommendedstockingdensityis2‐3hives/acre,actualmanagementrangesfrom0‐9
hives/acre.InChapter2,Iused11yearsofhistoricalyieldandhoneybeerecordsfrom
about40growerstodeterminetheeconomicallyoptimalstockingdensity.Theoptimal
stockingdensityisthepointatwhichanadditionalhiveresultsinnoadditionalincreasein
yield.Theresultsofthisresearchwillprovideanupdatedrecommendedhivestocking
densitythatcanbeusedbythegrowersfortheirpollinationmanagementpracticesinthe
future.
Anothersourceofbioticpollinationisnativebees.Inchapter3,Iinvestigatedthe
influenceoflocalfarmmanagement(i.e.,pesticideusage,plantuprightdensity)and
surroundinglandscapeonthesebees,aswellascranberryyield.Overthecourseofthree
fieldseasons,Ipantrappedbeesat49uniquesitesacrossalandscapegradient.In2011
and2012,Ialsocomparedyieldestimatesinopenplotsandpollinatorexclusioncagesto
determinewhetherlocalandlandscapefactorsinfluencethecontributionofbeestoyield.
Thischapterfurtheraddstotheliteratureontheeffectsoflocalandlandscapefactorson
nativepollinatorsandprovidesanewangleforlookingattheeffectofthesefactorsonthe
contributionofpollinatorstocropyield.
Whilewecanlearnalotaboutasystemthroughrigorousscientificstudies,
communicatingdirectlywiththepeoplewhomanagethesystemcanalsoprovidevaluable
3
4
insight.Ascientificstudytakesasnapshotofthesystem,butthegrowershavebeen
observingpatternsandactivityontheirfarmsforyearsandmayhaveamulti‐generational
perspectiveonmanagementdecisionsandoutcomes.Therefore,thefourthandfinal
chapterofmydissertationtakesasocialperspectiveoncranberrypollination,management
decisions,andon‐farmconservation.Thegoalofthischapterwastodeterminethe
barriersthatexistwhichpreventWisconsincranberrygrowersfromimplementingon‐
farmconservationprogramsfornativebees.Throughawrittensurveyofthegrowers,I
askedabouttheircurrentmanagementpractices,awarenessofnativepollinatorsand
pollinatorhabitat,andtheirattitudestowardandparticipationinon‐farmconservation
programs.Theresultsofthischapterwillprovideconcreterecommendationsforagency
personnelinordertotargetoutreachactivitiesandadjustprogramrequirementsto
increaseparticipationinUSDA–sponsoredon‐farmconservationprogramsfornativebees.
Fromthisresearch,wewillgainanunderstandingofthefactorsthatcontributeto
cranberrypollination,thevalueandoptimalstockingdensityofhoneybeesforcranberry
production,theinfluenceoflocalandlandscapefactorsonnativebeesandtheir
contributiontocranberryyield,andagrowerperspectiveonpollinationandon‐farm
conservation.Thisdissertationprovidestheresultsofrigorousscientificstudiesalong
withguidelinesforthepracticalapplicationoftheseresults.Itismyhopethattheresults
ofthisresearchwillbeusedbythescientific,grower,andagencycommunitiestoshape
furtherresearch,refinepollinationmanagementpractices,andimproveconservation
programstoincreasetheadoptionofon‐farmconservationfornativebeesinWisconsin
cranberry.
4
5
Literaturecited
Colquhoun,J.,andH.Johnson.2010.Sustainablecranberryproductionforavibrantfuture:TheWisconsinexperience.UniversityofWisconsin‐Extension.
Delaplane,K.S.,andD.F.Mayer.2000.Croppollinationbybees.CABIPublishing,New
York,NY.Eilers,E.J.,C.Kremen,S.SmithGreenleaf,A.K.Garber,andA.‐M.Klein.2011.Contribution
ofpollinator‐mediatedcropstonutrientsinthehumanfoodsupply.PLoSONE6:e21363.
Ellis,J.D.,J.D.Evans,andJ.Pettis.2010.Colonylosses,managedcolonypopulationdecline,
andColonyCollapseDisorderintheUnitedStates.JournalofApiculturalResearch49:134–136.
Fahrig,L.2003.Effectsofhabitatfragmentationonbiodiversity.AnnualReviewofEcology,
Evolution,andSystematics34:487–515.Foley,J.A.,R.DeFries,G.P.Asner,C.Barford,G.Bonan,S.R.Carpenter,F.S.Chapin,M.T.
Coe,G.C.Daily,H.K.Gibbs,andothers.2005.Globalconsequencesoflanduse.Science309:570.
Greenleaf,S.S.,N.M.Williams,R.Winfree,andC.Kremen.2007.Beeforagingrangesand
theirrelationshiptobodysize.Oecologia153:589–596.Klein,A.‐M.,B.E.Vaissière,J.H.Cane,I.Steffan‐Dewenter,S.A.Cunningham,C.Kremen,and
T.Tscharntke.2007.Importanceofpollinatorsinchanginglandscapesforworldcrops.ProceedingsoftheRoyalSocietyB:BiologicalSciences274:303–313.
Losey,J.E.,andM.Vaughan.2006.Theeconomicvalueofecologicalservicesprovidedby
insects.Bioscience56:311–323.Luck,G.W.,G.C.Daily,andP.R.Ehrlich.2003.Populationdiversityandecosystemservices.
TrendsinEcology&Evolution18:331–336.Potts,S.G.,J.C.Biesmeijer,C.Kremen,P.Neumann,O.Schweiger,andW.E.Kunin.2010.
Globalpollinatordeclines:trends,impactsanddrivers.TrendsinEcology&Evolution25:345–353.
Vitousek,P.M.,H.A.Mooney,J.Lubchenco,andJ.M.Melillo.1997.Humandominationof
Earth’secosystems.Science277:494.
5
6
Winfree,R.,N.M.Williams,H.Gaines,J.S.Ascher,andC.Kremen.2008.Wildbeepollinatorsprovidethemajorityofcropvisitationacrossland‐usegradientsinNewJerseyandPennsylvania,USA.JournalofAppliedEcology45:793–802.
Zurbuchen,A.,L.Landert,J.Klaiber,A.Müller,S.Hein,andS.Dorn.2010.Maximum
foragingrangesinsolitarybees:onlyfewindividualshavethecapabilitytocoverlongforagingdistances.BiologicalConservation143:669–676.
6
1
CHAPTER1
Bioticandnon‐bioticfactorscontributetocranberrypollination
Author:HannahR.GainesDay
Co‐author(s):ClaudioGratton
Forsubmissionto:JournalofPollinationEcologyorHortScience
Abstract:Asmanagedandnativebeepopulationscontinuetodecline,farmersface
possiblecropfailuresduetoinsufficientpollination.Crops,however,varyinthedegreeto
whichtheydependonpollinatorstomovepollenbetweenflowers,suggestingthatsome
cropsmaynotbeassensitivetovariationinpollinatoravailabilityorabundanceasothers.
Cranberry(Vacciniummacrocarpon)isgenerallyconsideredapollinator‐dependentcrop,
althoughestimatesofyieldlossintheabsenceofbeesvarygreatlyfrom30‐100%,raising
thepossibilitythatnon‐bioticfactorsmayalsobeimportantforcranberrypollination.The
objectiveofthisstudywastodeterminethecontributionofbioticandnon‐bioticfactorsto
cranberrypollination.Weperformedcomplementaryfieldandgreenhouseexperimentsto
comparetheeffectofbiotic(i.e.,beeorhandpollination)andnon‐bioticfactorssuchas
windormanualagitationonyield.Inthegreenhouse,wefoundthatplantsinthebiotic
treatmentproducedmoreberriesperupright,moreseedsperberry,andagreateroverall
yieldthanthenon‐biotictreatment.Plantsinthenon‐bioticagitationtreatment,however,
alsohadmoreberriesperuprightandoverallyieldthanplantsintheundisturbedcontrol
7
2
treatment.Inthefield,plantsinthebiotictreatmentproducedheavierberries,moreseeds
perberry,andagreateroverallyieldthanthenon‐bioticwindandagitationtreatments.
Plantsinthenon‐biotictreatmentsalsohadahigheryieldthantheundisturbedcontrol
treatment.Inboththegreenhouseandthefield,wefoundastrongcorrelationbetween
berryweightandnumberofseedsperberryfortheStevenscultivar.Theseresults
demonstratethatbothbioticandnon‐bioticfactorscontributesignificantlytocranberry
pollinationandthatcranberryisnotasdependentonpollinatorsaspreviouslybelieved.
Keywords(6):cranberry,Vacciniummacrocarpon,pollination,windpollination,thrips
1.0Introduction
Insectpollinationisanimportantecosystemservicerequiredbythemajorityof
floweringplants(Ollertonetal.2011).Beesarethemostimportantpollinators,butare
decliningworldwide(Biesmeijeretal.2006;NationalResearchCouncil2007;Pottsetal.
2010).InNorthAmerica,manyfarmersrelyonthemanaged,non‐nativehoneybee(Apis
mellifera)toprovidepollinationservices.Inthepast50years,however,managedhoney
beecolonieshavedeclinedover60%duetomites,disease,pesticideexposure,poor
nutrition,andstress(Ellisetal.2010).Nativebeesalsoprovideimportantpollination
servicesand,insomesystems,areabletoprovidesufficientpollinationwithoutthe
assistanceofhoneybees(e.g.,cranberry,Mohr&Kevan1987;watermelon,Winfreeetal.
2008).Unfortunately,nativebeesarealsoexperiencingdrasticdeclinesduetohabitat
destruction,intensifiedagriculture,agri‐chemicalexposure,anddisease(reviewedbyPotts
8
3
etal.2010).Withthedeclineofbees,farmersfacepossiblecropfailuresduetoinsufficient
pollination(Gallaietal.2009;Calderone2012).
Crops,however,varyinthedegreetowhichtheydependonpollinators,suggesting
thatnotallcropsareequallysusceptibletovariabilityintheabundanceofpollinators.
Whiletwothirdsofcropplantsrequireorbenefitfrominsectpollinationtoproduce
economicallyviablefruit(Kleinetal.2007),yieldresponsevariesbycrop.Intheabsence
ofpollinators,someplantsproducenofruit(e.g.,watermelon,Delaplane&Mayer2000),
whileothersproducemisshapen(e.g.,strawberries,Free1968;Jaycox1970)orsmallfruit
(e.g.,cherrytomato,Greenleaf&Kremen2006).Understandingthepollination
requirementsofspecificcropscanprovideinsightontheimpactofpollinatordeclineon
cropproduction.
Cranberry(Vacciniummacrocarpon)isconsideredtobeapollinatordependentcrop
(reviewedbyMcGregor1976;Eck1986,1990;Free1993;Delaplane&Mayer2000).
Althoughtheflowersareself‐compatible(Reader1977;Danaetal.1989;Sarracino&Vorsa
1991),pollenisreleasedbeforethestigmaisreceptive(Rigby&Dana1972),makingself‐
pollinationunlikely.Previousresearchshowsthatbeesareeffectivepollinatorsof
cranberry,capableofdepositingenoughpollenforfruitsetinoneortwovisits(Cane&
Schiffhauer2003).Asaresult,individualgrowersspendthousandsofdollarseachyearon
honeybeerentalstoensuresufficientpollination(USDANASSWASS2006).
Despitethewidespreaduseofhoneybeesandoveronehundredyearsofresearch,
thedegreetowhichcranberrydependsonpollinatorsremainsunclear.Althoughbeesare
efficientcranberrypollinatorsandcranberrydoesnotautogamouslyself‐pollinate(Cane&
9
4
Schiffhauer2003),itispossiblethatother,non‐bioticfactorscontributetocranberry
pollination.Fewstudies,however,haveconsiderednon‐beefactors.Ofthosethatdo,the
resultsareconflicting.Previousstudieshaveestimatedthataslittleas30%toasmuchas
100%ofthecranberrycropwouldbelostintheabsenceofbees(Robinsonetal.1989;
Southwick&Southwick1992;Williams1994;reviewedbyDelaplane&Mayer2000).At
oneendofthisrange,pollinatordeclineswouldhavelittleeffectoncranberryyield,while
attheotherendoftherange,therewouldbecropfailure.Furthermore,despiteclaimsthat
cranberrypollenistooheavytobetransferredbywind(Stricker1947),Papkeetal.(1980)
demonstratedthatthereisenoughcranberrypollenbeingcarriedinthewindandfalling
oncranberrymarshestopotentiallyprovideasignificantcontributiontopollination.And
Roberts&Struckmeyer(1942)demonstratedthatmanuallyagitatedcranberryplantsset
fruitintheabsenceofbees.Inlaterstudies,however,Filmer(1949)andMacKenzie(1994)
concludedthatwindorthemanualagitationofplantsdonotcontributetocranberry
pollination.Thepaucityofresearchinvestigatingnon‐beemechanismsofpollination,and
theinconsistencyinfindingsbetweenthestudiesthatdo,suggestthatwelackaclear
understandingofthefactorscontributingtocranberrypollination.
Thegoalofthisstudywastounderstandtherelativeimportanceofbioticandnon‐
bioticfactorstocranberrypollination.Wehypothesizedthatnon‐bioticfactorswould
resultinasufficientamountofpollentransfertoproduceaberrybutnotasmuchas
pollinationbybioticfactors.Toaddressthishypothesis,weestablishedfieldand
greenhouseexperimentsinwhichwemanipulatedbiotic(i.e.bees)andnon‐bioticfactors
(i.e.wind,manualagitation)thatmaycontributetocranberrypollination.Inthisstudywe
10
5
useyieldasametricofpollinationsuccesssincethisistheuniversalmetricusedin
agriculture.Weacknowledge,however,thatpollinationisonlyoneofmanyfactorsthat
mayaffectfinalyield.Bycomparingtheeffectofthesetreatmentsoncranberryyield,we
quantifiedtheindividualcontributionofeachfactortocranberryyield.
2.0Methods
2.1Cranberry
Cranberry(Vacciniummacrocarpon,Ericaceae)isaperennialfruitcropnativeto
NorthAmerica.Commerciallyitisgrowninartificiallycreatedmarsheswithsandy,acidic
soil.Cranberrygrowsinavininghabitalongthegroundandsendsupvegetativeand
floweringshoots(“uprights”).Eachfloweringuprightproducesupto8flowersthatbloom
sequentiallyfromthebottomoftheuprightupwardsoverseveralweeks(Eck1990).
Althoughcranberryflowerscontainbothmaleandfemaleparts,pollenisreleasedbefore
thestigmabecomesreceptive,makingautogamousself‐pollinationunlikely(Rigby&Dana
1972).Honeybeesarecommonlyusedforpollinationoncommercialcranberrymarshes
(Delaplane&Mayer2000).
2.2Greenhouse
Sixtydormantcranberryplantswithvisiblebuds,thirtyeachofthecultivars
“HyRed”and“Stevens”,weredugupfromacommercialcranberrymarshinlateMarchof
2012.Plantswererinsedthoroughlytoremoveallsandandpossiblepestsfromtheroots
andplantedinto15cmpotsofmoistpeatmoss.Potswerearrangedrandomlyina220C
11
6
greenhousewitha16hourphotoperiod.Approximatelyonemonthafterpotting,uprights
werethinnedsothatonlythosewith3‐5flowerbudsremained.Potswerefurtherthinned
to5‐6floweringuprightsperpotinordertoreducethetotalnumberofflowersinapot
andtoseparatethosethatremainedtomakehandpollinationpossible.Allbutthefirst
fourflowerstobloomperuprightweretrimmedoffastheyopenedsincecranberryplants
aremorelikelytosetfruitonthelower,earlierflowersthantheupperflowers(Birrenkott
&Stang1989).
Threetreatmentswereestablishedinthegreenhouse:(1)“hand”pollination,which
wasmeanttorepresentthebioticmovementofpollenbetweenflowers(mimickingbee
visitation),(2)“agitation”,whichwasmeanttorepresentthephysicalmovementofplants
bywind,and(3)anunmanipulatedcontrol,whichprovidedameasureofautogamousself‐
pollination.Toassessthepotentialforbioticpollination,flowerswithareceptivestigma
(i.e.thosethatweremoistandprotrudingfromthestamens)werehandpollinateddaily
duringbloombygentlyapplyingthestigmatoasmallaccumulationofpollencollected
fromyoungerflowersintothecapofamicro‐centrifugetube.Toassessthepotentialfor
non‐bioticpollination,plantsinthe“agitation”treatmentweregentlyjostleddailyduring
bloombymovingthepalmofthehandacrossthetopoftheuprights,causingtheplantsto
bumpagainsteachother.Thissimulatedthephysicalmovementofplantsasmaybe
causedbywindwhileexcludingthepossibilityofwindpollinationperseinwhichpollenis
movedthroughtheair.Plantsintheunmanipulatedcontroltreatmentwereleft
undisturbedthroughoutthestudytoassesswhetherfruitwouldbeproducedinthe
absenceofeitherbioticornon‐bioticfactors(i.e.,autogamousself‐pollination).Foreach
12
7
treatmentwehad10replicates(i.e.,pots)percultivar.Potswereplacedintwoparallel
rows~0.2mapartandspanning3‐3.5monbothsidesofasingleaisleofgreenhousetables
(onecultivarperside).Potsintherowclosesttotheaislewereassignedtothehand
pollinationtreatmentsothatallflowerswereeasilyaccessiblewithoutthepossibilityof
bumpingotherpots,andpotsinthebackrowwererandomlyassignedeithertothenon‐
bioticorundisturbedtreatment.Experimentaltreatmentswereinitiatedassoonasbloom
beganandwerecontinueddailyuntilallflowersweredoneblooming.Berrieswere
harvestedapproximatelytwomonthsafterbloombegan.Eachberrywasthenweighed
(wetweight,g)andthenumberofberriesperuprightwerecountedasaproxyforyield.In
ordertounderstandthelevelofpollinationreceivedineachtreatment,thenumberoffully
formedseedswascountedforeachberry.Cane&Schiffhauer(2003)demonstratedthat
thenumberofseedsisproportionaltotheamountofpollendepositedonthestigmaand
thereforerepresentsanindicationofpollinationsuccess.Averagesweretakenforeachpot
foratotalof10replicatespertreatment.
Toexaminedifferencesamongexperimentaltreatments(“hand”,“agitation”,
undisturbedcontrol)andcultivar(“Stevens”,“HyRed”),weusedafullyfactorialtwo‐way
analysisofvariance(ANOVA)tocompareberriesperupright,weightperberry,yield(total
berryweightperupright),andseedsperberry.Eachofthesevariablesindicatesomeform
ofpollinationsuccess.Differencesamongtreatmentsbycultivarweredeterminedusing
Tukey’sHonestlySignificantDifference(HSD)test(Hsu1996).Statisticalanalyseswere
performedwithJMPPro10.0.0(SASInstituteInc2007).
13
8
2.3Fieldcages
Toassesstheinfluenceofbioticandnon‐bioticfactorsoncranberryyieldinafield
setting,weestablishedacagestudyinacommercialcranberrymarshofthe“Stevens”
cultivar.Fourtreatmentswereestablished:(1)“open”,whichallowedbothbeevisitation
andagitationbywind,(2)“wind”,whichblockedinsectvisitationwithafinenylonmesh
(mattenylontulleinivory,JoannFabric)butallowedwindtoagitatetheplantsormove
pollen,(3)“closed”,whichpreventedinsectvisitationandwindusingfloatingrowcover
(Agribon+AG‐15InsectBarrier,Johnny’sSelectedSeeds)withacorrugatedplasticwind
blocksurroundingthecage(4mmcorrugatedplastic,LairdPlastics),and(4)“agitation”,
whichusedthesamecagedesignasthe“closed”treatmentbutreceivedmanualagitation
twiceperweekduringbloomtomimicmovementoftheplantsbywindbutexcludedthe
possibilityofwindpollinationperseinwhichpollenismovedthroughtheair.Each
treatmentwasreplicated10times.Cageframeswereconstructedusinglightweight,black
irrigationtubingand¾”PVCcutinto6”segmentswithapointatoneend(i.e.12”PVC
sectionswerecutinhalfata45oangle).ThePVCstakeswerehammeredintothegroundat
eachcornerofeachcageandtheirrigationtubingwasinsertedintothePVCtomakeatent
shapeframe.Meshwasattachedtothecageframesusing2”binderclips.Asmallsheetof
2mmplasticsheetingwasplacedbetweenthebinderclipandthemeshtopreventthe
meshfromtearing.Themeshwasfurthersecuredbyplacing½”‐¾”PCVsegmentsfilled
withgravelandsandatthebaseofeachsideofthecage.Inordertoaccountforpossible
differencesinplantsasafunctionoflocationwithinthecranberrybed,cageswere
arrangedinagridandtreatmentswereassignedusingamodifiedLatinSquaredesignwith
14
9
eachtreatmentoccurringoncepercolumnandtwiceperrow.Cagesweresetupbefore
cranberrybloom(May25,2012)andremovedwhenbloomwascomplete(July9,2012).
Berrieswereharvestedfroma0.3mx0.3m(1ft2)plotfromthecenterofeachcagebefore
harvest(September20,2012)andcountedandweighed(wet)toestimateyield.Wet
weightwasusedasthiscanbeeasilyconvertedtoyieldunitsusedbycranberrygrowers
(barrels/acre,1barrel=100lbs=45kg).Twentyberriesfromeachcagewerealsocutin
halfwitharazorbladeandallfullyformedseedswithinwerecounted.
Anadditional15cages(5each“open”,“wind”,and“closed”,treatmentsasdescribed
above)wereestablishedtomeasureenvironmentalvariablesthatmaybealteredbythe
cagesincludingtemperature,lightintensity,soilmoisture,andinsectcommunity.
Temperatureandlightintensityweremeasuredevery30minutesforthedurationofthe
cagestudyusingHOBOdataloggers(OnsetComputerCorporation)hunginsideStyrofoam
sunshields.SoilmoisturewasmeasuredusingaTDR300soilmoisturemeter(Spectrum
Technologies,Aurora,Illinois)twiceduringthegrowingseason(June14andJuly9,2012).
Fourmeasurementsweretakenpercage.Theinsectcommunitywasmeasured
continuouslyduringbloomusingyellowstickystrips(3”x5”,GreatLakesIPM025‐SS‐35)
andblue,yellow,andwhitepantraps(ACEBrandFluorescentpaint)containingsoapy
water(Dawnbluedishsoap).
Toexaminedifferencesamongtreatmentsforberriesperupright,weightperberry,
yield,andseedsperberry,weusedaone‐waymixedmodelANOVAwithrowandcolumn
locationsasrandomeffects.DifferencesamongtreatmentsweredeterminedusingTukey’s
15
10
HSDtest.StatisticalanalyseswereperformedusingJMPPro10.0.0(SASInstituteInc.
2007).
3.0Results
3.1Greenhouse
Inthegreenhouse,significantfruitsetoccurredonplantsinboththebiotic(“hand”)
andnon‐biotic(“agitation”)treatmentsbutnottheundisturbedcontrol.Ninety‐eight
percentof“HyRed”and92%of“Stevens”uprightsinthe“hand”pollinationtreatment
producedfruitand52%ofthe“HyRed”and30%ofthe“Stevens”uprightsinthe“agitation”
treatmentproducedfruit.Only2%of“HyRed”and5%of“Stevens”uprightsinthe
unmanipulatedcontrolproducedfruit,providingevidenceforthelackofautogamousself‐
pollinationincranberry.
Thenumberofberriesperuprightvariedsignificantlyamongtreatmentsinboth
cultivars(treatment,F2,58=248.6,p<0.0001,Fig.1a).Significantlymoreberriesperupright
wereproducedinthe“hand”pollinationtreatmentthaninthe“agitation”treatment,and
moreininthe“agitation”treatmentthanintheunmanipulatedcontrol.Thisdifference
wasgreaterinthe“HyRed”cultivarthanin“Stevens”(treatmentxcultivar,F2,58=3.9,
p=0.028).Berryweightdidnotvaryacrosstreatments(F2,41=2.4,p=0.11,Fig.1b)or
cultivar(F1,41=0.0021,p=0.96).Thenumberofseedsperberrywasgreaterinthe“hand”
pollinationtreatmentthanbothoftheothertreatmentsanddidnotvarybetweenthe
“agitation”treatmentandtheunmanipulatedcontrol(F2,58=136.3,p<0.0001,Fig.1c).The
numberofseedsperberryfortheundisturbedtreatmentin“HyRed”isbasedonasingle
16
11
berrythatformedinthistreatmentandisthereforenotrepresentativeofthetreatmentas
awhole.Yieldwasestimatedbytotalberryweightperuprightwhichisafunctionofberry
weightandberriesperupright.Overall,yieldswerehigherinthe“hand”pollination
treatmentthanintheothertreatments(treatment,F2,58=210.3,p<0.0001,Fig.1d).Total
berryweightperuprightwasalsosignificantlygreaterinthe“agitation”treatmentthanthe
“no‐pollination”treatmentfor“HyRed”butnotfor“Stevens”.
3.2Fieldexperiment
Inthefield,openplotshadthehighestfruitset,althoughevenwheninsectswere
excluded,therewassubstantialfruitsetinbothofthenon‐biotictreatments(“agitation”
and“wind”).Thenumberofberriesperuprightwasfairlyconsistentacrosstreatments
anddidnotvarybetweenthe“open”treatmentandanyothertreatmentbutwaslowerin
the“agitation”treatmentthaninthe“wind”or“closed”treatments(treatment,F3,24.6=4.6,
p=0.011,Fig.2a).Theweightperberrywassignificantlydifferentamongtreatments(F‐
3,24.2=36.6,p<0.0001,fig.2b)withtheheaviestberriesinthe“open”treatment,followedby
“agitation”,andlowestinthe“wind”treatmentand“closed”control.Thenumberofviable
seedsperberryvariedsignificantlybytreatment(F3,23.91=64.9,p<0.0001,Fig.2c)withthe
mostfoundinthe“open”treatmentfollowedby“agitation”andtheleastin“wind”and
“closed”treatment.Yield,asestimatedfromsmallplots,washighestforthe“open”
treatmentfollowedby“wind”and“agitation”andlowestforthe“closed”treatment
(F3,23.1=70.5,p<0.0001,fig.2d).
Becausethenumberofseedsisproportionaltotheamountofpollendeposited
(Cane&Schiffhauer2003),weusedseedsperberryasameasureofpollinationsuccess.
17
12
Wefoundastrongrelationshipbetweenberryweightandseednumberfor“Stevens”in
boththegreenhouse(p=0.0011,R2=0.42,fig.3a)andthefield(R2=0.83,p<0.0001,fig.3a)
butnotfor“HyRed”(p=0.39,R2=0.04,fig.3b).Thispatternsuggeststhatthenumberof
pollentetradsdepositedonflowersvariedsignificantlyamongtreatmentsandthatevenin
the“open”treatment,pollencanbelimited,preventingtheberryfromreachinga
maximumweight.
3.2.1Environmentaleffects
Theenvironmentalvariableswemeasuredtotestcageeffectsprovideevidencethat
thecagessuccessfullyexcludedpollinatorswhilemaintainingcomparableconditions
withineachcagetype.Theabundanceofinsectsvariedsignificantlyamongalltreatments
withthelowestabundanceinthe“closed”(rowcover)treatment,suggestingthatcage
treatmentseffectivelyexcludedallbutthesmallestinsects(e.g.,thrips).Soilmoisturedid
notvaryamongcagetreatments(F2,23=0.17,p=0.84).Thetemperaturewasonaverage
1.1oChigherinthe“closed”treatmentcagethantheopenor“wind”(nylonmesh)
treatments(F2,7038=13.5,p<0.0001).Lightvariedsignificantlyaswellwiththemostlightin
theopentreatmentfollowedequallybythe“wind”and“closed”treatments(F2,13163=42.4,
p<0.0001).Shadingcanresultinahigherproductionofvegetativematerialresultingin
loweredyields(RobertsandStruckmeyer1942).
4.0Discussion
Therequirementofbeesforcranberrypollinationisgenerallyassumedalthough
theevidenceislacking.Inoveronehundredyearsofresearch,onlyfourstudieshave
18
13
consideredthecontributionofnon‐beefactorstocranberrypollination(Roberts&
Struckmeyer1942;Filmer1949;Papkeetal.1980;MacKenzie1994).Inthisstudywe
demonstratethatintheabsenceofbees,cranberryisstillabletoproducefruit.Inafield
setting,wefoundthatplants,fromwhichinsectpollinators,butnotwind,wereexcluded,
producedagreateroverallyieldthanplantsfromwhichbothinsectsandwindwere
excluded.Thisfinding,combinedwithmanualagitationtreatmentsinthefieldand
greenhouse,challengesthenotionthatbeesaretheonlywayforcranberrytoproducefruit
(reviewedbyMcGregor1976;Eck1990;Free1993;DelaplaneandMayer2000),and
suggeststhatbothbioticandnon‐bioticfactorscontributesignificantlytocranberry
pollination.
Previousresearchoncranberrypollinationhasmainlycomparedtheeffectofbee
pollinationversusautogamousself‐pollination.Themajorityofthesestudieshave
concludedthatbeesarerequiredtoproducefruit,althoughwhattheirdataactually
demonstrateisthatcranberriesarenotautogamousandareunabletoself‐pollinatewithin
asingleflower(Hutson1925;Farrar&Bain1947;Filmer&Doehlert1959;Marucci1966;
Marucci&Moulter1977;Cane&Schiffhauer2003).Ourstudysupportsthefindingthat
cranberrydoesnotautogamouslyself‐pollinate,butalsoprovidesevidencethatother,non‐
beefactorsplayaroleincranberrypollination.
Theincreasedyieldinthenon‐biotictreatments(i.e.agitation,wind)ascompared
totheclosedtreatmentinthefieldmaybeattributabletouncontrolledvariablesinthefield
suchasnon‐beeinsectpollinators,agronomicvariation(i.e.uprightdensity),ora
combinationofthetwo.Inthegreenhouse,theagitationtreatmentproduced
19
14
approximately25%oftheyieldcomparedtothehandpollinationtreatment,whereasthe
no‐pollinationtreatmenthadvirtuallyzeroyield.Inthefield,thenon‐biotictreatments
producedapproximately50%oftheyieldcomparedtotheopentreatment,andeventhe
no‐pollinationtreatmenthadsomeyield.Throughcarefulcageconstructionandpan
trappingwithincages,weareconfidentthatbeeswerenotcontributingtotheincreased
yieldwithinthefieldcagetreatments.However,thrips,tinypollen‐eatinginsects,were
foundinalltreatmentsinthefieldandhavebeenshowntobeeffectivepollinatorsinother
systems,includingEricaceousplantswhichareinthesamefamilyascranberry(Hagerup&
Hagerup1953;Kirk1988;Baker&Cruden1991;Ananthakrishnan1993),althoughtheir
influenceincranberrypollinationhasnotbeeninvestigated(butseeAppendix1.1).The
absenceofthripsinthegreenhouse(pers.obs.)andabundanceinthefieldsuggeststhat
theymaybeacontributingfactor.
Anotherfactorthatmayhavecausedincreasedyieldinthenon‐bioticandclosed
treatmentsinthefieldisfloweringuprightdensity.Inthegreenhouse,floweringuprights
werethinnedtoaverylowdensitytomakehandpollinationmanageable(279flowering
uprights/m2),whereasinthefield,floweringuprightsweresixtimesasdense(1705
floweringuprights/m2).Withahigherdensityoffloweringuprights,pollenismore
abundantandflowersareclosertogether,increasingthelikelihoodthatpollencouldbe
transferredbetweenflowerswithevenslightagitation.Ahighdensityofuprightsnotonly
increasestheprobabilityofpollentransferbetweenflowerswhenagitated,butmayalso
makemovementofthripsamongflowerseasier.
20
15
Anotherdifferencebetweenthefieldandgreenhousewasthepatternamong
treatmentsinberryweight.Inthefield,berryweightwassignificantlydifferentamong
treatments,whereasinthegreenhouse,berryweightdidnotvaryamongtreatments.
Previousstudieshavefoundastrongcorrelationbetweenberryweightandseednumber
(Hall&Aalders1965;Rigby&Dana1971),demonstratingthatwhenpollenislimited,
berryweightislow.Inourstudy,wefoundasignificantrelationshipbetweenberryweight
andseednumberintheStevenscultivarinthefieldandthegreenhouse,althoughthe
relationshipwasweakerandshallowerinthegreenhouse.Despiteevidenceinprevious
studiesthatcranberrycanproduceseedlessfruitthroughparthenogenesis(Roberts&
Struckmeyer1942),evenberriesintheno‐pollinationtreatmenthadsomeviableseeds,
suggestingthatparthenogenesisisnotoccurringhere.Additionally,thenumberofseeds
perberrywassimilarbetweentreatmentsinthefieldandgreenhouse,suggestingthat
otherfactors,suchaswaterandnutrientavailabilityandnotjustpollenlimitation,may
playaroleinberryweight.Thelackofrelationshipbetweenberryweightandseed
numberintheHyRedcultivarinthegreenhousefurthersuggeststhatberryweightis
influencedbyfactorsotherthanseednumber.CaneandSchiffhauer(2003)foundthatin
theStevenscultivar,theberryreachesamaximumweightwhentheflowerreceives8
tetradsofpollenresultinginabout15seeds.Inourstudy,berriesintheStevenscultivar
onlyreached15seedsperberryinthehandpollinatedtreatmentinthegreenhouse,
suggestingthatevenintheopentreatmentinthefield,pollenmaybelimited.InHyRed,
however,therelationshipwasflatsuggestingthattheseednumber/berryweight
relationshipmayvaryamongcultivars.
21
16
Previousstudies,onwhichcurrentpollinationmanagementrecommendationsare
derived(McGregor1976),weredesignedtotesttheimportanceofbeestocranberry
pollinationbutdidnotaccountforthepossibilityofnon‐beepollinationvectors.Other,
morerecentstudies(e.g.,CaneandSchiffhauer2003)havedemonstratedtheefficiencyof
beepollinatorsandthelackofautogamousself‐pollinationincranberry,butstilloverlook
alternativemechanismsforpollination.Phillips(2011)alsodemonstratedthelackof
autogamousself‐pollinationincranberrybutincorrectlyconcludedthatcranberryrequires
insectpollination.Inthesestudies,individualcranberryflowerswereisolatedto
demonstratetheinabilitytoself‐pollinate.However,inthefield,flowersarenotisolated,
rathertheyareexposedtoavarietyofbioticandnon‐bioticfactors.Inseveralinfluential
studies,Hutson(1925)andFilmerandDoehlert(1955)demonstratedthedependenceof
cranberryonbeesinafieldcagestudy,butwhenberriesformedwithinthecages,
attributedtheberriestofaultycagesratherthanconsideringalternativemechanismsof
pollination.
Ourstudywasspecificallydesignedtotesttheco‐existenceofpollination
mechanismsincludingbees,wind,mechanicalaction,andautogamousself‐pollination.We
foundthatmechanicalagitationbywindorhand,caneffectivelymovepollenbetween
cranberryflowers.Insmallplotsatourstudysite,asmuchas50%ofyieldcouldbe
attributedtonon‐beepollinators,resultingineconomicallycompetitiveyields.Theresults
ofthisstudysuggestthatintheabsenceofbeepollinators,cranberrygrowersmaystillbe
abletoproduceasignificantcropyield.Theseresultsfurthersuggestthatmechanical
22
17
agitationmaybemoreprevalentinothercroppingsystemsthanpreviouslybelievedand
providesfurtherrationalefordetailedpollinationstudiesinothercropplants.
Acknowledgements
TheauthorswouldliketothankC.J.SearlesCranberryInc.forallowingustoconductour
fieldcagestudyontheirmarshandGaynorCranberryforprovidingplantmaterialforthe
greenhousestudy.WethankEricZeldinforhisassistancediggingplantmaterialand
providingguidanceonestablishingtheplantsinthegreenhouse.Wealsothankthefield
andlabassistantswhoassistedwiththisproject:RachelMallinger,AmandaRudie,Collin
Schwantes,EricWiesman,ChristopherWatson,SachaHorn,andScottLee.Fundingforthis
researchwasprovidedbyaHatchGranttoC.G.
Literaturecited
AnanthakrishnanTN(1993)Theroleofthripsinpollination.CurrentScience65:262–264.
BakerJD,CrudenRW(1991)Thrips‐mediatedself‐pollinationoftwofacultativelyxenogamouswetlandspecies.AmericanJournalofBotany78:959–963.
BiesmeijerJC,RobertsSPM,ReemerM,OhlemüllerR,EdwardsM,PeetersT,SchaffersAP,PottsSG,KleukersR,ThomasCD,SetteleJ,KuninWE(2006)Paralleldeclinesinpollinatorsandinsect‐pollinatedplantsinBritainandtheNetherlands.Science313:351–354.
BirrenkottBA,StangEJ(1989)Pollinationandpollentubegrowthinrelationtocranberryfruitdevelopment.JournaloftheAmericanSocietyforHorticulturalScience114:733–737.
CalderoneNW(2012)Insectpollinatedcrops,insectpollinatorsandUSagriculture:Trendanalysisofaggregatedatafortheperiod1992‐2009.PlosOne7:e37235.
CaneJH,SchiffhauerD(2003)Dose‐responserelationshipsbetweenpollinationandfruitingrefinepollinatorcomparisonsforcranberry(Vacciniummacrocarpon[Ericaceae]).AmericanJournalofBotany90:1425–1432.
23
18
DanaMN,SteinmannS,GobenL(1989)Pollensourceandfruitsetofcranberry.Cranberries53:10–14.
DelaplaneKS,MayerDF(2000)Croppollinationbybees.CABIPublishing,NewYork,NY.
EckP(1986)Cranberry.In:Monselise,S.P.(ed)CRChandbookoffruitsetanddevelopment.CRCPress,BocaRaton,FL,pp109–117.
EckP(1990)TheAmericancranberry.RutgersUniversityPressNewBrunswick,NJ.
EllisJD,EvansJD,PettisJ(2010)Colonylosses,managedcolonypopulationdecline,andColonyCollapseDisorderintheUnitedStates.JournalofApiculturalResearch49:134–136.
FarrarCL,BainHF(1947)Honeybeesaspollinatorsofthecranberry.Cranberries11
FilmerRS(1949)Cranberrypollinationstudies.In:AmericanCranberryGrowers’AssociationProceedingsoftheAnnualMeeting.pp14–20.
FilmerRS,DoehlertCA(1959)Useofhoneybeesincranberrybogs.NewJerseyAgriculturalExperimentStation,NewBrunswick,NJ.
FilmerRS,Doehlert.CA(1955)Useofhoneybeesincranberrybogs.GleaningsinBeeCulture:265–268.
FreeJB(1968)Thepollinationofstrawberriesbyhoney‐bees.JournalofHorticulturalScience48:107–11.
FreeJB(1993)Insectpollinationofcrops,2nded.AcademicPressInc.,London.
GallaiN,SallesJ‐M,SetteleJ,VaissièreBE(2009)Economicvaluationofthevulnerabilityofworldagricultureconfrontedwithpollinatordecline.EcologicalEconomics68:810–821.
GreenleafSS,KremenC(2006)WildbeespeciesincreasetomatoproductionandresponddifferentlytosurroundinglanduseinNorthernCalifornia.BiologicalConservation133:81–87.
HagerupE,HagerupO(1953)ThripspollinationofEricatetralix.NewPhytologist52:1–7.
HallIV,AaldersLE(1965)Therelationbetweenseednumberandberryweightinthecranberry.CanadianJournalofPlantScience45:292–292.
HsuJC(1996)Multiplecomparisons:theoryandmethods.Chapman&Hall,London.
HutsonR(1925)Thehoneybeeasanagentinthepollinationofpears,applesandcranberries.JournalofEconomicEntomology18:387–391.
JaycoxER(1970)Pollinationofstrawberries.AmericanBeeJournal:176–177.
24
19
KirkWDJ(1988)Thripsandpollinationbiology.In:Ananthakrishnan,T.N.,Raman,A.(eds)DynamicsofInsect‐PlantInteraction:RecentAdvancesandFutureTrends.Oxford&IBHPublishingCompany,NewDelhi
KleinA‐M,VaissiereBE,CaneJH,Steffan‐DewenterI,CunninghamSA,KremenC,TscharntkeT(2007)Importanceofpollinatorsinchanginglandscapesforworldcrops.ProceedingsoftheRoyalSocietyB‐BiologicalSciences274:303–313.
MacKenzieKE(1994)Theforagingbehaviourofhoneybees(ApismelliferaL)andbumblebees(Bombusspp)oncranberry(VacciniummacrocarponAit).Apidologie25:375–383.
MarucciPE(1966)CranberryPollination.Cranberries30:11–13.
MarucciPE,MoulterHJ(1977)CranberrypollinationinNewJersey.ActaHorticulturae(ISHS)61:217–222.
McGregorSE(1976)Insectpollinationofcultivatedcropplants.AgriculturalResearchService,USDepartmentofAgriculture,Washington,D.C.
MohrNA,KevanPG(1987)Pollinatorsandpollinationrequirementsoflowbushblueberry(VacciniumangustifoliumAit.andV.myrtilloidesMichx.)andcranberry(V.macrocarponAit.)inOntariowithnotesonhighbushblueberry(V.corymbosumL.)andlingonberry(V.vitis‐ideaeL.).In:ProceedingsoftheEntomologicalSocietyofOntario.pp149–154.
NationalResearchCouncil(2007)StatusofPollinatorsinNorthAmerica.CommitteeontheStatusofPollinatorsinNorthAmerica,TheNationalAcademiesPress,Washington,D.C.
OllertonJ,WinfreeR,TarrantS(2011)Howmanyfloweringplantsarepollinatedbyanimals?Oikos120:321–326.
PapkeAM,EatonGW,BowenPA(1980)Airbornepollenaboveacranberrybog.HortScience15:756.
PhillipsKN(2011)Acomparisonofbumblebees(Bombusspp.)andhoneybees(Apismellifera)forthepollinationofOregoncranberries(Ericaceae:Vacciniummacrocarpon).M.S.Thesis.OregonStateUniversity
PottsSG,BiesmeijerJC,KremenC,NeumannP,SchweigerO,KuninWE(2010)Globalpollinatordeclines:trends,impactsanddrivers.TrendsinEcology&Evolution25:345–353.
ReaderRJ(1977)Bogericadflowers:self‐compatibilityandrelativeattractivenesstobees.CanadianJournalofBotany55:2279–2287.
RigbyB,DanaMN(1971)Seednumberandberryvolumeincranberry.HortScience6:495–496.
RigbyB,DanaMN(1972)Floweropening,pollenshedding,stigmareceptivity,andpollentubegrowthinthecranberry.HortScience7:84–85.
25
20
RobertsRH,StruckmeyerBE(1942)Growthandfruitingofthecranberry.In:ProceedingsoftheAmericanSocietyforHorticulturalScience.pp373–379.
RobinsonWS,NowogrodzkiR,MorseRA(1989)ThevalueofhoneybeesaspollinatorsofU.S.crops.AmericanBeeJournal129:411–423,477–487.
SarracinoJM,VorsaN(1991)Selfandcrossfertilityincranberry.Euphytica58:129–136.
SASInstituteInc.(2007)JMPPro10.SASInstituteInc.,Cary,NC.
SouthwickEE,SouthwickL(1992)EstimatingtheEconomicValueofHoneyBees(Hymenoptera:Apidae)asAgriculturalPollinatorsintheUnitedStates.JournalofEconomicEntomology85:621–633.
StrickerMH(1947)Beesandpollinatingcranberries.AmericanCranberryGrowers’AssociationProceedingsoftheAnnualMeeting77
USDANASSWASS(2006)2005Cranberryproductionandpollinationsurvey.
WilliamsI(1994)ThedependenceofcropproductionwithintheEuropeanUniononpollinationbyhoneybees.AgriculturalZoologyReviews6:229–257.
WinfreeR,WilliamsNM,GainesH,AscherJS,KremenC(2008)Wildbeepollinatorsprovidethemajorityofcropvisitationacrossland‐usegradientsinNewJerseyandPennsylvania,USA.JournalofAppliedEcology45:793–802.
26
21
Figurecaptions
Figure1:Metricsofpollinationsuccessforthecranberrycultivar“HyRed”and“Stevens”as
measuredinthegreenhouse:(a)thenumberofberriesperupright(mean±SE),(b)the
berryweight(mean±SE),(c)thenumberofseedsperberry(mean±SE),and(d)theyield
estimate(meanwetweightperupright±SE).Treatmentsincluded“hand”pollination
whichrepresentedbioticpollination,“agitation”whichrepresentednon‐bioticpollination,
andanundisturbedcontrol.
Figure2:Metricsofpollinationsuccessforthecranberrycultivar“Stevens”asmeasuredin
thefield:(a)thenumberofberriesperupright(mean±SE),(b)theberryweight(mean
±SE),(c)thenumberofseedsperberry(mean±SE),and(d)theyieldestimate(mean±SE).
Treatmentsincluded“open”pollinationwhichrepresentedambientpollination,“wind”
whichrepresentednon‐bioticpollinationthroughwindpollinationperseandphysical
movementoftheplantsduetowind,“agitation”whichrepresentednon‐bioticpollination
duetophysicalmovementoftheplantsintheabsenceofwindpollinationperse,anda
closedcontrol.
Figure3:Relationshipbetweenberryweightandseedsperberryforthecranberry
cultivars(a)“Stevens”and(b)“HyRed”.
27
Figure 1
A.
C.
B.
D.
Ber
ries/
uprig
ht
A
C
D
B
CDD
agitationhandpollination
undisturbed
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
HyRed cultivarStevens cultivar
agitationhandpollination
undisturbedW
et w
eigh
t per
ber
ry (g
)
A A
A
AA
A1.50
1.25
1.00
0.75
0.50
0.25
0.00
A
B
Cagitationhand
pollinationundisturbed
A
BCC
15
10
5
0
20
Tota
l ber
ry w
et w
eigh
t/u
prig
ht (g
)
See
ds/b
erry
A
C
C
15
10
5
0
20
agitationhandpollination
undisturbed
B
CABC
28
A
CC
B
Ber
ries/
uprig
ht
A A
B
Yiel
d es
timat
e (b
arre
ls/a
cre)
A
B
C
B
Cage Treatment
agitationwindopen closed
0.0
0.5
1.0
1.5
2.0AB
2.5
See
ds p
er b
erry
A
C
C
B
Figure 2
Cage Treatment
agitationwindopen closed
agitationwindopen closed agitationwindopen closed
6
4
2
0
8
12
10
150100500
200
300250
350
400
A.
C.
B.
D.
1.50
1.00
0.50
0W
et w
eigh
t per
ber
ry (g
)
29
(a) Stevens cranberry cultivar
(b) HyRed cranberry cultivar
Figure 3
Greenhouse
Field
Greenhouse
Seeds/berry
Seeds/berry
Ber
ry w
et w
eigh
t (g)
Ber
ry w
et w
eigh
t (g)
2.0
0.0
0.5
1.0
1.5
200 5 10 15
2.0
0.0
0.5
1.0
1.5
200 5 10 15
Field Greenhouse
openagitation
closedwind
openagitationundisturbed
30
1
Appendix1.1:Assessingtheroleofthrips(Thysanoptera)incranberrypollination
Introduction
Cranberryisconsideredapollinatordependentcrop(DelaplaneandMayer2000),
butinfieldexperimentsin2011and2012,caged(“closed”)cranberryplantsproduceda
significantamountoffruit.Inafollow‐upgreenhousestudy,plantsinthe“closed”
treatmentproducednearlynofruitatall.Theseresultssuggestthatadditional,
uncontrolledfactorsexistinthefieldthateffectcranberrypollination.Onepossiblefactor
isthepresenceofthrips,tinypollen‐eatinginsects.Thripsareknowntobeimportant
pollinatorsinsomesystems(Kirk,W.D.J.1988,BakerandCruden1991,Ananthakrishnan
1993)andwerefoundonstickycardsinallfieldtreatmenttypes(Gaines,unpubl.data).
Anotherfactorthatmayinfluencepollinationisthedensityoffloweringuprights.Field
plotshadasignificantlyhigherdensityoffloweringuprightsascomparedtothe
greenhouse.Ahigherfloweringuprightdensitycouldmakemovementofthripsfromone
planttoanothereasierorsimplyincreasetheprobabilitythatflowersbumpintoeach
otherwhenagitated.Therefore,theobjectiveofthisstudywastodeterminethe
contributionofthripsandtheinfluenceofuprightdensityoncranberrypollination.
Methods
Cranberries
CranberryplantsofthecultivarStevensweredugfromacommercialcranberry
marshinlatefall2012.Sincethegroundinthemarshwasalreadyfrozen,plantswereonly
31
2
Table1:Numberofpotspertreatmentforintendedstudydesign
collectedattheedgeofthemarshwherechunksofplantscouldbeextractedwithanice
pick.Plantswerebroughtbacktothegreenhouseandkeptincoldstorage(4oCwithan8‐
hourphotoperiod)foronemonth.Icewasplacedovertheplantsonceperweektokeep
theplantsfromdryingout.
Afteronemonthincoldstorage,theplantswererinsedoffandtransplantedinto6”
potsfilledwithmoistpeat.Potsweresetina220Cgreenhousewitha16hourphotoperiod
andassignedatreatment(table1).Plantswerewateredevery4daysandexcessivevine
growthtrimmed.Afteraboutthreeweeks,plantswerefertilizedwith½teaspoon
fertilizerpergallonwater.
Thestudywasintendedtobea2‐
factor,fullfactorialdesigntocomparefruit
setduetohandpollination,thrips
pollination,agitation,andnopollinationas
afunctionofuprightdensity.Tenpots
wereassignedtoeachtreatment(table1).
Approximatelyfourweeksafter
potting,plantswereexpectedtobeginshowingsignsofflowerdevelopment.
Unfortunately,veryfewplantsdidsoandtheexperimentaldesignhadtobedrastically
reduced(table2).Thiswaslikelyduetopoorplantmaterialdugfromthemarshinthefall.
Becauseofthefrozenground,wewereunabletobeselectiveinwhichplantswedug.The
plantsthatdidshowsignsoffloweringwereplacedintolargecageswithfinemeshtokeep
Uprightdensity
Low High
Pollination
treatment
Hand 10 10
Thrips 10 10
Agitation 10 10
None 10 10
32
3
Table2:Numberofpotspertreatmentforactualstudydesign.
thripsinorout.Eachcagehadtwopotswithsixcagespertreatment(n=12potsper
treatment).
Thrips
Oncethecranberryplantswerere‐
pottedintopeat,athripscolonywas
establishedinthelab.Tostartthecolony,a
fewdozengreenbeanscontainingthrips
eggswereacquiredfromanexistingcolony.
Thecolonywasmaintainedusing
previouslyestablishedmethods(Tom
German,pers.comm.)asfollows.Twotothreeofthebeansfromtheestablishedcolony
wereplacedinplasticdelicupswithmeshtops.Freshbeanswerepurchasedandsoaked
for10minutesinableachsolution(50mLbleachto500mLwater)tokillany
microorganismsandpreventmoldingonthebeans.Oncedry,twotothreefreshbeans
wereaddedtoeachdelicuptoprovidefoodforemergingthrips.Freshbeanswereadded
andoldbeansremovedevery~4days.
Whencranberrybloombegan,plantsinthe“thrips”treatmentwereinoculatedwith
3lateinstaroradultthripsperflower.Thripsweretransportedtothegreenhouseusing
centrifugetubes.Tubeswerestuckintothesoilofthepotsinthethripstreatmentandthe
capsremovedtoallowthethripstocrawlontothecranberryplants.Thiswasdonetwice
duringthethreeweekbloom.Plantsinthenopollinationtreatmentwereleftundisturbed.
Uprightdensity
Low High
Pollination
treatment
Hand 0 0
Thrips 0 12
Agitation 0 0
None 0 12
33
4
Afterbloom,bluestickycardswereplacedineachcagetoverifythepresenceof
thrips(Chenetal.2009).Wealsointendedtocountandmeasureberriesoncetheywere
fullsize.Thisneverhappenedbecausenoberriesformedonanyplantsineithertreatment.
Results
Nocranberryfruitswereproducedineithertreatment(table3).Wewerealso
unabletorecaptureanythripsonthebluestickycardsineitherthecagesorintheopen
greenhouse.
Pots Berriesperupright
Thripsperstickycard
Pollination
treatment Thrips 12 0±0 0±0
None 12 0±0 0±0
Conclusions
Theresultsofthisstudyareunabletosupportthehypothesisthatthripscontribute
tocranberrypollination.Itispossiblethatthripsarenotcontributingtocranberry
pollinationorthatexperimentaldesignfailedtoeffectivelytestthis.Futurestudiesshould
digcranberryplantsearlierinthefallorinthespringtoallowforbetterplantmaterial
selection(i.e.,plantswithlargebudspresent).Additionally,insteadofinoculatingplants
withthrips,placingplantsingreenhouseswithknownthripsinfestationsmaybemore
Table3:Resultsfrompollinationstudy
34
5
effective.Thripsareaseriousgreenhousepestandfindingagreenhousewiththrips
problemshouldnotbedifficult.
Literaturecited
Ananthakrishnan,T.N.1993.Theroleofthripsinpollination.CurrentScience65:262–264.Baker,J.D.,andR.W.Cruden.1991.Thrips‐MediatedSelf‐PollinationofTwoFacultatively
XenogamousWetlandSpecies.AmericanJournalofBotany78:959–963.Chen,T.‐Y.,C.‐C.Chu,G.Fitzgerald,E.T.Natwick,andT.J.Henneberry.2009.Trap
EvaluationsforThrips(Thysanoptera:Thripidae)andHoverflies(Diptera:Syrphidae).EnvironmentalEntomology:1416–1420.
Delaplane,K.S.,andD.F.Mayer.2000.Croppollinationbybees.CABIPublishing,New
York,NY.Kirk,W.D.J.1988.Thripsandpollinationbiology.inAnanthakrishnan,T.N.andRaman,A.,
editors.DynamicsofInsect‐PlantInteraction:RecentAdvancesandFutureTrends.Oxford&IBHPublishingCompany,NewDelhi.
35
1
CHAPTER2
Thecontributionofhoneybeestocranberryyieldvarieswiththeamountof
woodlandinthesurroundinglandscape
Author:HannahR.Gaines
Co‐author(s):PaulD.MitchellandClaudioGratton
Forsubmissionto:JournalofEconomicEntomology(“apicultureandsocialinsects”or
“horticulturalentomology”section),EcologicalApplications,FrontiersinEcology,or
Ecosphere
Abstract:Two‐thirdsofcropplantsrequireorbenefitfrominsectpollination.Toensure
sufficientpollinationoftheircrops,manyfarmersrenthivesofhoneybees.Cranberry
(Vacciniummacrocarpon),aperennialfruitcropnativetoNorthAmerica,isgenerally
consideredpollinatordependentandmostcranberrygrowersrenthoneybeesfor
pollination.Theeconomicallyoptimalhivestockingdensity,orthepointatwhichthecost
ofaddinganadditionalhiveisgreaterthantheeconomicreturn,however,isunknown,
resultinginwidevariationinmanagementpractices.Therecommendedstockingdensity
is4.9‐7.4hives/ha(2‐3hives/ac),butsomeWisconsincranberrygrowersrentasmanyas
22hives/ha(9hives/ac).Anotherexplanationforthisvariationinmanagementisthatthe
economicallyoptimalstockingdensityvariesdependingonthesurroundinglandscapedue
todifferinglevelsofbackgroundpollinationbynativebees.Therefore,theobjectiveofthis
36
2
studywastodeterminetherelationshipbetweencranberryyieldandhoneybeehive
stockingdensityandwhetherthisrelationshipisconsistentacrossthelandscape.To
addressthisobjective,wecollectedhistoricalproductionrecordsfromWisconsin
cranberrygrowersregardingtheuseofhoneybeesandcranberryyieldfor2000‐2011.We
foundastrongpositiverelationshipbetweenyieldandhives/habutonlywhenthe
proportionofwoodlandwithin1kmofthecranberrymarshwaslow(<41.6%).
Furthermore,thevariationinyieldatmarshesinlow‐woodlandlandscapesdecreasedas
thenumberofhivesincreased,indicatingthatincreasinghivesleadstohigheryieldswith
lessvariationinthatyield.Althoughtherewasnoevidencethatincreasinghiveswas
beneficialatmarshesinhigh‐woodlandlandscapes,theyieldatthesemarsheswhenhoney
beeswereabsentwasmarginallyhigherthanatmarshesinlow‐woodlandlandscapes.
Overtherangeofhivesobserved,wewereunabletodeterminetheoptimalhivestocking
densityalthoughWisconsincranberrygrowerswithmarshesinlow‐woodlandlandscapes
canexpectincreasingreturnsonyielduptoandabove17hives/ha(7.6hives/ac).The
valueofeachadditionalhivewas1098kg/hive(24.2bbl/ac)atmarshesinlow‐woodland
landscapesand0kg/hiveinhigh‐woodlandlandscapes.Theseresultssuggestthatthe
habitatsurroundingacranberrymarshinfluencestheeffectivenessandeconomically
optimalstockingdensityofhoneybeesforcranberrypollination.
Keywords:honeybees,Apismellifera,cranberry,hivestockingdensity,pollinator
37
3
1.0 Introduction
Pollinationisanimportantecosystemservicevaluedat$15billion/yeartoUS
agriculturealone(Calderone2012).Two‐thirdsofcropplantsrequireorbenefitfrom
insectpollination(Kleinetal.2007)andprovidethemainsourceofseveralkeyvitamins
andmineralsinthehumandiet(Eilersetal.2011).Asaresult,managedhoneybeesare
widelyusedinmanyagriculturalsystemsforcroppollination.Honeybees(Apismellifera)
occurinlargecolonies,areeasytomanage,andwillvisitmanydifferentplantspecies,
makingthemanidealspeciesforpollinationinmodernagriculturalsystems(Kremenand
Chaplin‐Kramer2007,Calderone2012).Nativebeesarealsoeffectivecroppollinators,but
occurinsmallnumbers,aremoredifficulttomanage,andoftentimesarespecialized
pollinatorsofoneplantfamilyorgenus(DelaplaneandMayer2000,Michener2000).Even
so,nativebeesaloneareabletoprovidefullpollinationservicesinsomesystems(e.g.,
watermelon,Winfreeetal.2008,cranberry,MohrandKevan1987).
Cranberryisonecropwherehoneybeesarecommonlyusedtofulfillpollination
requirements,althoughhivestockingdensitiesvarygreatly.Currentmanagement
recommendationsforWisconsincranberrycallfor4.9‐7.4hives/ha(2‐3hives/ac,
McGregor1976),butgrowersuseanywherefrom0‐22hives/ha(0‐9hives/ac,H.Gaines,
pers.obs.).Oneexplanationforthevariationinmanagementpracticesmaybethatsome
marsheshaveahigherlevelofbackgroundpollinationratesduetothepresenceofnative
beesandthereforedonotrequireasmanyhoneybees.Previousstudieshavealsoshown
thatsomenativebeesare,infact,moreefficientcranberrypollinatorsthanhoneybees
(CaneandSchiffhauer2003).Additionally,currentmanagementrecommendationsare
38
4
basedonresearchfromover50yearsagowhenyieldsweremuchlower,landscapeswere
lessdeveloped,andthepollinationrequirementsofcranberrywerenotfullyunderstood,
resultinginconflictingrecommendations.Forexample,FilmerandDoehlert.(1955)
suggestusing0.5hives/ha(1hive/5ac)whereasFarrarandBain(1947)suggestfrom2.5‐
24.7hives/ha(1‐10hives/ac).Changesintechnologyandmanagementpracticesalong
withabetterunderstandingofthepollinationrequirements(Gaineschp.1)suggestthat
updatedrecommendationsmaybeneeded.Furthermore,economictheoryregarding
optimalinputuseimpliesthattheoptimalstockingdensitywilldependonthecropprice
andthecostofrentinghives(Beattieetal.2009).Theoptimalinputisreachedwhenthe
costofaddinganadditionalhiveexceedstheeconomicreturnofaddingthathive.
Onefactorthatmayinfluencetheoptimalstockingdensityforhoneybeesisthe
amountofnaturalhabitatinthesurroundinglandscape.Previousstudieshaveshownthat
nativebeesrespondpositivelytoamountofandproximitytonaturalhabitatandthat
pollinationservicesandyieldincreasewithbeediversity(Kremenetal.2002,Rickettsetal.
2008).Therefore,cranberrymarsheswithmorenaturalhabitatinthesurrounding
landscapemayhavehigherbackgroundpollinationratesthanmarsheswithlittlenatural
habitat,reducingtheamountofhoneybeesrequired.Thehypothesisthatoptimalhive
stockingdensitiesvarywithlocalconditionsissupportedbythelargevariationin
managementpracticesobservedinWisconsinandthefactthataveragestockingdensities
forcranberryalsovarybygrowingregion(R.Serres,pers.comm.).
Cranberrygrowersspendasignificantamountofmoneyonpollinationservices,but
sincetheoptimalhivestockingdensityisunknown,theymaynotbespendingtheirmoney
39
5
inanefficientway.Becauseofthemagnitudeofthecostofhoneybeehiverentals,
determiningtheoptimalhivestockingdensitywouldallowgrowerstomaximizetherateof
returnontheirinvestmentinhivesandpotentiallysavethemthousandsofdollarsevery
year.Thegoalofthisstudywastodeterminetheeconomicallyoptimalhoneybeehive
stockingdensityforWisconsincranberryandwhetherthisrateisconsistentacrossthe
landscape.WeusedhistoricaldatafromWisconsincranberrygrowerstodetermine(1)the
relationshipbetweencranberryyieldandhoneybeehivestockingdensity,(2)themarginal
valueofeachadditionalhiveofhoneybeesforcranberryyield,and(3)iftheeconomically
optimalrateremainsconstantacrossthelandscape.
Methods
2.1Growerdata
Datawerecollectedfrom38cranberrygrowers,eitherthroughpreviously
establishedgrower‐collaboratorsorothergrowersreachedthrougharequestattheannual
meetingoftheWisconsinStateCranberryGrowersAssociation(WSCGA).Growerswere
askedtocompleteaworksheetincludinginformationonmarshlocation,establishment
year,varietiesgrown,areaincranberryproduction,honeyandbumblebeeuse,andyield
recordsfortheperiod2000‐2011.Becausecranberrytakesabout3yearsfrommarsh
establishmenttoproduceacrop,onlydatafrommarshesmorethan3yearsoldwereused.
Datafromyearsinwhichsignificantdamagetothecropduetopestsorhailwasrecorded
wereexcluded.Datafromyearsinwhichgrowerswerefollowingamarketingorder
volumecontroltoreduceyieldwerealsoexcluded.Amarketingordervolumecontrolisa
40
6
toolusedbyagriculturalindustriestoreducetheover‐productionofacropbylimitingthe
amountofcropagrowercansell(JesseandRogers2006).Thisresultedinadatasetwith
38uniquegrowerswithanaverageof9.6years(min.1‐max.12years)ofyieldandhive
data.Thedatathatwerereportedincludedtotalacresofcranberrycultivated,total
numberofhivesrented,andtotalcranberryharvestinbarrels(1bbl=100lbsor45.4kg)
foreachyear.Ithenconvertedthesedatatohives/haandyield/ha.
2.2Landscapedata
Sinceotherdatasetssuggestthatwoodlandisagoodpredictorvariablefornative
beesandtheirpollinationservices(Gaineschp.3),Icalculatedtheproportionwooded
habitatwithin1kmofeachmarshusingtheCroplandDataLayer(USDA2011)andArcGIS
10.0(ESRI2011).Acentroidwaslocatedinthecenterofeachmarshanda1kmradius
drawnaroundthatpoint.Nativebeeabundanceandspeciesrichnessispositively
associatedwithnaturalhabitatinthesurroundinglandscape(Gaineschp.3),sowooded
habitatwasusedasaproxyforbackgroundnativepollinatorspresentinthelandscape.
2.3Statisticalanalyses
Todeterminetherelationshipbetweencranberryyield(kg/ha)andhoneybeehive
stockingdensity(hives/ha)asafunctionofthepercentwoodlandwithina1kmradius
aroundeachmarsh,weusedamultiplemixedmodellinearregressionwithhives/ha,
percentwoodland,andtheirinteractionasfixedeffectsandfarmasarandomeffect.Since
theinteractiontermbetweenhives/haandwoodlandcategorywassignificant(p=
0.0003),theanalysiswasalsoconductedseparatelyusingthedatafrommarshesinhigh‐
andlow‐woodlandlandscapes.AShapiro‐WilkGoodnessofFittestonyieldindicatedthat
41
7
thedatawerenormallydistributed(W=0.993,p=0.0757),sonotransformationswere
used.Atermforyearandsecondordertermforhives(anindicationofcurvature)were
initiallyincludedinthemodelbutwerefoundnotsignificantandwerethereforedropped
fromthefinalmodel(year,p=0.095F1,212.1=2.82,hives2,p=0.50F1,205.1=0.45).
Todeterminewhetherthevariationinyieldvariedasafunctionofhivestocking
density,weusedalinearmodelwiththecoefficientofvariation(standarddeviationof
yielddividedbythemean)asthedependentvariableandthehoneybeestockingdensityas
theindependentvariable.Separateregressionswereestimatedforsitesinhigh‐woodland
areasandlow‐woodlandareas.
Todeterminethedensityatwhichtheeffectofhoneybeesonyieldwassignificantly
greaterthanyieldwithnohives,weusedthepredictionequationtopredictyieldatagiven
hivedensity.Wevariedhives/hatoidentifytheloweststockingdensityatwhichtheupper
limitofthe95%confidenceintervalwithzerohives/hawaslessthanthelowerlimitofthe
95%confidenceinterval.AllstatisticalanalysiswasdoneusingJMPPro10(SASInstitute
Inc.2007).
2.0 Results
Cranberryyieldwas,onaverage,positivelyassociatedwiththenumberofhoney
beehives/ha(p<0.0001,Adj.R2=0.49,fig.1),butthisrelationshipvarieddependingonthe
amountofwoodlandinthesurroundinglandscape(hivesxlandscape,p=0.0091).Atfarms
withalowamountofwoodedhabitat(lessthanaverage,<41.6%)withina1kmradius,the
relationshipwasstrongandpositive(p<0.0001,Adj.R2=0.58),butatsiteswithahigh
42
8
amountofwoodedhabitat(aboveaverage,≥41.6%)within1km,therelationshipwasnot
significant(p=0.51,Adj.R2=0.25).Fromtheslopeoftheserelationships,wedetermined
thatthemarginalvalueofeachadditionalhivewas1098kg/hive(24.2bbl/hive)atlow‐
woodlandsitesand0kg/hive(nosignificantrelationship)athigh‐woodlandsites.
Variationincranberryyieldwasalsoeffectedbyhivedensity,butagain,onlyatsites
inlow‐woodlandlandscapes.Whentheproportionofwoodedhabitatishigh,increasing
hivedensityhasnoeffectonthecoefficientofvariation(CV,p=0.3088,Adj.R2=0.006,fig.
2).Whentheproportionofwoodedhabitatislow,increasinghivedensityresultsin
decreasedyieldCV(p=0.0392,Adj.R2=0.10).Theseresultssuggestthatgrowersathigh‐
woodlandsitesdonotgainanybenefitfromtheuseofhoneybeesbutgrowersatlow‐
woodlandsitesseeanincreaseinyieldandadecreaseinvariationbyusinghoneybees.
Theminimumnumberofhivesneededtorealizeasignificantyieldincreasewas4.2
hives/ha(1.7hives/ac)forsitesinlow‐woodlandlandscapes.Growersinhigh‐woodland
landscapesneverrealizedastatisticallysignificantyieldincreaseregardlessofthenumber
ofhives/haused.However,growersinahigh‐woodlandlandscapehadamarginallyhigher
yieldatzerohives/hathangrowersinalow‐woodlandlandscapeandallgrowershada
significantyieldatzerohives/ha(intercept,p<0.0001).Atzerohives/ha,yieldwas21.5
metrictons/ha(95%CI,19.1‐24.0)(192.4bbl/ac,95%CI,170.9–213.9)atlow‐woodland
sitesand25.3metrictons/ha(95%CI,22.6‐28)(226.2bbl/ac,95%CI202.2‐250.1)athigh‐
woodlandsites,whichrepresentsan18%increaseinyield.
Basedonthesedata,theeffectofvariationinhivestockingdensitydependsonthe
surroundinglandscape,however,anoptimalstockingdensity,thepointatwhichyieldis
43
9
maximized,couldnotbeestimatedsinceyieldscontinuetoincreasepasttheranges
reportedinthisstudyforfarmsinlow‐woodlandlandscapes.Wefoundnoevidenceof
yieldincreaseswiththeuseofhoneybeesatanyobservedstockingdensityforfarmssetin
high‐woodlandlandscapes.Wewereunabletodeterminetheoptimalhoneybeehive
stockingdensityforlow‐woodlandsitesbecausetherelationshipbetweenyieldand
hives/hadidnotreachapointofdiminishingreturns(fig.1).Therelationshipbetween
yieldandhives/haislinearfrom0hives/hato18.8hives/ha,thehigheststockingdensity
ofreportedbythegrowersforthetwelveyearsofdatareported.
3.0 Discussion
Individualcranberrygrowersspendtensofthousandsofdollarseachyearrenting
honeybeesforpollinationandstockingdensitiesvarywidely,suggestingthatgrowersmay
notbeoptimizingtheiruseofhoneybees.Bytakingadvantageofawealthofdataalready
heldbythegrowers,wewereabletoassembleapowerfuldatasetconsistingofmorethana
decadeofharvestdatafromnearly40differentmarshes.Thesedatarevealedthatyieldis
positivelyassociatedwithhivestockingdensity,butthatthisrelationshipdependsonthe
surroundinglandscape.Ourresultssuggestthatgrowerswhosemarshesaresetin
landscapeswithahighproportionofwoodlandmaybewastingmoneyonhoneybees,
whereincreasinghoneybeestockingdensitiesdidnotsignificantlyincreasecranberry
yield.Incontrast,atcurrentcranberrypricesandhivecosts,growerswhosemarshesare
setinlandscapeswithalowproportionofwoodlandshouldincreasetheirnetincomeby
usingahivestockingdensitygreaterthanthemaximumof18.8hives/ha(7.6hives/ac)
44
10
seeninourdata.Thisstudyprovidesthefirstrigorousassessmentofthevalueofhoney
beehivesforcranberryyieldthatweareawareof.
Themostobviouspatterninourdatasetwasthepositiverelationshipbetween
cranberryyieldandhives/hainsimplifiedlandscapes.Furthermore,formarshesinthese
low‐woodlandlandscapes,therelationshiphasnotreachedapointofdiminishingreturns,
suggestingthatgrowerscouldincreasestockingdensitiestoachievehigheryields.This
relationshipsupportstheuseofmanagedhoneybeesforcranberryyield,butalsosuggests
thatthepreviousrecommendationof4.9‐7.4hives/ha(2‐3hives/ac)requiresupdating.
Ourdatasuggestthatgrowersinlow‐woodlandlandscapescanincreasetheircranberry
yieldbyover1000kg/hiveforeachadditionalhoneybeehive/acused.Beyond18.8
hives/ha(7.6hives/ac),wecannotpredictwhatwillhappen.
Fromaneconomicperspective,cranberrygrowersappeartobebelowthemaximum
economicreturntheycouldachieveusingmanagedhoneybees.Growersinhigh‐woodland
landscapesusing10hives/hamaybewasting$700/ha($380/ac)whichonanaveraged
sizedmarshof32ha(80ac),amountsto$22,400/year.Inlow‐woodlandlandscapesatthe
stockingdensityreportedinourdata,however,theeconomicreturnsarestillincreasing
(BeattieandTaylor2009).Forthemosteconomicallyprofitablelevelofproduction,
growersshouldoperateatthepointwherediminishingreturnstoinputuseoccurs.A
recentstudyofblueberrypollinationfoundasimilarresultwithastronglinear
relationshipbetweenyieldandhivestockingdensity,suggestingthatblueberrygrowers
arealsonotmaximizingtheiruseofhoneybees(EatonandNams2012).
45
11
Economictheoryrecommendsincreasinguseofaninputuntilthemarketvalueofthe
additionalyieldgeneratedbytheadditionalinputjustequalsthecostoftheadditional
input.Atcurrentmarketpriceforcranberries($1.21/kgforcooperativeand$0.49/kgfor
independentgrowers,$65/bbland$22/bblrespectively)andhiverentalfees($70/hive),
theadditionofasinglehive/haprovidesan18:1returnoninvestmentsforcooperative
growersanda6.6:1returnoninvestmentsforindependentgrowersinlow‐woodland
landscapes.Withalinearresponseofyieldtohives/ha,growersinlow‐woodland
landscapeshavenotreachedtheeconomicallyoptimalhivestockingdensityandcould
continuetoincreasestockingdensityabovethemaximumof18.8hives/ha(7.6hives/ac)
reportedinthesedata.Theinteractingeffectofthelandscapealsosuggests,however,that
thereisnot,infact,asingleoptimalstockingdensityforcranberryindustry‐wide.
Theinteractionbetweenhivestockingdensityandsurroundinglandscapesuggeststhat
honeybeesarelesseffectiveatincreasingcranberrypollinationinhigh‐woodland
landscapes.Onepossibleexplanationforthisisthathoneybeesaremovingfromthecrop
intothenon‐cropinordertoaccessfloralresources.Vaissière(1990)foundthathoney
beeforagingdensitywasgreaterincottonfieldssurroundedby“poor”habitatwhichmay
explainthepatternseeninourstudyofahighercontributionofhoneybeestocranberry
pollinationinlow‐woodlandlandscapes.Pettisetal.(2013)furtherdemonstratedthat
honeybeesonaNewJerseycranberrymarshcollectedminimalamountsofcranberry
pollen,furthersupportingtheideathathoneybeesareforaginginnon‐crophabitats
surroundingthecranberrymarshes.
46
12
Anotherpatternevidentfromourdatasetisthatevenwhentherearenohivesofhoney
beespresent,growersstillachieveasignificantyield.Thiscouldbeduetoanumberof
factorsincludingferalhoneybees,nativebees,orothernon‐bioticvectorsofpollination.
Basedonadditionaldatacollectedbytheauthors,marsheswherehoneybeeswerenot
rentedhadveryfewifanyhoneybeespresent,nativebeeswerepresentinhighabundance
anddiversity(nearly200species)especiallyatmarshesinhigh‐woodlandlandscapes
(Gaineschapter3),andother,non‐bioticfactorscancontribute20‐80%oftotalcranberry
yield(Gaineschapter1).Previousstudieshavedemonstratedthatnativebeesareefficient
cranberrypollinators(CaneandSchiffhauer2003,Cariveauunpubl.data),suggestingthat
nativebeesmaybecontributingtothisyieldintheabsenceofhoneybees.Additionally,
assumingthatpollinationservicesalsoincreasewithnativebeeabundanceanddiversity,
thiscouldexplainwhythehigh‐woodlandmarsheswherenohoneybeesareusedhave
about20%greateryieldsthanlow‐woodlandsites,sincenativebeesarepositively
associatedwithnaturalhabitat(Rickettsetal.2008).
Despitethestrongandsignificantpatternsfoundinthisstudy,otherunmeasured
correlatesoflandusecouldcontributetotheobservedpatterns.Hivestockingdensitymay
becorrelatedwithothermanagementpracticesthatalsoaffectyield.Forexample,theuse
ofhighhoneybeestockingdensitiescouldalsobeassociatedwithmoreintensivefarming
practiceswithhigheroverallinputs(e.g.,fertilizer,pesticide,irrigation).Moredata
regardingmanagementpractices,suchastheapplicationoffertilizerorpesticides,
irrigation,andtheintensityofcropscoutingwhichmaybecorrelatedwithhivedensity
couldbeobtainedfromgrowers.Thesedatawouldallowfortheeffectofhivedensitytobe
47
13
disentangledfromothermanagementpractices.Finally,smaller‐scalefieldstudiesthat
experimentallyvarythehivestockingdensitycouldbeusedtoverifythepatternsobserved
inthishistoricaldataset.
Thisstudyrepresentsthefirstattempttodetermineeconomicallyoptimaluseofhoney
beesforcranberryproductionusingdatafromalargetimespanandacrossmanysites.
OurresultssuggestthattheoptimalstockingdensityhasnotbeenreachedonWisconsin
cranberrymarshesandfurtherresearchshouldexploreatwhatpointtheyieldresponse
curvereachesarateofdiminishingmarginalreturns.Ourresultsfurthersuggestthat
Wisconsincranberrygrowersinlow‐woodlandlandscapesshouldbeusingmorethan18.8
hives/ha(7.6hives/ac)tooptimizetheeconomicvalueofhoneybees.
Acknowledgements
Theauthorswouldliketothankallofthecranberrygrowerswhosharedtheiryieldand
honeybeerentalrecordswithus.ThisresearchwasfundedbyaHatchGrantawardedto
C.G.
Literaturecited
Beattie,B.R.,C.R.Taylor,andM.J.Watts.2009.Theeconomicsofproduction,2ndedition.KriegerPublishingCompany,Malabar,Fla.
Calderone,N.W.2012.Insectpollinatedcrops,insectpollinatorsandUSagriculture:Trend
analysisofaggregatedatafortheperiod1992‐2009.PlosOne7:e37235.Cane,J.H.,andD.Schiffhauer.2003.Dose‐responserelationshipsbetweenpollinationand
fruitingrefinepollinatorcomparisonsforcranberry(Vacciniummacrocarpon[Ericaceae]).AmericanJournalofBotany90:1425–1432.
48
14
Delaplane,K.S.,andD.F.Mayer.2000.Croppollinationbybees.CABIPublishing,NewYork,NY.
Eaton,L.J.,andV.O.Nams.2012.Honeybeestockingnumbersandwildblueberry
productioninNovaScotia.CanadianJournalofPlantScience92:1305–1310.Eilers,E.J.,C.Kremen,S.SmithGreenleaf,A.K.Garber,andA.‐M.Klein.2011.Contribution
ofpollinator‐mediatedcropstonutrientsinthehumanfoodsupply.PLoSONE6:e21363.
ESRI.2011.ArcGIS.EnvironmentalSystemsResearchInstitute,Redlands,CA.Farrar,C.L.,andH.F.Bain.1947.Honeybeesaspollinatorsofthecranberry.Cranberries
11.Filmer,R.S.,andC.A.Doehlert.1955.Useofhoneybeesincranberrybogs.GleaningsinBee
Culture:265–268.Jesse,E.V.,andR.T.Rogers.2006.ThecranberryindustryandOceanSprayCooperative:
Lessonsincooperativegovernance.FoodSystemResearchGroupMonographs19.Klein,A.‐M.,B.E.Vaissière,J.H.Cane,I.Steffan‐Dewenter,S.A.Cunningham,C.Kremen,and
T.Tscharntke.2007.Importanceofpollinatorsinchanginglandscapesforworldcrops.ProceedingsoftheRoyalSocietyB:BiologicalSciences274:303–313.
Kremen,C.,andR.Chaplin‐Kramer.2007.Insectsasprovidersofecosystemservices:Crop
pollinationandpestcontrol.Pages349–382inA.J.A.Stewart,T.R.New,andO.T.Lewis,editors.InsectConservationBiology.CABIPublishing,Wallingford,Oxfordshire.
Kremen,C.,N.M.Williams,andR.W.Thorp.2002.Croppollinationfromnativebeesatrisk
fromagriculturalintensification.ProceedingsoftheNationalAcademyofSciences99:16812–16816.
McGregor,S.E.1976.Insectpollinationofcultivatedcropplants.AgriculturalResearch
Service,USDepartmentofAgriculture,Washington,D.C.Michener,P.C.D.2000.Thebeesoftheworld,2ndedition.TheJohnsHopkinsUniversity
Press,Baltimore,Maryland.Mohr,N.A.,andP.G.Kevan.1987.Pollinatorsandpollinationrequirementsoflowbush
blueberry(VacciniumangustifoliumAit.andV.myrtilloidesMichx.)andcranberry(V.macrocarponAit.)inOntariowithnotesonhighbushblueberry(V.corymbosumL.)
49
15
andlingonberry(V.vitis‐ideaeL.).Pages149–154ProceedingsoftheEntomologicalSocietyofOntario.
Pettis,J.S.,E.M.Lichtenberg,M.Andree,J.Stitzinger,R.Rose,andD.vanEngelsdorp.2013.
CroppollinationexposeshoneybeestopesticideswhichalterstheirsusceptibilitytothegutpathogenNosemaceranae.PLoSONE8:e70182.
Ricketts,T.,J.Regetz,I.Steffan‐Dewenter,S.Cunningham,C.Kremen,A.Bogdanski,B.
Gemmill‐Herren,S.Greenleaf,A.Klein,M.Mayfield,L.Morandin,A.Ochieng,andB.Viana.2008.Landscapeeffectsoncroppollinationservices:Aretheregeneralpatterns?EcologyLetters11:499–515.
SASInstituteInc.2007.JMPPro10.SASInstituteInc.,Cary,NC.USDA.2011.NationalAgriculturalStatisticsServiceCroplandDataLayer.USDA‐NASS,
Washington,D.C.Vaissière,B.E.1990.Honeybeestockingrate,pollinatorvisitation,andpollination
effectivenessinuplandcottongrownforhybridseedproduction.ActaHorticulturae(ISHS)288:359–363.
Winfree,R.,N.M.Williams,H.Gaines,J.S.Ascher,andC.Kremen.2008.Wildbee
pollinatorsprovidethemajorityofcropvisitationacrossland‐usegradientsinNewJerseyandPennsylvania,USA.JournalofAppliedEcology45:793–802.
50
16
Figurecaptions
Figure1:Cranberryyieldversushoneybeehivestockingdensity.Eachpointrepresentsa
singlemarshinasingleyear.Therewasastrongpositiverelationshipbetweenyieldand
hivestockingdensitybutonlywhenthepercentwoodlandinthesurroundinglandscape
waslow.
Figure2:Coefficientofvariationinyieldversushoneybeehivestockingdensity.Variation
inyielddecreasedwithincreasinghivestockingdensityatcranberrymarsheswherethe
surroundinglandscapewaslowinwoodland.Atmarshessetinhigh‐woodlandlandscapes,
therewasnorelationshipbetweenvariationinyieldandhivestockingdensity.
51
Yiel
d (m
etric
tons
/hec
tare
)
Hives/hectare
Figure 1
0 5 10 15 200
10
20
30
40
50
<41.6% woodland
≥41.6% woodland
Lowwoodland
Highwoodland (ns)
52
Lowwoodland
Highwoodland (ns)
0 5 10 15 200
5
10
15
20
25
30
35
40
Hives/hectare
Coe
ffici
ent o
f Var
iatio
nFigure 2
<41.6% woodland
≥41.6% woodland
53
1
CHAPTER3
Localandlandscapefactorsinfluencenativebees,yield,
andthecontributionofbeestoyield
Author:HannahR.Gaines
Co‐author(s):ClaudioGratton
Forsubmissionto:LandscapeEcology,BiologicalConservation,EcologicalApplications,
JournalofAppliedEcology
Abstract:Habitatlossandfragmentation,changesinlanduse,andintensivefarming
practicesareleadingcausesofbiodiversitylossaroundtheworld.Ofparticular
importanceinagroecosystemsisthelossofbeneficialinsectsthatprovideservicesto
farmersincludingpollination.Nativebeesprovidethemajorityofpollinationservicesin
somefarmingsystemsandhavethepotentialtoactasabufferagainstthecurrentdecline
inmanagedhoneybees.Inordertoensurethepersistenceofnativebeesandthe
pollinationservicestheyprovideinagroecosystems,itisimportanttounderstandhow
theyrespondtofeaturesofthesurroundinglandscapeandlocalfarmpractices.Therefore,
theobjectiveofthisstudywastodeterminetowhatextentnativebees,yield(ameasureof
pollinationsuccess),andthecontributionofbeestoyieldrespondtosurroundinghabitat
andlocalfactorsinaperennialcroppingsystem.Toaddressourobjective,wepan‐trapped
beesat49commerciallymanagedcranberrymarshesincentralWisconsinacrossa
54
2
landscapegradientfromhighlyagriculturaltohighlywooded.Yieldwasestimatedata
localfieldlevelandcollectedfromgrowersasamarsh‐wideaverage.Nativebee
abundanceandspeciesrichnesswerepositivelyassociatedwithopen,herbaceousand
woodedhabitat.Speciesrichnesswasalsopositivelycorrelatedtowoodededgedensity
andanindexoflocalpesticideuse.Yieldwaspositivelycorrelatedwithhoneybeehive
stockingdensityandcranberryfloraldensitybutwasnotassociatedwithnativebeesor
landscapefactors.Cagemanipulationstudiesweredoneatsitesalongalandscapegradient
andatsitesvaryinginlocalhoneybeeuse.Thecontributionofbeestocranberry
pollinationdidnotvaryasafunctionofsurroundinglandscapebutincreasedwith
cranberryfloweringuprightdensity.Theseresultssuggestthatbothlocalandlandscape
factorsareimportantpredictorsofbothnativebeeabundanceandrichnesswhilelocal
factorsaremoreimportantpredictorsofyieldandthecontributionofbeestoyield.Our
resultsdonotprovidestrongevidencethatnativebeesareastrongpredictorofyieldin
thissystem.
Keywords:nativebees,landscapeecology,EnvironmentalImpactQuotient,cranberry,
Vacciniummacrocarpon,structuralequationmodeling,lavaan
1.Introduction
Theexpansionofmodernagricultureandhumanactivityhaveledtothealterationand
fragmentationofthelandscapeanddestructionofnaturalhabitat.Agricultureiscurrently
thedominantland‐use,covering38%oftheterrestrialareaoftheglobe(Foleyetal.2005,
2011).Thislargescalealterationofthelandscapehasprofoundeffectsontheabilityof
55
3
organismstosurvive.Insectsareonegroupoforganismsthatprovidevaluableecosystem
servicesinagroecosystemssuchaspollination.Inordertofulfilltheirfeedingandnesting
resourcerequirements,manyofthesebeneficialinsectsmovebetweencropandnon‐crop
habitatsthroughouttheirlifecycle(Kremenetal.2007,Blitzeretal.2012).Becauseof
theirlimitedabilitytotravellongdistances(fromtensofmeterstoafewkilometers,
Zurbuchenetal.2010),theavailabilityofnon‐crophabitatwithintheagricultural
landscapeisofvitalimportance(Banaszak1992).
Althoughchangestothelandscapecanaffectthemovementandsurvivaloforganisms,
localfarmmanagementpracticescanalsohaveastronginfluenceonthepersistenceof
organismsinthearea,especiallythosewithlimitedflightdistancesthatmayexistprimarily
withinthefarmlandscape.Forexample,previousresearchhasshownthatfarmsthatuse
organicpracticesorhaveabundantfloweringnon‐cropplantssupportamorediverse
beneficialinsectcommunity(Bengtssonetal.2005,MorandinandKremen2013).
Additionally,thetiminganduseofpesticidescanhaveasignificantlethalorsub‐lethal
impactonbeneficialinsects(e.g.,bees,Desneuxetal.2007,Brittainetal.2010).Local
managementpractices,suchastillageandherbicideuse,affectthefloralandnesting
resourceavailabilityonthefarmwhichmaydirectlyaffectlocalscalepersistence.
Bees,onesuchexampleofmobileorganismsthatdeliverkeyservicestoagriculture
(Kremenetal.2007),arethemostimportantpollinatorsforcropproductionandare
sensitivetothecompositionandconfigurationofthelandscape,exposuretoagri‐
chemicals,andlocalresourceavailability(Pottsetal.2010).Historically,farmershave
reliedupononespecies,thenon‐nativehoneybee(Apismellifera),fortheirpollination
56
4
requirements.Inrecentyears,however,honeybeeshaveexperienceddrasticdeclinesasa
resultofmites,disease,andtherecentemergenceofColonyCollapseDisorder(CCD)(Ellis
etal.2010).Ashoneybeesdecline,farmerspayahigherpriceforhivesand,forsome
crops,mayneedtoseekalternativewaysofpollinatingtheircrops.
Nativebeesalsoprovidevaluablepollinationservices(LoseyandVaughan2006,
Winfreeetal.2008)andinsomesystemsareabletofullymeetthepollination
requirementsintheabsenceofhoneybees(watermelon,Winfreeetal.2008,cranberry,
MohrandKevan1987).Additionally,researchershavedemonstratedthatasnativebee
visitationincreases,cropyieldalsoincreasesregardlessofhoneybees(Garibaldietal.
2013).Byusingmanagementpracticesthatprotectorincreaseforageandnesting
resourcesfornativebees,farmersmayalsobeabletoincreasetheiryields.Although
farmersarenotlikelytoabandontheuseofhoneybeesforpollination,nativebeesmay
provideinsuranceagainstthedeclineinhoneybees(Winfreeetal.2007).Unfortunately,
declinesinnativebeepopulations(Pottsetal.2010,Cameronetal.2011)threatenthe
continuedprovisioningofthesepollinationservices,makingunderstandinghowlocaland
landscapefactorsinfluencenativebeepopulationsofgreatimportance.
Multiplestudieshavedocumentedthattheabundanceanddiversityofnativebees
increasewithproximitytonaturalhabitat(Kremenetal.2004,MorandinandWinston
2006)andareaswithdiversefloralresources(Pottsetal.2003,MorandinandKremen
2013).Currentresearchprovidesincreasingevidencethatthetypeandavailabilityof
habitatinthelandscape(i.e.,landscapecomposition)isimportantindeterminingnative
beeabundanceanddiversity.Arecentstudyprovidesevidencethatthearrangementof
57
5
habitat(i.e.landscapeconfiguration)isalsoimportant(Kennedyetal.2013).Fewstudies,
however,haveconsideredhowtheselandscapefactorsrelatetoyield(butseeRickettset
al.2004)orthecontributionofbeestoyield.Sinceyieldisauniversalmetricindicating
pollinationsuccess,understandinghowitrelatesdirectlyorindirectlytolocaland
landscapefactorscouldprovideinsightintofuturemanagementdecisionsinagro‐
ecosystems.
Inthisstudyweexaminedhownativebees,cropyield,andthecontributionof
nativeandmanagedbeestocropyieldvaryasafunctionoflocalandlandscapefeaturesin
theWisconsincranberryagro‐ecosystem.Previousstudieshaveshownthatnative,wild
beesareefficientcranberrypollinators(CaneandSchiffhauer2003)suggestingthattheir
conservationcouldprovideeconomicbenefitstocranberrygrowers.Theobjectivesofthis
studyweretodeterminetheinfluenceoflocalmanagementandlandscapefactorson(1)
nativebeeabundanceandspeciesrichness,(2)cranberryyield,and(3)thecontributionof
beestocranberryyield.Anunderstandingoftherelationshipsbetweenthesevariables
couldhaveimplicationsforfuturemanagementandconservationdecisions.
2.Methods
2.1StudySystemandSiteSelection
ThisstudywasconductedoncommercialcranberrymarshesincentralWisconsin
(USA),themaincranberrygrowingregioninthecountry.Thisregionisknownasthe
CentralSandPlainsandischaracterizedbysandysoilandflatopenterrain(DottandAttig
58
6
2004).Theareaisheavilyagriculturalandproducesmostofthestates’cranberryand
potatocrops.
Commercialcranberrymarshes(sites)wereselectedtospantheexistinglandscape
gradientfromhighlywoodedtohighlyagricultural(fig.1).Weuseddigitalorthophotosto
identifyallpossiblesitesandthencharacterizedthelandscapesurroundingeachsiteusing
remotelysensedland‐coverdata(USDA2011).Potentialsiteswereplottedonagraphof
percentwoodlandversuspercentagricultureandasubsetofthesesiteswerechosento
representtheentiregradient.Siteswereatleast2kmapartfromeachothertoensurea
minimumdegreeofspatialindependenceamongbeecommunities(Greenleafetal.2007,
Zurbuchenetal.2010).Wesampledat15sitesin2008,30sitesin2010,and20sitesin
2011foracombined49uniquesites(i.e.somesitesweresampledmultipleyears).
Samplingtookplaceinacornerofthemarshclosesttonon‐crophabitat.Eachmarshis
composedofagroupofindividualbeds(~40mx200m)andsamplingwasdonewithinthe
cornerbedandthebedimmediatelyadjacenttoit.
2.2Localvariables
Todeterminecranberryfloraldensity,flowerswithintenrandomlyplaced1ft2
quadratspersitewerecountedduringpeakbloomin2010and2011.
Localpesticidetoxicityscoreswerecalculatedforeachsitebasedonpesticidespray
recordscollectedfromthegrowersforeachsamplingyear.Weusedamodifiedversionof
theEnvironmentalImpactQuotientofPesticidesFieldScore(Kovachetal.1992)which
simplifiessprayrecordstoasingleindexbasedonavarietyoffactorsincludingtoxicitytoa
targetorganism,activeingredients,andthenumberofapplicationsmade.
59
7
mEIQ=∑(Z*P*3)(%activeingredient/100)(applicationrate)(#ofapplications)
whereZ=beetoxicity
andP=plantsurfacehalf‐life
Whenanapplicationratewasnotindicatedonsprayrecords,theraterecommendedonthe
materiallabelwasused.ThepercentactiveingredientwasextractedfromtheMaterial
SafetyDataSheet(MSDS)foreachchemical.Whenratewasinpounds/acre,onepound
wasconsidered16oz.
2.3Landscapevariables
Todescribethelandscapearoundeachsite,wecalculatedbothcompositionand
configurationmetricsusingPatchAnalyst4(Rempeletal.2008)inArcGIS10(ESRI2011).
Thecompositionprovidesanindexofhabitatandresourceavailabilitywhileconfiguration
describestheaccessibilityofthathabitattoorganismsinthelandscape.Forcomposition
(i.e.percentcover),wenarrowedtheoriginal56CDLland‐covercategoriesdownto14
categorieswhichrepresentedatleast1%ofthetotallandareawithin1kmofallsites.
These14categoriescombinedrepresent96.4%oftheland‐coverwithin1kmofallsites.
Tofurthersimplifythelandscapecompositiontojusttwoaxes,weperformedaPrinciple
ComponentsAnalysis(PCA)usingthe14categories(fig.2).ThePCAvaluesforaxis1and
axis2werethenusedinourstatisticalanalysistorepresentlandscapecomposition.
Landscapeconfigurationwasrepresentedbywoodededgedensity(i.e.lengthofwooded
edgeswithin1km).Woodededgesareexpectedtoprovidegoodbeehabitatwithaccessto
60
8
multiplehabitattypes(woodlandandopen/agriculture)andcreateauniquehabitattype
withanabundanceofnestingandforagingresources(Diaz‐Foreroetal.2013).
2.4Nativebeeabundanceandspeciesrichness
Tomeasurenativebeeabundanceandspeciesrichness,wepantrappedspecimens
fromMaytoAugustof2008,2010,and2011.Eachyear,samplesweretakenbefore,during
andaftercranberrybloom.Pantrapswerefilledwithsoapywaterandsetdowninthe
cranberryvinesintwoparalleltransectslocatedeitherimmediatelyadjacenttoanon‐crop
marshedgeor50metersawayfromtheedge.Thedecisiontolocatethetransectsneara
non‐farmedgewasbasedonresearchindicatingthatbeesrespondpositivelytoproximity
tonaturalhabitat(reviewedbyRickettsetal.2008),soiftherewereanyresponsebybees
tolandscape,itwouldbecapturedinourstudydesign.Eachtransectincluded5setsof
threepantraps(blue,yellow,andwhite,ACEFluorescentspraypaint)setat10mintervals.
Specifictrappingmethodsvariedslightlyfromyeartoyear.In2008,trapswere
madefrom5ozplasticbowls(Chinettebrand).Trapswereleftineachmarshfor6‐hour
intervalsbetween8amand4pmandallsiteswerevisitedwithinthreedays(5sitesper
day).Datawereonlycollectedondaysdeemedtobegoodbeedays(windbelow2.5m/s,
temperatureabove15oC,skyconditionssunnytobrightovercast).In2010and2011,traps
weremadefrom12ozplasticcups.Trapswereleftineachmarshfor3‐dayintervalsand
allsiteswerevisitedwithintwodays(~15sitesperday).
Specimenswerestoredinto70%alcoholthendried,pinned,labeled,andsortedto
morphospeciesinthelabbeforebeingsenttoataxonomistforspecieslevelidentification.
2.5Yield
61
9
Yielddatawascollectedbothempiricallyonasmallplotscaleinthefieldandata
marsh‐widescalefromgrowers.Toestimatetheplotlevelyield,four1ft2berrysamples
weretakenfromeachsiteduringthe2011fieldseason.Thismethodisusedbycropscouts
andfarmerstoestimateyieldbeforeharvest(J.Sojka,pers.comm.).Allberrieswithin
randomlyselected1ft2quadratswereharvestedinlateSeptemberbeforecranberry
harvestandweighedinthelab.Wetweightwasusedasthisiseasilyconvertedtothe
industrystandardyieldmetricofbarrels/acre(1bbl=100lbs).
Todeterminethedirecteffectsoflocalandlandscapeeffectsonnativebee
abundanceandspeciesrichnessandthedirectandindirecteffectsonyield,weusedpath
analysiswhichisaformofStructuralEquationModeling(SEM).SEMisausefulwayto
analyzecomplex,multivariatelinearrelationships,intercorrelationofpredictors,and
providesamethodtoallowseparationofdirectandindirecteffects(Grace2006).Fordata
collectedattheplotlevel(n=17),thesmallnumberofvariablesandpathswaslimitedby
oursamplesize.Incontrast,ourmarsh‐widedatahadfarmoreobservations(n=42)
allowingustoincludemorevariablesandpathsinourmodel.Ourhypothesizedmodelfor
ourestimatedplotlevelyieldincludeddirecteffectsofhoneybeehivestockingdensity,
nativebeeabundance,andnativebeerichnessonyield;directeffectsofPC1,PC2,and
woodededgedensityonnativebeeabundanceandspeciesrichness;andindirecteffectsof
PC1,PC2,andwoodededgedensityonyield(fig.3).Ourhypothesizedmodelforthe
marsh‐wideaverageyieldincludeddirecteffectsofhoneybeehivestockingdensity,
cranberryfloraldensity,mEIQ,nativebeeabundanceandnativebeespeciesrichnesson
yield;directeffectsofmEIQ,cranberryfloraldensity,woodededgedensity,PC1,andPC2
62
10
onnativebeeabundanceandspeciesrichness;andindirecteffectsofmEIQ,PC1,PC2,and
woodededgedensityonyield(fig.4).Alllocalandlandscapevariablesusedinboth
modelswerefromthesamedata.Inordertodetermineindirecteffects,wemultipliedthe
standardizedcoefficientsofdirecteffects(Mattesonetal.2013).Aχ2goodnessoffittest
wasusedtodeterminewhetherthemodelwasagoodfitofthedata.AllSEManalysiswas
doneusingtheLavaanpackage(Rosseel2012)inRStudio0.97.551(RStudio2012).
2.6Contributionofbeestoyield
Todeterminehowthecontributionofbeestocranberrypollinationvariesasa
functionoflocalorlandscapefactors,wedidcagemanipulationstudiesin2011and2012.
Cageswere50cmx50cmandwerecoveredwithnylontulletoexcludepollinators.Cages
weresetupbeforecranberrybloomandremovedoncebloomwascomplete.Allberries
withina1ft2areainthecenterofthecagesandinseparateopenplotswereharvestedin
lateSeptemberandcountedandweighed(wet)toestimateyield.In2011,3cages/site
wereestablishedat30sites(n=90cages)alongalandscapegradientofhighwoodedto
highagriculture.In2012,10cages/sitewereestablishedat11sites(n=110cages),5of
whichhadhoneybeespresentand6absent.Allsitesin2012werelocatedinhighwooded
landscapestocontrolforlandscapeeffectsonnativebeesoryield.
Todeterminewhetherthecontributionofbeestoyieldvariesasafunctionoflocal
andlandscapefactors,wedidamultipleregressiononyieldwithcagetreatment,local,and
landscapefactorsaspredictorvariables.Forour2011cagemanipulationacrossa
landscapegradient,ourpredictorvariablesincludedcagetreatment,PC1,PC2,cage*PC1,
andcage*PC2aspredictorvariablesandsiteasarandomterm.Tounderstandthe
63
11
influenceoflocaleffects,wedidamultipleregressiononouryielddatacollectedinour
2012cagemanipulationstudythatwasonfarmssetinhighwoodlandlandscapesonly.
Ourpredictorvariablesincludedcagetreatment,honeybeehivestockingdensity,native
beeabundance,densityofcranberryfloweringuprights,cage*honeybees,cage*nativebee
abundance,cage*uprightdensity,andfarmsiteasarandomterm.Thisanalysiswasdone
usingJMP10.0.0(SASInstituteInc.2007).
3.Results
Wecollectedatotalof6673specimensrepresenting182speciesofbeesoverthree
fieldseasons(Appendix3.1).TheseincludeseveralnewrecordsforthestateofWisconsin
includingCoelioxysimmaculata,Lasioglossumlusorium,Megachileaddenda,andXenoglossa
kansensisaswellas19specimensofBombusterricola,aspeciesbelievedtobeindecline
throughouttheMidwesternUnitedStates(Cameronetal.2011).Overallthemost
abundantspeciescollectedwereLasioglossumadmirandum(n=523representing8%ofthe
totalsample),Agapostemontexanus(n=425,6%),andLasioglossumpilosum(n=367,5%).
ThesedatarepresentanexpandedspecieslistandspeciesrangesforWisconsin(Wolfand
Ascher2008)andprovideacomprehensivelistofbeespeciesfoundintheWisconsin
cranberrysystem.
3.1PrincipleComponentsAnalysisonlandscape
OurPCAindicatedastronggradientoflandcoverfromwoodedhabitattorowcrop
agriculture(e.g.,corn,soybeans)(PC1,26.1%)andfromwetlandandcranberrytoopen,
herbaceousandwoodedhabitat(PC2,15.8%)(fig.2).Becauseofthewaterrequirements
64
12
forcranberryproductionandthehistoricpracticeofmanagingnaturalwetlandsfor
cranberryproduction,cranberrymarshes,herbaceouswetland,andopenwaterwere
clusteredtogether.
3.2Beeabundance,speciesrichness,andyield
3.2.1Modelbasedonmarshlevelestimates
Ourmodelbasedoncranberryyieldasreportedbygrowersatthemarshlevel
foundastrongpositivecorrelationbetweennativebeeabundanceandopen,herbaceous
andwoodedhabitatinthelandscape(PC2,R2=0.30,standardizedpathcoefficient=0.52,
p<0.001,fig.3).Therewasalsoamarginallysignificantpositiveassociationbetweenbee
abundanceandlocalpesticideintensity(mEIQ,standardizedpathcoefficient=0.25,
p=0.06).Beespeciesrichnesswasalsoassociatedwithbothlandscapecharacteristicsand
localfactors.Open,herbaceousandwoodedhabitatinthelandscapehadastrongpositive
associationwithnativebeespeciesrichness(PC2,R2=0.45,standardizedpathcoefficient=
0.61,p<0.001).Localpesticideintensityalsohadapositiveassociationwithbeespecies
richness,possiblyasaresultofacorrelationbetweensprayintensityandbee‐friendly
managementpracticessuchastheuseoflesstoxicchemicals(mEIQ,standardizedpath
coefficient=0.32,p=0.005).Cranberryfloraldensityhadamarginallysignificant,weak
negativeassociationwithnativebeespeciesrichness(standardizedpathcoefficient=‐0.11,
p=0.11).
Cranberryyieldasmeasuredatthemarshscalewasdirectly,positivelycorrelated
withhoneybeehivestockingdensity(R2=0.56,standardizedpathcoefficient=0.65,
p<0.001)andmarginallywithcranberryfloraldensity(standardizedpathcoefficient=
65
13
0.18,p=0.13).Therewerenoindirecteffectsoflandscapeonyield.The∆AICbetweenthe
hypothesizedmodelandourfinalmodelwas12.64.Anon‐significantχ2goodnessoffittest
indicatedthatourmodelwasagoodfit(p=0.62).
3.2.2Modelbasedonplotlevelestimates
Nativebeeabundancewasstronglycorrelatedwithincreasingwoodedandopen,
herbaceoushabitatinthesurroundinglandscape(PC2,R2=0.43,standardizedpath
coefficient=0.66,p<0.001,fig.4)butnotwithanylocalvariables.Nativebeespecies
richnesswasalsostronglycorrelatedwithincreasingwoodedandopen,herbaceous
habitat(PC2,R2=0.46,standardizedpathcoefficient=0.52,p=0.002),aswellasthelength
ofwoodededgesinthelandscape(woodededgedensity,standardizedpathcoefficient=
0.25,p=0.014).Richnesswasnotassociatedwithlocalfactors.
Cranberryyieldasmeasuredattheplotscalewasstronglyassociatedwithlocal
factorsbutonlymarginallywithlandscapefactors.Therewasadirect,positiveassociation
betweenyieldandhoneybeehivestockingdensity(R2=0.32,standardizedpathcoefficient
=0.43,p=0.02).Yieldwasmarginally,negativelyassociatedwithnativebeeabundance
(standardizedpathcoefficient=‐0.36,p=0.054)andindirectly,negativelywithincreasing
woodedandopen,herbaceoushabitatinthelandscape(PC2,standardizedpathcoefficient
=‐0.24).The∆AICbetweenthehypothesizedmodelandourfinalmodelwas7.84.Anon‐
significantχ2goodnessoffittestindicatedthatourmodelwasagoodfit(p=0.81).
3.3Contributionofbeestoyield
Excludingbeeshadasignificanteffectonyield,asdidthesurroundinglandscape.
Fromour2011cagemanipulationalongalandscapegradient,wefoundthatexcluding
66
14
pollinators(cage,p=0.0012,F1,25=14.0)andincreasingwoodlandinthesurrounding
landscape(PC2,p=0.0165,F1,25=6.6,fig.5)hadsignificantnegativeeffectsonyield.The
effectofexcludingbeeshadthesameeffectwhetherwoodlandinthesurrounding
landscapewashighorlow(cage*PC2,p=0.12,F1,25=2.6).
Whenwehadfinescalecranberryfloraldensitydata,wefoundthatfloraldensity
andthepollinatorexclusionhadaninteractingeffectonthecontributionofbeestoyield.
Theresultsofour2012cagemanipulationindicatethatthemostimportantfactors
affectingyieldwerepollinatorexclusion(cage,p<0.0001,F1,9=58.3),floweringupright
density(p<0.0001,F1,9=96.6),andtheirinteraction(p=0.0051,F1,9=13.5),indicatingthat
thecontributionbybeesisdifferentwhenfloraldensityishighcomparedtolow(fig.6).
4.Discussion
Previousresearchhasinvestigatedtheinfluenceoflocalandlandscapefactorson
nativebeeabundanceandspeciesrichness,butfewhaveconsideredhowthesefactors
influenceyieldorthecontributionofbeestoyield.Inthisstudy,wefoundthatnativebees
respondstronglytolandscapefactors(i.e.,open,herbaceousandwoodedhabitat)but
weaklytolocalmanagement(i.e.,sprayintensity,cranberryfloraldensity).Cranberry
yield,incontrast,respondsstronglytolocalmanagement(i.e.,honeybeehivedensity,
cranberryfloraldensity)butweaklytolandscapefactorsandnativebees.Thecontribution
ofnativeandmanagedbeestocranberryyieldwasinfluencedbylocalfactors(i.e.,
cranberryfloraldensity)butnotlandscapefactors.Asthedensityofcropflowers
increased,thecontributionofbeesalsoincreased.Theseresultssuggestthat,atleastin
67
15
thissystem,theeffectoflocalmanagementonyieldismuchstrongerthanthatofthe
landscape.Wefoundnoevidencethatnativebeescontributetocranberryyield.
Theclearestpatternfromourstudywasthepositiveinfluenceofwoodlandon
nativebees.Bothnativebeeabundanceandspeciesrichnesswerestronglyassociatedwith
increasingwoodlandinthesurroundinglandscape.Thissupportspreviousstudies
showingthatbeesrespondpositivelytonaturalhabitatinthesurroundinglandscape
(reviewedbyRickettsetal.2008,Kennedyetal.2013).Thispatternislikelyduetothefact
thatnaturalhabitatsprovideforageandnestingresourcesthroughouttheseason,whereas
agriculturallandwithfloweringcrops,likecranberry,onlyprovideforageresourcesfora
veryshorttimespan.Wealsofoundapositiveinfluenceofwoodededgedensityonspecies
richness.Woodededgeslikelyprovidegoodnestingsitesforstem,ground,andcavity
nestingbeesincloseproximitytoopenhabitatswhereforageresourcescouldbeabundant.
Thepositiverelationshipbetweenlocalpesticideintensity(i.e.,mEIQ)andnative
beespeciesrichnesswasunexpected.Wehypothesizedthatlocalpesticidesprayintensity
wouldbenegativelyassociatedwithnativebeessincemoreexposuretopesticidesshould
haveadetrimentaleffectonbees(Desneuxetal.2007,Brittainetal.2010).Onepossibility
isthatsummarizingsprayintoasingleindexisaninadequatemethodforrepresenting
actualsprayexposure.Previousstudiesthatusedsprayindicesasanindexofpesticideuse
havealsofoundunexpected(Cramptonetal.2010),inconsistent(Singleton2010),orno
effects(JenkinsandIsaacs2007)ofsprayindexonbeneficialinsectssuggestingthatthis
maybethecase.Anotherpossibilityisthatahighsprayindexratingiscorrelatedwith
otherbee‐consciouspracticessuchassprayingatnightorusinglessharmfulspraysduring
68
16
bloom.Evenifgrowersspraymoreoftenoruseharsherchemicals,whenandhowthey
applythechemicalswillhaveabiginfluenceonhowitaffectsbeneficialinsects.Wisconsin
cranberrygrowershaveadocumentedrecordoffollowingenvironmentallyconscious
managementdecisions(ColquhounandJohnson2010)soitmightbeexpectedthatthey
wouldfollowpollinator‐consciouspracticesaswell.
Anotherclearpatterninourstudywasthestrongrelationshipbetweencranberry
yieldandhoneybeestockingdensity,butthelackofassociationbetweenyieldandnative
beeabundance.Whilethereisastrongrelationshipbetweencranberryyieldandhive
stockingdensity(Gaines,chp.2),thepatternheremaypartlybeanartifactofscale.The
associationbetweenyieldandhoneybeehivestockingdensitywasstrongerinourmodel
basedonthemarsh‐wideyieldestimatesthaninourmodelbasedonplotscaleyield
estimates.Notsurprisingly,honeybeehivestockingdensitywerealsobasedonmarsh‐
wideaveragesprovidedbythegrowers,whereasnativebeeabundancewasbasedonaplot
levelsampletakeninthesameareaofthemarshastheplotscaleyieldestimates.
Additionally,honeybeehivesareoftendispersedacrosstheentiremarshwhereasnative
beeabundancedecreasestowardsthecenterofthemarshandawayfromanon‐cropfarm
edge(Gaines,unpubl.data).Thismayallowhoneybeestohaveamoreeveninfluenceon
marsh‐wideyield,whereasthenativebeeswouldhaveastrongerinfluenceattheedgeof
themarsh(EvansandSpivak2006).
Thelackofinfluenceoflandscapeonthecontributionofbeestocranberryyieldwas
surprising.Weexpectedthatthecontributionbybeeswouldincreasewithincreasing
woodlandinthesurroundinglandscapebecausetherewouldbeagreaterabundanceand
69
17
richnessofbeespresent.Furthermore,previousstudieshavefoundapositiveinfluenceof
naturalhabitatoncropyield(Ricketts2004,MorandinandWinston2006).Instead,we
foundthatyielddecreasedaswoodedandopen,herbaceoushabitatinthelandscape
increasedandthatthecontributionofbeestoyielddidnotchange.Thispatternsuggests
eitherthatthenativebeesarenotcontributingtocranberryyieldorthatthecontribution
oflocalhoneybeesmasksanycontributionbynativebees.
Despitealackofinfluenceofthelandscapeonbeecontributionstoyield,thelocal
factorcranberryfloraldensitywasanimportantpredictorofthecontributionbybeesto
yield.Asfloraldensityincreased,thecontributionbybeesalsoincreased.Beesare
attractedtolargefloraldisplays,soincreasingthedensityofflowersmaymakethe
cranberrymarshmoreattractivetobees.Alternatively,floraldensitymaybecorrelated
withothermanagementpracticessuchashigherhoneybeehivestockingdensitiesand
moreintensivepracticesthatareassociatedwithincreasedyield.
Theresultsofthisstudysuggestthatlocalfarmmanagement,includingtheuseof
managedbees,hasastrongeffectoncropyield.Althoughwefoundnoevidencethatnative
beesarecontributingtothepollinationofcommercialcranberry,someWisconsin
cranberrygrowersdonotusemanagedhoneybeesforpollination,suggestingthatthe
nativebeesdocontribute.Futureresearchshouldconsideralternativemethodsfor
determiningthecontributionofnativebeestoyield.
70
18
Acknowledgements
Wewouldliketoespeciallythankallofthecranberrygrowerswhoallowedusto
conductresearchontheirfarmsandwhoprovidedyielddatatous.Wecouldnothave
collectedandprocessedallofourdatawithoutthehelpofmanyfieldandlabassistants:
EmilyFricke,AdamHiggins,CarlKaiser,MerrittSingleton,AllisonDehnel,Collin
Schwantes,BrianHedberg,AmandaRudie,ChaseFritz,MollyWaytes,GavinJones,Federica
Lacasella,JuliaTiede,RachelMallinger,andAliNelson.JayneSojkaandDanMahrassisted
withinitialsiteselection.BeeidentificationwasverifiedbyMikeArduser,JasonGibbs,and
JohnAscher.StatisticalassistancewasprovidedbyBrianSpiesmanandTaniaKim.This
researchwasfundedbyaHatchGrantawardedtoCG.
Literaturecited
Banaszak,J.1992.Strategyforconservationofwildbeesinanagriculturallandscape.Agriculture,Ecosystems&Environment40:179–192.
Bengtsson,J.,J.Ahnström,andA.‐C.Weibull.2005.Theeffectsoforganicagricultureon
biodiversityandabundance:ameta‐analysis.JournalofAppliedEcology42:261–269.
Blitzer,E.J.,C.F.Dormann,A.Holzschuh,A.‐M.Klein,T.A.Rand,andT.Tscharntke.2012.
Spilloveroffunctionallyimportantorganismsbetweenmanagedandnaturalhabitats.AgricultureEcosystems&Environment146:34–43.
Brittain,C.A.,M.Vighi,R.Bommarco,J.Settele,andS.G.Potts.2010.Impactsofapesticide
onpollinatorspeciesrichnessatdifferentspatialscales.BasicandAppliedEcology11:106–115.
Cameron,S.A.,J.D.Lozier,J.P.Strange,J.B.Koch,N.Cordes,L.F.Solter,andT.L.Griswold.
2011.PatternsofwidespreaddeclineinNorthAmericanbumblebees.ProceedingsoftheNationalAcademyofSciences108:662–667.
71
19
Cane,J.H.,andD.Schiffhauer.2003.Dose‐responserelationshipsbetweenpollinationandfruitingrefinepollinatorcomparisonsforcranberry(Vacciniummacrocarpon[Ericaceae]).AmericanJournalofBotany90:1425–1432.
Colquhoun,J.,andH.Johnson.2010.Sustainablecranberryproductionforavibrantfuture:
TheWisconsinexperience.UniversityofWisconsin‐Extension.Crampton,L.A.,G.M.Loeb,K.A.Hoelmer,andM.P.Hoffmann.2010.EffectofInsecticide
RegimensonBiologicalControloftheTarnishedPlantBug,Lyguslineolaris,byPeristenusspp.inNewYorkStateAppleOrchards.JournalofInsectScience10.
Desneux,N.,A.Decourtye,andJ.‐M.Delpuech.2007.Thesublethaleffectsofpesticideson
beneficialarthropods.AnnualReviewofEntomology52:81–106.Diaz‐Forero,I.,V.Kuusemets,M.Mänd,A.Liivamägi,T.Kaart,andJ.Luig.2013.Influenceof
localandlandscapefactorsonbumblebeesinsemi‐naturalmeadows:amultiple‐scalestudyinaforestedlandscape.JournalofInsectConservation17:113–125.
Ellis,J.D.,J.D.Evans,andJ.Pettis.2010.Colonylosses,managedcolonypopulationdecline,
andColonyCollapseDisorderintheUnitedStates.JournalofApiculturalResearch49:134–136.
ESRI.2011.ArcGIS.EnvironmentalSystemsResearchInstitute,Redlands,CA.Evans,E.C.,andM.Spivak.2006.Effectsofhoneybee(Hymenoptera:Apidae)andbumble
bee(Hymenoptera:Apidae)presenceoncranberry(Ericales:Ericaceae)pollination.Journalofeconomicentomology99:614–620.
Foley,J.A.,R.DeFries,G.P.Asner,C.Barford,G.Bonan,S.R.Carpenter,F.S.Chapin,M.T.
Coe,G.C.Daily,H.K.Gibbs,andothers.2005.Globalconsequencesoflanduse.Science309:570.
Foley,J.A.,N.Ramankutty,K.A.Brauman,E.S.Cassidy,J.S.Gerber,M.Johnston,N.D.
Mueller,C.O’Connell,D.K.Ray,P.C.West,C.Balzer,E.M.Bennett,S.R.Carpenter,J.Hill,C.Monfreda,S.Polasky,J.Rockström,J.Sheehan,S.Siebert,D.Tilman,andD.P.M.Zaks.2011.Solutionsforacultivatedplanet.Nature478:337–342.
Garibaldi,L.A.,I.Steffan‐Dewenter,R.Winfree,M.A.Aizen,R.Bommarco,S.A.
Cunningham,C.Kremen,L.G.Carvalheiro,L.D.Harder,O.Afik,I.Bartomeus,F.Benjamin,V.Boreux,D.Cariveau,N.P.Chacoff,J.H.Dudenhöffer,B.M.Freitas,J.Ghazoul,S.Greenleaf,J.Hipólito,A.Holzschuh,B.Howlett,R.Isaacs,S.K.Javorek,C.M.Kennedy,K.M.Krewenka,S.Krishnan,Y.Mandelik,M.M.Mayfield,I.Motzke,T.Munyuli,B.A.Nault,M.Otieno,J.Petersen,G.Pisanty,S.G.Potts,R.Rader,T.H.Ricketts,M.Rundlöf,C.L.Seymour,C.Schüepp,H.Szentgyörgyi,H.Taki,T.
72
20
Tscharntke,C.H.Vergara,B.F.Viana,T.C.Wanger,C.Westphal,N.Williams,andA.M.Klein.2013.WildPollinatorsEnhanceFruitSetofCropsRegardlessofHoneyBeeAbundance.Science339:1608–1611.
Grace,J.B.2006.Structuralequationmodelingandnaturalsystems.CambridgeUniversity
Press,NewYork.Greenleaf,S.S.,N.M.Williams,R.Winfree,andC.Kremen.2007.Beeforagingrangesand
theirrelationshiptobodysize.Oecologia153:589–596.Jenkins,P.E.,andR.Isaacs.2007.Reduced‐riskInsecticidesforControlofGrapeBerry
Moth(Lepidoptera:Tortricidae)andConservationofNaturalEnemies.JournalofEconomicEntomology100:855–865.
Kennedy,C.M.,E.Lonsdorf,M.C.Neel,N.M.Williams,T.H.Ricketts,R.Winfree,R.
Bommarco,C.Brittain,A.L.Burley,D.Cariveau,L.G.Carvalheiro,N.P.Chacoff,S.A.Cunningham,B.N.Danforth,J.‐H.Dudenhöffer,E.Elle,H.R.Gaines,L.A.Garibaldi,C.Gratton,A.Holzschuh,R.Isaacs,S.K.Javorek,S.Jha,A.M.Klein,K.Krewenka,Y.Mandelik,M.M.Mayfield,L.Morandin,L.A.Neame,M.Otieno,M.Park,S.G.Potts,M.Rundlöf,A.Saez,I.Steffan‐Dewenter,H.Taki,B.F.Viana,C.Westphal,J.K.Wilson,S.S.Greenleaf,andC.Kremen.2013.Aglobalquantitativesynthesisoflocalandlandscapeeffectsonwildbeepollinatorsinagroecosystems.EcologyLetters16:584–599.
Kovach,J.,C.Petzoldt,J.Degni,andJ.Tette.1992.Amethodtomeasuretheenvironmental
impactofpesticides.NewYork’sFoodandLifeSciencesBulletin139.Kremen,C.,N.M.Williams,M.A.Aizen,B.Gemmill‐Herren,G.LeBuhn,R.Minckley,L.
Packer,S.G.Potts,andothers.2007.Pollinationandotherecosystemservicesproducedbymobileorganisms:aconceptualframeworkfortheeffectsofland‐usechange.EcologyLetters10:299–314.
Kremen,C.,N.M.Williams,R.L.Bugg,J.P.Fay,andR.W.Thorp.2004.Thearea
requirementsofanecosystemservice:croppollinationbynativebeecommunitiesinCalifornia.EcologyLetters7:1109–1119.
Losey,J.E.,andM.Vaughan.2006.Theeconomicvalueofecologicalservicesprovidedby
insects.Bioscience56:311–323.Matteson,K.C.,J.B.Grace,andE.S.Minor.2013.Directandindirecteffectsoflanduseon
floralresourcesandflower‐visitinginsectsacrossanurbanlandscape.Oikos122:682–694.
73
21
Mohr,N.A.,andP.G.Kevan.1987.Pollinatorsandpollinationrequirementsoflowbushblueberry(VacciniumangustifoliumAit.andV.myrtilloidesMichx.)andcranberry(V.macrocarponAit.)inOntariowithnotesonhighbushblueberry(V.corymbosumL.)andlingonberry(V.vitis‐ideaeL.).Pages149–154ProceedingsoftheEntomologicalSocietyofOntario.
Morandin,L.A.,andC.Kremen.2013.Hedgerowrestorationpromotespollinator
populationsandexportsnativebeestoadjacentfields.EcologicalApplications23:829–839.
Morandin,L.A.,andM.L.Winston.2006.Pollinatorsprovideeconomicincentiveto
preservenaturallandinagroecosystems.Agriculture,Ecosystems&Environment116:289–292.
Potts,S.G.,J.C.Biesmeijer,C.Kremen,P.Neumann,O.Schweiger,andW.E.Kunin.2010.
Globalpollinatordeclines:trends,impactsanddrivers.TrendsinEcology&Evolution25:345–353.
Potts,S.G.,B.Vulliamy,A.Dafni,G.Ne’eman,andP.Willmer.2003.Linkingbeesand
flowers:Howdofloralcommunitiesstructurepollinatorcommunities?Ecology84:2628–2642.
Rempel,R.S.,A.P.Carr,andD.Kaukinen.2008.PatchAnalystextensionforArcMap:Version4.2.Ontario.MinistryofNaturalResources.
Ricketts,T.H.2004.TropicalForestFragmentsEnhancePollinatorActivityinNearby
CoffeeCrops.ConservationBiology18:1262–1271.Ricketts,T.H.,G.C.Daily,P.R.Ehrlich,andC.D.Michener.2004.Economicvalueoftropical
foresttocoffeeproduction.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica101:12579–12582.
Ricketts,T.,J.Regetz,I.Steffan‐Dewenter,S.Cunningham,C.Kremen,A.Bogdanski,B.
Gemmill‐Herren,S.Greenleaf,A.Klein,M.Mayfield,L.Morandin,A.Ochieng,andB.Viana.2008.Landscapeeffectsoncroppollinationservices:Aretheregeneralpatterns?EcologyLetters11:499–515.
Rosseel,Y.2012.AnRpackageforstructuralequationmodeling.JournalofStatistical
Software48:URLhttp://www.jstatsoft.org/v48/i02/.RStudio.2012.IntegrateddevelopmentenvironmentforR.RStudio,Boston,MA.SASInstituteInc.2007.JMPPro10.SASInstituteInc.,Cary,NC.
74
22
Singleton,M.2010.BiologicalImpactsofInsecticideUseintheWisconsinCranberryIndustry.UniversityofWisconsin,Madison,Madison,WI.
USDA.2011.NationalAgriculturalStatisticsServiceCroplandDataLayer.USDA‐NASS,
Washington,D.C.Winfree,R.,N.M.Williams,J.Dushoff,andC.Kremen.2007.Nativebeesprovideinsurance
againstongoinghoneybeelosses.EcologyLetters10:1105–1113.Winfree,R.,N.M.Williams,H.Gaines,J.S.Ascher,andC.Kremen.2008.Wildbee
pollinatorsprovidethemajorityofcropvisitationacrossland‐usegradientsinNewJerseyandPennsylvania,USA.JournalofAppliedEcology45:793–802.
Wolf,A.T.,andJ.S.Ascher.2008.BeesofWisconsin(Hymenoptera:Apoidea:Anthophila).
GreatLakesEntomologist41:129–168.Zurbuchen,A.,L.Landert,J.Klaiber,A.Müller,S.Hein,andS.Dorn.2010.Maximum
foragingrangesinsolitarybees:onlyfewindividualshavethecapabilitytocoverlongforagingdistances.BiologicalConservation143:669–676.
75
23
Figurecaptions
Figure1:Thedistributionofcranberrymarshes(sites)alongalandscapegradientof
woodedhabitatandagriculture.Siteswerechosentospanasmuchofthegradientas
possible.
Figure2:APrincipleComponentsAnalysisshowingtherelationshipbetweenlandcover
typessurroundingoursites.Eachlandcovertyperepresentsatleast1%ofthetotalarea
withina1kmradiusofallsites.
Figure3:Hypothesized(gray)andfinal(black)structuralequationmodelexplainingnative
beeabundance,speciesrichness,andthemarsh‐wideaveragecranberryyield.Gray
arrowsindicateconnectionsthatwereremovedfromthemodel.Solidblackarrows
indicatepositiverelationshipswhiledashedarrowsindicatenegativerelationships.
Numbersadjacenttoarrowsarestandardizedregressioncoefficientsalongwiththeir
significance(*).Thecoefficientindicatesthestrengthoftherelationship.
Figure4:Hypothesized(gray)andfinal(black)structuralequationmodelexplainingnative
beeabundance,speciesrichness,andfieldestimatedcranberryyield.Grayarrowsindicate
pathsthatwereremovedfromthemodel.Solidblackarrowsindicatepositive
relationshipswhiledashedarrowsindicatenegativerelationships.Numbersadjacentto
arrowsarestandardizedregressioncoefficientsalongwiththeirsignificance(*).The
coefficientindicatesthestrengthoftherelationship.
Figure5:Yieldfromthe2011cagemanipulationexperimentacrossalandscapegradient
wasbestestimatedbyamodelthatincludedcagetreatmentandprinciplecomponent2.
76
24
Thecontributionofbeestoyielddoesnotvaryasafunctionofthelandscapeasindicated
byanon‐significantinteractionterm(cage*PC2,p=0.12).
Figure6:Yieldfromthe2012cagemanipulationexperimentwasbestestimatedbya
modelthatincludedcagetreatment,thedensityoffruitinguprightsandtheirinteraction
(cage*uprights,p=0.0054),indicatingthattheinfluenceofinsectpollinatorsonyieldis
greaterwhenthedensityoffloweringuprightsishigher.
77
Figure 1
Agriculture within 1km (%)
0 10 20 30 40 50 60 70 800
10
20
30
40
50
60
70
80
90
Woo
ded
habi
tat w
ithin
1km
(%)
78
Figure 2
-1.0
-0.5
0.0
0.5
1.0
CORNPASTURE/HAY
SOYBEANSALFALFA
PASTURE/GRASS
CRANBERRYMARSH
DEVELOPED/OPEN SPACE
OPENWATER
HERBACEOUSGRASSLAND
HERBACEOUSWETLAND
WOODEDWETLAND
MIXED FOREST
EVERGREENFOREST
DECIDUOUSFOREST
-1.0 -0.5 0.0 0.5 1.0
Component 1 (26.1 %)
Com
pone
nt 2
(15
.8 %
)79
Cranberry
Honey beehives/acre
mEIQ
PC1
PC2
Native BeeAbundance
Native BeeRichness
Cranberry yield(marsh avg.)
Figure 3
Removed path
0.65***
0.18
0.61***
0.52***
0.32**
R2=0.30
R2=0.45
R2=0.56
0.80***
† p<0.10** p<0.01*** p<0.001
wood edge density
0.25†
-0.11
Land
scap
eLo
cal m
anag
emen
tMarsh Scale Model
80
PC1
Honey beeabundance
PC1
PC2
Native BeeAbundance
Native BeeRichness
Cranberry yield(estimate)
Figure 4
Removed path
† p<0.10* p<0.05** p<0.01*** p<0.001
-0.36†
0.52**
0.66***
R2=0.43
R2=0.46
R2=0.32
0.85**
wood edge density
0.43*
0.25*
Land
scap
eLo
cal
man
agem
ent
Plot Scale Model
81
Figure 5
Open
Caged
0
50
100
150
200
250
300
350
400
450
Yie
ld e
stim
ate
(bbl
/acr
e)
-5 -4 -3 -2 -1 0 321
Principle Component 2 (PC2)
82
Fruiting uprights/36 sq. in.
0 10 20 30 40 50 60 70 80
Yiel
d es
timat
e (b
bl/a
cre)
0
100
200
300
400
500
600
Figure 6
Open
Caged
83
Appendix3.1:ListofbeespeciesandtheirabundancescollectedinWisconsincranberrybetween2008and2011.SpecimenidentificationwasverifiedbyMikeArduser,JohnAscher,andJasonGibbs.
Species 2008 2010 2011 TOTAL
Agapostemon sericeus 33 80 38 151
Agapostemon splendens 6 38 54 98
Agapostemon texanus 173 131 121 425
Agapostemon texanus or splendens 1 1
Agapostemon virescens 19 117 125 261
Andrena (Ptilandrena) erigeniae 9 9
Andrena (Trachandrena) sp. 1 2 2
Andrena alleghaniensis 3 1 1 5
Andrena arabis 1 1 2
Andrena barbilabris 5 1 6
Andrena carlini 20 13 1 34
Andrena carolina 2 2
Andrena ceanothi 7 1 8
Andrena chromotricha 1 1
Andrena commoda 1 2 3 6
Andrena crataegi 1 5 2 8
Andrena cressonii 17 16 4 37
Andrena distans 1 2 3
Andrena erythrogaster 1 1
Andrena forbesii 2 2
Andrena fragilis (cf) 1 1
Andrena hippotes 1 1
Andrena imitatrix 1 1
Andrena krigiana 1 1
Andrena milwaukeensis 1 1
Andrena miranda 2 4 4 10
Andrena miserabilis 4 1 1 6
Andrena nasonii 1 4 5
Andrena nivalis 11 164 28 203
Andrena nr dunningi 1 1
Andrena perplexa 1 1 4 6
Andrena platyparia 2 2
Andrena rufosignata 3 3
Andrena rugosa 1 1
Andrena sigmundi 2 2
Andrena vicina 11 9 3 23
Andrena violae 1 1
84
Andrena wilkella 5 8 2 15
Andrena wilmattae 1 1
Andrena w‐scripta 2 2
Anthophora terminalis 5 10 15
Augochlora pura 2 1 1 4
Augochlorella aurata 14 65 79 158
Augochloropsis fulgida 13 13
Augochloropsis metallica 5 9 14
Augochloropsis sumptuosa 1 1 2
Bombus auricomus 1 1
Bombus bimaculatus 1 24 13 38
Bombus borealis 4 12 8 24
Bombus fernalde 2 2
Bombus fervidus 1 6 11 18
Bombus griseocollis 5 25 18 48
Bombus impatiens 5 90 35 130
Bombus pensylvanicus 1 1
Bombus rufocinctus 5 1 6
Bombus sandersoni 6 4 10
Bombus sp. 1 1
Bombus ternarius 5 56 79 140
Bombus terricola 17 2 19
Bombus vagans 2 13 18 33
Calliopsis andreniformis 3 3
Ceratina calcarata 1 1
Ceratina calcarata or dupla or micmaqi 1 1
Ceratina dupla 1 1
Ceratina dupla or calcarata 8 8
Ceratina sp. 3 3
Coelioxys immaculata* 2 1 3
Colletes inaequalis 20 20
Dufourea monardae 3 3
Eucera atriventris 1 1
Eucera hamata 26 60 86
Halictus confusus 73 66 49 188
Halictus ligatus 6 71 40 117
Halictus parallelus 19 8 27
Halictus rubicundus 8 11 8 27
Hoplitis pilosifrons 2 2 1 5
Hoplitis producta 4 1 5
Hoplitis spoliata 3 1 4
85
Hoplitis truncata 1 1
Hylaeus (Prosopis) nr modestus 2 2
Hylaeus affinis/modestus 2 2
Hylaeus basalis 1 1
Hylaeus mesillae 2 1 3 6
Hylaeus modestus 1 1
Lasioglossum (Dialictus) lineatulum 17 85 30 132
Lasioglossum (Dialictus) MA WI sp. A 128 128
Lasioglossum (Dialictus) MA WI sp. B 99 99
Lasioglossum (Dialictus) MA WI sp. D 12 12
Lasioglossum (Dialictus) MA WI sp. E 1 1
Lasioglossum (Dialictus) MA WI sp. F 1 1
Lasioglossum (Dialictus) MA WI sp. G 1 1
Lasioglossum (Dialictus) MA WI sp. H 1 1
Lasioglossum (Dialictus) nr MA WI sp. B 3 3
Lasioglossum (Dialictus) perpunctatum 5 9 3 17
Lasioglossum (Dialictus) pilosum 66 215 86 367
Lasioglossum (Dialictus) rohweri 53 53
Lasioglossum (Dialictus) sp. 3 3
Lasioglossum (Dialictus) viereckii 5 2 7
Lasioglossum (Dialictus) zephyrum 40 46 4 90
Lasioglossum ?unknown 1 1
Lasioglossum acuminatum 70 131 33 234
Lasioglossum admirandum 392 131 523
Lasioglossum albipenne 3 10 5 18
Lasioglossum anomalum 1 1
Lasioglossum athabascense 1 13 1 15
Lasioglossum bruneri 12 22 13 47
Lasioglossum cf. floridanum 1 1
Lasioglossum cinctipes 7 13 6 26
Lasioglossum coreopsis 1 1
Lasioglossum coriaceum 32 149 119 300
Lasioglossum cressonii 4 11 12 27
Lasioglossum ellisiae 1 2 3
Lasioglossum ephialtum 18 18
Lasioglossum fedorense 32 4 36
Lasioglossum floridanum 1 1
Lasioglossum forbesii 1 1 2
Lasioglossum forbesii or paraforbesii 1 1
Lasioglossum foxii 1 1
Lasioglossum heterognathum 1 1
86
Lasioglossum laevissimum 2 2
Lasioglossum leucocomum 127 73 200
Lasioglossum leucozonium 30 265 66 361
Lasioglossum lusorium* 1 1
Lasioglossum lustrans 1 29 3 33
Lasioglossum lustrans (3 cells) 2 2
Lasioglossum michiganense 1 1
Lasioglossum mitchelli 3 1 4
Lasioglossum nelumbonis 6 40 46 92
Lasioglossum nigroviride 1 2 3
Lasioglossum novascotiae 1 1
Lasioglossum nymphaearum 31 85 116
Lasioglossum oblongum 7 1 8
Lasioglossum oceanicum 121 121
Lasioglossum paraforbesii 5 47 9 61
Lasioglossum pectorale 16 52 54 122
Lasioglossum pictum 1 1
Lasioglossum pruinosum 1 1
Lasioglossum smilacina 4 4
Lasioglossum sp. 1 1
Lasioglossum subviridatum 16 1 17
Lasioglossum swenki 1 17 34 52
Lasioglossum taylorae 2 2
Lasioglossum timothyi 2 1 3
Lasioglossum versatum 95 25 120
Lasioglossum viridatum 8 11 19
Lasioglossum weemsi 2 2
Lasioglossum zonulum 30 105 102 237
Macropis nuda 1 1 1 3
Megachile addenda* 11 21 7 39
Megachile brevis 2 2 4
Megachile gemula 7 6 1 14
Megachile latimanus 5 17 22
Megachile latimus 21 21
Megachile relativa 1 1 2
Melissodes (Eumelissodes) sp. 9 9
Melissodes agilis 10 10
Melissodes bimaculata 15 8 23
Melissodes desponsa 1 11 19 31
Melissodes druriella 20 20
Melissodes subillata 3 3
87
Melissodes trinodis 6 6
Melissodes agilis or trinodis 14 14
Nomada sp. 3 3
Nomada sp. 1 1 1
Osmia albiventris 1 1
Osmia atriventris 1 2 3
Osmia distincta 9 7 2 18
Osmia georgica 2 2
Osmia inspurgens 6 6
Osmia lignaria 1 1 2
Osmia pumila 21 5 1 27
Osmia simillima 1 3 4
Osmia tersula 2 1 3
Osmia virga 1 1
Peponapis pruinosa 6 8 14
Perdita maculigera 1 1
Pseudopanurgus helianth 1 1
Psithyrus sp. 4 4
Sphecodes atlantis 1 1 2
Sphecodes confertus 2 2
Sphecodes coromus 1 1
Sphecodes davisii 2 1 3
Sphecodes dichrous 6 6
Sphecodes levis 1 1
Sphecodes mandibularis 1 1 2
Sphecodes rannacali 2 2
Sphecodes solonis 1 1
Stelis (Stelis) labiata 1 1
Stelis foederalis? 1 1
Tripeolis sp. 1 1
Xenoglossa kansensis* 11 5 16
*indicates new record for the state of Wisconsin
88
1
CHAPTER4
Understandingthebarrierstoimplementationofon‐farmhabitatconservationof
nativebeesinWisconsincranberry
Author:HannahR.Gaines
Co‐author(s):ClaudioGratton
Forsubmissionto:JournalofInsectConservation
Abstract:Theexpansionofmodernagriculturehasledtothelossandfragmentationof
naturalhabitat,resultinginaglobaldeclineinbiodiversity.Onewaytomitigatethe
negativeeffectsofagricultureonbiodiversityisthroughtheconservationofnaturalhabitat
withinagriculturallandscapes.Inmanycountries,farmerscanreceivecost‐sharefunding
tocreatethesehabitats.Thesecost‐shareprogramscouldbeespeciallybeneficialfor
farmerswithpollinator‐dependentcrops,sincelocalhabitatcanprotectandenhance
nativebeesandthepollinationservicestheyprovide.Wisconsincranberrygrowersshould
beanespeciallylikelygrouptoparticipateinsuchprogramsduetotheirdemonstrated
commitmenttoenvironmentalstewardship,dependenceonpollinators,andavailabilityof
non‐croplandontheirfarms.Unfortunately,theiradoptionhasbeenverylow.Inorderto
increaseparticipation,wehavetounderstandwhygrowersarenotparticipating.
Therefore,theobjectiveofthisstudywastounderstandthebarriersthatpreventthe
implementationoffederallyfundedon‐farmconservationprogramsfornativebeeson
89
2
Wisconsincranberrymarshes.Wesentapapersurveyregardingfarmingpractices,
pollinators,conservation,anddemographicstothemailinglistoftheWisconsinState
CranberryGrowersAssociation(n=250).Wefoundthatwhileonly10%ofgrowerswere
awareofthefederalpollinatorcost‐shareprograms,onethirdofthemweremanaging
habitatforpollinatorswithoutfederalaidand50%wereinterestedinparticipatinginthe
program.Wealsofoundthat57%ofgrowersmanagehabitatforwildlife,althoughnone
receivecost‐sharefundingtodosoeither.Basedonresponsestoadditionalquestions,
participationincost‐shareprogramscouldbenefitfromoutreachactivitiesthatpromote
thecost‐shareprograms,areductionofbureaucratichurdlestoparticipate,andtechnical
supporttogrowersonhowtomanagehabitatforwildbees.
Keywords(4‐6):conservation,pollinator,agriculture,EQIP,classificationtreeanalysis
1.0Introduction
Theexpansionofmodernagriculturehasledtothedestructionandfragmentationof
habitat,resultinginamassivelossofbiodiversity(Bentonetal.2003,Foleyetal.2005).
Someofthisbiodiversityalsoprovidesvaluableecosystemservicestohumansandasthis
biodiversityislost,sotooaretheservicesitprovides.Insectsareonegroupoforganisms
thatarenegativelyaffectedbytheexpansionofagricultureandalsoprovidevaluable
servicestohumansincludingpestsuppressionandpollination(LoseyandVaughan2006,
Isaacsetal.2009).Byprovidinghabitatwithintheagriculturallandscape,someofthe
negativeeffectsofagricultureonbiodiversitymaybealleviated,allowingbeneficial
organismsandtheservicestheyprovidetopersist.
90
3
On‐farmconservationpracticestoestablishnon‐crophabitatwithinthefarming
landscaperepresentsoneapproachtolessenthenegativeeffectsofagriculturalexpansion.
Previousstudieshavedocumentedthevalueofsuchnon‐crophabitattobeneficialinsects
andtheservicestheyprovide(Merckxetal.2009,Scheperetal.2013,Morandinand
Kremen2013).BecauseofthelargeamountoftheEarth’slandareathatisusedfor
agriculture(Foleyetal.2011),incorporatingon‐farmconservationpracticeshasthe
potentialtomakeasubstantialcontributiontoprotectingbiodiversity.Becauseofthe
potentialvalueofon‐farmconservationtobiodiversity,manygovernmentsaroundthe
worldhavedevelopedconservationincentivesprogramstoencouragefarmerstomanage
non‐crophabitatinanefforttoprotectbiodiversity(Wadeetal.2008).Throughthese
programsfarmersreceivefinancialsupporttotakecroplandoutofproductionandinstead,
manageitforbiodiversitybyplantingnon‐crophabitat.
IntheUnitedStates,conservationincentivesprogramsexisttoencouragefarmersto
stopfarminghighlyerodibleland(e.g.,steephills),protectwaterwaysbyinstallingriparian
buffers,andmorerecently,providinghabitatforpollinators(VaughanandSkinner2008).
Farmerscanreceivecost‐sharefundingforpollinatorhabitatthroughanumberofdifferent
USDAconservationprogramssuchastheConservationReserveProgram(CRP)andthe
EnvironmentalQualityIncentivesProgram(EQIP)andpracticeswithinthoseprograms
(e.g.,fieldborders,conservationcover,hedgerows).Inaddition,asofthe2008USDAFarm
Bill,USDAconservationistsmustnowprioritizepollinatorsandpollinatorhabitatwhen
determiningpaymentsorreviewinganddevelopingconservationpractices(XercesSociety
2009).
91
4
Withtherecentdeclineinbothmanagedhoneybeesandwild,nativebees(Biesmeijer
etal.2006,Ellisetal.2010,Pottsetal.2010,Cameronetal.2011),theseprogramscouldbe
especiallybeneficialtofarmersgrowingpollinator‐dependentcrops,suchasWisconsin
cranberrygrowers.Cranberryisapollinator‐dependentcrop(DelaplaneandMayer2000)
forwhichgrowersspendthousandsofdollarseachyearonhoneybeehiverentals.
Additionally,Wisconsincranberrygrowersoftenownsignificantamountsofnon‐crop
supportlandthatcouldbeusedforpollinatorhabitat(ColquhounandJohnson2010).
Furthermore,theyhaveademonstratedappreciationforenvironmentalstewardshipwith
theirhighadoptionrateofintegratedpestmanagement(IPM)practices(DanMahr,pers.
comm.).Forexample,almostallofthesegrowershirecropscoutsinordertodetermine
whentosprayforpests,applyreducedriskpesticidesontheirmarshes,andusealternative
culturalpracticestocontrolpests(e.g.,flooding).Bycreatingpollinatorhabitatontheir
farms,cranberrygrowerscouldreceiveaneconomicallysignificantreturnontheirefforts
intheformofcroppollinationservicesprovidedbynativebees.
Despitethepossiblebenefitstogrowers,participationbyWisconsincranberry
growersinUSDAprogramstocreatepollinatorhabitatisnearlynon‐existent(JulieAmmel,
pers.comm.).Inordertoconvincegrowerstoparticipateintheseprograms,wehaveto
understandwhytheyaren’tcurrentlyparticipating.Somepossiblefactorsmayincludea
perceivedpestproblemfromconservationhabitat,lackofawarenessoftheprograms,lack
oftechnicalknowledgeorsupporttoimplementtheproject,oraperceptionthatthe
programsrequiretoomuchtime,space,ormoney.Byunderstandingwhygrowersarenot
92
5
participatingandwhichfactorsaremostimportantintheirdecisionmakingprocess,we
canmoreeffectivelyaddresstherootoftheproblem.
Inordertodeterminetheobstaclespreventinggrowersfromparticipatinginon‐
farmconservationprogramsfornativebeesandprovidepracticalsolutionsbasedon
growerinput,weconductedastate‐widegrowersurvey.Ourspecificobjectiveswereto
determinewhichfactorsaremostimportantinpredictingwhetherWisconsincranberry
growersare(1)activelymanagingtheirfarmstoencouragenativebeepollinators,(2)
activelymanaginghabitattoprotectotherwildlife,and(3)participatinginon‐farm
conservationincentiveprograms.Theresultsofthisstudywillinformtheoutreachand
educationeffortsofuniversityandagencypersonnelandidentifywaystoimprovefarmer
participationtosupportconservationpractices.
2.0 Methods
2.1Growersurvey
Wecreateda50‐questionwrittensurveyregardingcurrentfarmingpractices,
pollination,on‐farmconservation,anddemographics(Appendix4.1).Thesurveywas
reviewedbyseveralcranberrygrowers,conservationprofessionals,andextension
professorstoensureclarityandcompleteness.Questionswereadded,removed,oredited
basedonfeedbackfromthesegroups.ThesurveywasthensenttotheUniversityof
WisconsinSurveyCenterforreviewofquestionstructuretoensureinterpretability.The
finalsurveywasmailedtoeverycranberrygrowerontheWisconsinStateCranberry
GrowersAssociation(WSCGA)mailinglist(n=250)inJuneof2011.Duetoconfidentiality
93
6
policies,thelistwasnotsharedwiththeauthors,ratherthesurveysweredeliveredtothe
WSCGAofficeandtheofficestaffaddressedtheenvelopes.Acoverletterexplainingthe
purposeofthesurveyandastampedaddressedreturnenvelopewereincludedinthe
mailing.Areminderpost‐cardwassenttothesamerecipientsoneweekfollowingthe
initialmailing.
2.2Statisticalanalysis
Theresultsofthissurveyweresummarizedusingbothdescriptivestatisticsand
classificationandregressiontree(CART)analysis(R2.15.1,libraryrpart).Descriptive
statisticswereusedtodeterminethefrequencyofresponsestosurveyquestions.CART
analysiswasusedtodeterminewhichfactorsweremostimportantinpredictingthe
responseofgrowerstospecificquestionsregardingactivemanagementofhabitatforwild,
nativebees(Q22);alterationofmanagementtoprotectwild,nativebees(Q25);interestin
managingfornativebees(Q27);managementofhabitatforwildlife(Q29);andinterestin
participatinginaconservationincentivesprogramtocreatehabitatforwild,nativebees
(Q39).ACARTmodelisanon‐parametricmethodusedinordertodeterminewhich
factorsaremostimportantinpredictingvaluesofanotherfactor(De’athandFabricius
2000,ChangandWang2006).Forfactorswithdiscretevariable(e.g.,yes/noresponses)
theclassificationtreeisusedratherthanaregressiontree.Allothersurveyquestions
excepttheonedesignatedasthepredictedvariablewereincludedaspossiblepredictorsin
theCARTmodel.Theresultsofthisanalysishelpusidentifyfactorswhichclassifygrowers
intogroupswithsimilarresponsestoagivenquestion.Forexample,ifinterestlevelsin
conservationofwild,nativebeesisgroupedbycountybutparticipationislowinacounty
94
7
whereinterestishigh,extensionactivitiescanfocusonprovidinginformationand
technicalsupporttogrowersinthatcounty.Byunderstandingwhichfactorsbestpredict
currentmanagementorinterestbygrowers,futureextensionandoutreacheffortscanbe
bettertargetedtoaddressboththedesiredconservationoutcomesandthefactors
contributingtowhythegrowersdecidetoparticipateinon‐farmconservation.
3. Results
3.1 Responserateanddemographics
Theresponseratetooursurveywashigh(49%ofsurveysreturned,n=122).Survey
responsescamefromgrowersrepresentingabroadrangeofgeographiclocationsand
demographics.Growersrespondedfrom13countieswhichrepresentedboththecentral
andnortherngrowingregionsinthestate.Forty‐fourpercentofresponseswerefrom
WoodCountywhichhasthehighestdensityofcranberrygrowersinthestate.Theaverage
ageofrespondentswas45‐54yearsoldandtheaveragelengthoftimetheyhadbeen
growingcranberrieswas24.5years(±1.15SD,median23,range4‐77).Theaverage
propertysizewas1,115acres(±159SD,median450,range11‐11,000)includingon
average50‐149acresincranberryproduction.Inaddition,84%ofrespondentsindicated
thatgrowingcranberriesistheirprimarysourceofincome.
3.2 Activemanagementforbees
Despitealackofparticipationinpollinatorhabitatcost‐shareprogramsamong
cranberrygrowers(0%),33%ofgrowersactivelymanagehabitat(e.g.,plantflowering
shrubsandtrees,provideartificialnestboxes,providebrushpilesandmattressesfor
95
8
nesting)forwildbeeswithoutoutsideassistancefromaUSDAprogram.Additionally,30%
reportedthattheyhavealteredtheirmanagementinsomeotherway(e.g.,reduced‐risk
pesticides,timingofspray,delayedmowingofdikes)toencouragewildbeesontheir
property.Thethreefactorsthatweremostoftenselectedasveryorextremelyimportant
ingrowers’decisiontomanageforbeeswere(1)theimportanceofpollinationfor
cranberries,(2)environmentalstewardship,and(3)knowledgeaboutpollinatorhabitat
(fig.1).Ofthegrowerswhodonotmanageforbees,themostimportantfactorsinfluencing
theirdecisionwere(1)theimportanceofpollinationforcranberries,(2)theavailabilityof
technicalsupport,and(3)timecommitment.TheresultsofourCARTanalysisshowedthat
themostimportantpredictorofwhethergrowersmanagehabitatforbeeswaswhether
theyhadalsoalteredtheirmanagementinotherwaystoencouragewildbees(fig.2).Of
therespondentswhomanagehabitatforbees,56%ofthemhavealsoalteredtheir
managementinsomeotherway.Thenextbestpredictorvariablewaswhetherornotthe
growermanageshabitatforotherwildlife.Ofthosewhomanagehabitatforbees,51%also
managehabitatforwildlife.Regardlessofwhetherornotthegrowersmanageforwild
bees,87%respondedthatbeesareveryorextremelyimportanttocranberrypollination
and89%reportedthattheycurrentlyrenthoneybeesforpollination.
Thebestpredictorofwhethergrowersaltertheirmanagementinanywayother
thanhabitatmanagementtoprotectbees(e.g.,timingofspray,useofreducedtoxicity
pesticides)waswhethertheymanagehabitatforbees(fig.3).Thenextbestpredictorwas
howimportanttheyrated“environmentalstewardship”.Forty‐onepercentofgrowers
whoaltertheirmanagementforwildbeesratedenvironmentalstewardshipas
96
9
very/extremelyimportantintheirdecisiontomanageforbees.Theseresultsshowthat
growerswhomanagehabitatforwildlifeandvalueenvironmentalstewardshiparemore
likelytomanagehabitatforbeesoraltertheirmanagementinotherwaystoprotectbees.
3.3 Activemanagementofwildlifehabitat
Tounderstandfurtherhowgrowersperceiveon‐farmconservationingeneral,we
askedaseriesofquestionsregardingmanagementofhabitattoprotectwildlife.Wefound
that57%ofgrowersactivelymanagehabitatforthespecificgoalofprotectingwildlife,
althoughnonearereceivingcost‐sharefundingtodoso.Thefactorsthatweremostoften
selectedasveryorextremelyimportantininfluencingagrower’sdecisiontomanage
wildlifehabitatwere(1)knowledgeaboutwildlifehabitat,(2)knowledgeofwildlife,
environmentalstewardship,andrecreationsuchashuntingorhikingequallytiedfor
second,and(3)timecommitment(figure4).Ofthegrowerswhodonotmanagefor
wildlifehabitat,themostimportantfactorsintheirdecisionmakingwas(1)time
commitment,(2)financialcommitment,and(3)perceivedpestproblemsfromwildlife
habitat.TheCARTanalysisfurtherindicatedthatthemostimportantpredictorvariablefor
whetherornotgrowersmanagehabitattoprotectwildlifewashowimportanttheyrated
recreation(fig.5).Ofthosewhomanagehabitat,55%alsoratedrecreationas
very/extremelyimportant.Ofthosewhodidnotraterecreationasimportant,thenext
mostimportantpredictorvariablewasthenumberofyearsthegrowerhadbeengrowing
cranberries.Thirty‐twopercentofgrowerswhomanagehabitatforwildlifehavebeen
growingcranberriesformorethan28.5yearswhereasonly13%havebeengrowing
cranberriesforlessthan28.5years.
97
10
Theseresultssuggestthatgrowerswhomanageforwildlifehaveknowledgeofwildlife
andtheirhabitatrequirements,astrongenvironmentalstewardshipethicandinterestin
outdoorrecreationsuchashuntingorhiking.Whiletheyacknowledgethetime
commitmentasanimportantfactor,theotherfactorsoutweighthetimecommitment.
Theseresultsfurthersuggestthattheperceivedtimeandfinancialcommitment,inaddition
topossiblepestproblems,causesomegrowerstonotmanagewildlifehabitatontheir
farm.Growerswhovaluerecreationsuchashuntingorhikingaremostlikelytoalso
managewildlifehabitatbutofthosewhodon’tvaluerecreationashighly,thegrowerswho
havebeenfarmingforalongerperiodoftimearemorelikelytomanagewildlifehabitat.
3.4Participationinconservationincentivesprograms
Thirty‐onepercentofrespondentsindicatedthattheycurrentlyorhavepreviously
participatedinUSDA‐sponsoredconservationincentivesprogramsincludingEQIP(22%),
CRP(8%),theConservationReserveEnhancementProgram(CREP,2%),andtheWildlife
HabitatIncentivesProgram(WHIP,2%).Growerswere,however,splitonwhetherthey
wouldparticipateinconservationincentivesprogramsinthefuture,andmorewere
unlikelytoparticipateinaUSDA‐sponsoredprogram(36%)thananon‐USDA‐sponsored
program(26%).Wealsofoundthatthemostimportantfactorsdeterminingwhethera
growerparticipatedinaconservationincentivesprogramwere(1)theamountof
paperwork,(2)timecommitment,and(3)financialcommitment(figure6).Ofthegrowers
whodidnotparticipateinconservationincentiveprograms,themostimportantfactors
influencingtheirdecisionwere(1)amountofpaperwork,(2)timecommitmentand
environmentalstewardshipequally,and(3)awarenessoftheprograms.
98
11
Whilenoneoftherespondentsparticipateinanincentivesprogramforpollinator
habitat,andonly10%areawarethatsuchprogramsexist,50%ofgrowerswouldbe
interestedinparticipating.ACARTanalysisindicatedthatthemostimportantpredictor
variableforwhetherornotgrowersareinterestedinparticipatinginacost‐shareprogram
toinstallpollinatorhabitatontheirpropertywastheirinterestinmanaginghabitatfor
beesinthefuture(fig.7).Thenextbestpredictorvariablewaswhetherthegrower
respondedpositivelytotheusefulnessofaninformationalpamphletaboutUSDA
conservationincentivesprogramsforwildbees.Thefinalpredictorvariablewashow
importanttherespondentratedbeautyorlandscapingintheirdecisionprocess.Ofthe2%
ofrespondentswhoratedbeautyandlandscapingas“extremelyimportant”,nonewere
interestedinmanaginghabitatforbeesinthefuture,possiblybecauseofaperceptionthat
thishabitatwouldbeunattractive.
Theseresultssuggestthattheamountofpaperworkandtimecommitmentwere
enoughtodiscouragemanygrowersfromparticipatingintheprogramsdespitethe
importanceofenvironmentalstewardshiptothesegrowers.
3.5 Sourcesofmanagementinformationforgrowers
Inordertoaddressthefindingsabove,extensionandoutreacheffortsmustreach
thegrowers.Byunderstandingwhichsourcesofinformationgrowers’useorwanttouse,
extensioneffortscanbemoreeffectivebycommunicatingtogrowersthroughthese
channels.Currently,themostcommonsourcesofinformationarepapernewsletters
(90%),cropscouts(82%),andtheannualmeetingoftheWSCGA(CranberrySchool,81%).
Inthefuture,growersareinterestedinusingemaillistservs(38%)inadditiontopaper
99
12
newsletters(36%),andcropscouts(35%).Wealsocollectedinformationonsourcesthe
growersdonotwanttouseinthefuturewiththetopthreebeingsocialnetworkingsites
(85%),automatictextmessagealerts(61%),andemaillistservs(44%).Thisinformationis
alsousefulsothatextensionandoutreachisnotmisdirectedtothesesourcesto
disseminateinformation.
4.0Discussion
Despitetheirdemonstratedcommitmenttoenvironmentalstewardship,participation
inpollinatorhabitatcost‐shareprogramsbyWisconsincranberrygrowersisverylow.
ThroughawrittensurveyofWisconsincranberrygrowers,wefoundthatawarenessof
pollinatorhabitatincentivesprogramsisalsoverylow.Despitethelackofparticipationin
formalconservationprograms,athirdofgrowersarecurrentlymanaginghabitatfor
pollinatorsanyway.Thesesamegrowerswerealsomorelikelytomanagehabitatfor
wildlife.Growerswhowerenotmanaginghabitatforbeesorwildlifeweredeterredbya
lackoftechnicalsupport,andtheperceivedtimeandfinancialcommitments.Thissuggests
thatoutreachandextensioneffortsshouldfocusbothonpromotingpollinatorhabitat
incentivesprogramsaswellasprovidinginformationandtechnicalsupporttogrowers
interestedincreatingpollinatorhabitat.
Themoststrikingfindinginoursurveywasthat,despitealackofparticipationinthe
cost‐shareprogramsforpollinatorhabitat,athirdofWisconsincranberrygrowersare
managinghabitatforwildbeesanyway.Thesemanagementactivitiesincludeddelaying
mowingofnon‐cropareasuntilcranberrywasinbloom,plantingflowersandtreesthat
100
13
bloomthroughouttheseason,andprovidingnestinghabitatsuchasbrushpilesandbare
soil.Morethanhalfofthegrowerssurveyedalsomanagehabitatforwildlife,yetnone
receivedcost‐sharefundingtodoso.Thegrowerswhomanagedhabitatforwildlifewere
morelikelytovalueoutdoorrecreationsuchashuntingorhadbeengrowingcranberries
formanyyears.Wisconsincranberrygrowersappeartobemotivatedmorebytheir
commitmenttoenvironmentalstewardshiporanappreciationoftheoutdoorsthanby
financialincentivesdespitethecostofhabitatmanagement.Similartoourresults,Banack
andHvenegaard(2010)foundthatfarmerswhoparticipatedinbiodiversity‐friendly
farmingpracticesweremostcommonlymotivatedbyamoralobligationtocareforthe
environmentovereconomicfactors.Thegoalofon‐farmconservationistoprovidehabitat
inagriculturallandscapestoprotectandenhancebiodiversity.Whilefinancialincentives
programsareonewaytomotivategrowerstoprovidethishabitat,appealingtotheirsense
ofenvironmentalstewardshipmaybeanequallystrongandeffectivewaytoincreasethe
conservationofnon‐crophabitatonfarms.
TheminimalparticipationinconservationincentivesprogramsbyWisconsincranberry
growersseemstobeduemainlytoanaversiontobureaucratichurdlesandalackof
awarenessoftheprogramsratherthanlackofinterest.Lemkeetal.(2010)alsofoundthat
farmerswerediscouragedfromparticipatinginconservationpracticesontheirfarms
becauseofthecomplexityofpaperworkinvolved.InMichigan,whereadoptionofthe
pollinatorhabitatprogramshasbeenhigh,thekeytosuccesshasbeenactivepromotionof
theprograms(EricMader,pers.comm.)aswellasgoodfinancialincentivesofuptoa90%
cost‐share(FSA2010).Inotherregionsofthecountryandtheworld,successful
101
14
implementationofconservationprogramshasbeenaresultofengagedconservation
professionalsdevelopingpersonalrelationshipswiththefarmersandpromotingthe
benefitsofconservationthroughworkshopsandone‐on‐oneassistance(Mendhametal.
2007,Lemkeetal.2010,Whittenetal.2012).CurrentlyinWisconsin,minimaltechnical
supportisavailableforgrowersinterestedinparticipatinginon‐farmpollinator
conservationpractices,althoughinterestinpromotingtheseprogramsisgrowingatthe
statelevel(J.Ammel,pers.comm.).
Thisstudyhighlightsfourkeyareasthatcanbeaddressedtoincreaseparticipationin
cost‐shareprogramsaimedatcreatinghabitat.First,thespecificprogramsandpractices
intendedforpollinatorconservationneedtobepromotedinWisconsin.Onlyasmall
percentageofcranberrygrowerswereawareoftheprograms,althoughhalfofthegrowers
wereinterestedinparticipatinginacost‐shareprogramforpollinatorhabitat.Since80‐
90%ofthecranberrygrowersgettheirmanagementinformationfromindustry
newsletters,cropscouts,andCranberrySchool,thesewouldbeappropriateoutlets
throughwhichtopromoteconservationprograms.
Second,thepaperworkrequiredbythegrowerstoparticipateintheseprogramsneeds
tobereduced.Loweringthebureaucratichurdlestoparticipationwouldgreatlyreduce
thetimecommitmentandmakeparticipationeasier.InordertoparticipateinUSDA‐
sponsoredconservationincentivesprograms,growershavetogothroughalong,
bureaucraticprocessofdeterminingeligibilitythroughtheFarmServiceAgency(FSA),
developingconservationplans,selectingappropriateprogramsandpractices,and
submittingproposalstotheFSAthroughtheNRCS(USDA2013).Ontopofallofthe
102
15
paperworkalreadyrequiredbythegrowersregardingapplicationofpesticides,wateruse,
andpermitsformarshrenovationandexpansion,additionalpaperworkcandiscourage
growersfromparticipatingincost‐shareprograms(Lemkeetal.2010).Streamliningthe
processbyreducingtheamountofpaperworkcouldhaveabigimpactonparticipation.
Third,creatingpollinatorhabitatmayseemoverwhelmingbecauseoftheknowledge
requiredforestablishingandmaintainingnativeplants.Thereisaneedfortechnical
supporttodesignandimplementconservationplansthatincludepollinatorhabitat
(Traoréetal.1998,WhiteandSelfa2013),butwithoutapersonavailabletodirectthe
growers,answertheirquestions,andprovidepracticalon‐the‐groundsolutions,few
growerswillattempttoinstallpollinatorhabitat.Severalinformationalpamphletsandjob
sheetsexisttohelpgrowersplantheirpollinatorplanting(VaughanandMader2008,
VaughanandSkinner2008,Vaughanetal.2012),butwithmanyfarmerswithouthigh
speedinternet(VolenbergandJensen2010)accessingtheseresourcesisdifficult.
Andfinally,NRCSprogramadministratorsshouldestablishapeer‐mentoringprogram
toconnectgrowerswhoarecurrentlymanagingpollinatorhabitatwithgrowerswhoare
interestedinmanagingpollinatorhabitat.Sixty‐fourpercentofgrowersindicatedthatthey
getinformationregardingmanagementpracticesfromtheirneighborsandfriends,
suggestingthatpeermentoringcouldbeaneffectivewaytoincreaseparticipation.
Providinggrowerswithanexampleofanestablishedpollinatorhabitatplantingand
connectingthemwithagrowerwhohasgonethroughtheprocesshasthepotentialto
increasethesuccessofinterestedgrowerswhilereducingthedemandputonagency
personaltoaddresseveryconcernfromnewparticipants.Althoughitisbeyondthescope
103
16
ofthisresearchproject,furtherthoughtanddevelopmentwillneedtobegiventocreate
incentivesforgrowersactingasmentorstobepartoftheprogram.
Inordertomakeiteasierforgrowerstomanagehabitatforpollinatorsandparticipate
incost‐shareprogramstodoso,weneedtopromotetheavailableprograms,simplifythe
process,providedirection,andencouragepeermentorship.Thesestepshavethepotential
togreatlyincreaseparticipationamongWisconsincranberrygrowersaswellasother
pollinator‐dependentfruitcrops.Withoutthesesteps,theincentiveprogramsfor
pollinatorhabitatwilllikelycontinuestrugglingtogainparticipants.Byfollowingthesteps
outlinedabove,wecanmakeiteasierforfarmerstoparticipateandshowthaton‐farm
conservationisaprioritytotheUSDA.
Acknowledgements
Wethankallofthecranberrygrowerswhorespondedtooursurvey.Wealsothank
JulieDollandtheUniversityofWisconsinSurveyCenterforhelpwithsurveydevelopment,
JaneAndersonformailingthesurveys,TomEbert,JohnO’Day,JayneSojka,MaceVaughan,
andEricMaderforcommentingonearlydraftsofthesurvey,andDavidDuncanfor
assistantswiththestatistics.ThisresearchwasfundedbyaNorthCentralRegion
SustainableAgricultureResearchandEducation(SARE)GraduateStudentGranttoH.R.G.
Literaturecited
Banack,S.A.,andG.T.Hvenegaard.2010.Motivationsoflandownerstoengageinbiodiversity‐friendlyfarmingpracticesinAlberta’scentralparklandregion.HumanDimensionsofWildlife15:67–69.
104
17
Benton,T.G.,J.A.Vickery,andJ.D.Wilson.2003.Farmlandbiodiversity:ishabitatheterogeneitythekey?TrendsinEcology&Evolution18:182–188.
Biesmeijer,J.C.,S.P.M.Roberts,M.Reemer,R.Ohlemüller,M.Edwards,T.Peeters,A.P.
Schaffers,S.G.Potts,R.Kleukers,C.D.Thomas,J.Settele,andW.E.Kunin.2006.Paralleldeclinesinpollinatorsandinsect‐pollinatedplantsinBritainandtheNetherlands.Science313:351–354.
Cameron,S.A.,J.D.Lozier,J.P.Strange,J.B.Koch,N.Cordes,L.F.Solter,andT.L.Griswold.
2011.PatternsofwidespreaddeclineinNorthAmericanbumblebees.ProceedingsoftheNationalAcademyofSciences108:662–667.
Chang,L.‐Y.,andH.‐W.Wang.2006.Analysisoftrafficinjuryseverity:Anapplicationof
non‐parametricclassificationtreetechniques.AccidentAnalysis&Prevention38:1019–1027.
Colquhoun,J.,andH.Johnson.2010.Sustainablecranberryproductionforavibrantfuture:
TheWisconsinexperience.UniversityofWisconsin‐Extension.De’ath,G.,andK.E.Fabricius.2000.Classificationandregressiontrees:Apowerfulyet
simpletechniqueforecologicaldataanalysis.Ecology81:3178–3192.Delaplane,K.S.,andD.F.Mayer.2000.Croppollinationbybees.CABIPublishing,New
York,NY.Ellis,J.D.,J.D.Evans,andJ.Pettis.2010.Colonylosses,managedcolonypopulationdecline,
andColonyCollapseDisorderintheUnitedStates.JournalofApiculturalResearch49:134–136.
Foley,J.A.,R.DeFries,G.P.Asner,C.Barford,G.Bonan,S.R.Carpenter,F.S.Chapin,M.T.
Coe,G.C.Daily,H.K.Gibbs,andothers.2005.Globalconsequencesoflanduse.Science309:570.
Foley,J.A.,N.Ramankutty,K.A.Brauman,E.S.Cassidy,J.S.Gerber,M.Johnston,N.D.
Mueller,C.O’Connell,D.K.Ray,P.C.West,C.Balzer,E.M.Bennett,S.R.Carpenter,J.Hill,C.Monfreda,S.Polasky,J.Rockström,J.Sheehan,S.Siebert,D.Tilman,andD.P.M.Zaks.2011.Solutionsforacultivatedplanet.Nature478:337–342.
FSA.2010.CRP‐SAFENativePollinatorPlanting.FarmServiceAgency,Michigan
DepartmentofAgriculture.Isaacs,R.,J.Tuell,A.Fiedler,M.Gardiner,andD.Landis.2009.Maximizingarthropod‐
mediatedecosystemservicesinagriculturallandscapes:theroleofnativeplants.FrontiersinEcologyandtheEnvironment7:196–203.
105
18
Lemke,A.M.,T.T.Lindenbaum,W.L.Perry,M.E.Herbert,T.H.Tear,andJ.R.Herkert.
2010.EffectsofoutreachontheawarenessandadoptionofconservationpracticesbyfarmersintwoagriculturalwatershedsoftheMackinawRiver,Illinois.JournalofSoilandWaterConservation65:304–315.
Losey,J.E.,andM.Vaughan.2006.Theeconomicvalueofecologicalservicesprovidedby
insects.Bioscience56:311–323.Mendham,E.,J.Millar,andA.Curtis.2007.Landholderparticipationinnativevegetation
managementinirrigationareas.EcologicalManagement&Restoration8:42–48.Merckx,T.,R.E.Feber,P.Riordan,M.C.Townsend,N.A.D.Bourn,M.S.Parsons,andD.W.
Macdonald.2009.Optimizingthebiodiversitygainfromagri‐environmentschemes.Agriculture,Ecosystems&Environment130:177–182.
Morandin,L.A.,andC.Kremen.2013.Hedgerowrestorationpromotespollinator
populationsandexportsnativebeestoadjacentfields.EcologicalApplications23:829–839.
Potts,S.G.,J.C.Biesmeijer,C.Kremen,P.Neumann,O.Schweiger,andW.E.Kunin.2010.
Globalpollinatordeclines:trends,impactsanddrivers.TrendsinEcology&Evolution25:345–353.
Scheper,J.,A.Holzschuh,M.Kuussaari,S.G.Potts,M.Rundlöf,H.G.Smith,andD.Kleijn.
2013.EnvironmentalfactorsdrivingtheeffectivenessofEuropeanagri‐environmentalmeasuresinmitigatingpollinatorloss–ameta‐analysis.EcologyLetters16:912–920.
Traoré,N.,R.Landry,andN.Amara.1998.On‐farmadoptionofconservationpractices:The
roleoffarmandfarmercharacteristics,perceptions,andhealthhazards.LandEconomics74:114–127.
USDA.2013.NaturalResourcesConservationService.www.nrcs.usda.gov.Vaughan,M.,andE.Mader.2008.Wisconsinbiologytechnicalnote8:Pollinatorbiologyand
habitat.USDANRCSWashington,DC.Vaughan,M.,E.Mader,J.Guisse,J.Goldenetz‐Dollar,J.Hopwood,B.Borders,andR.Hirsch.
2012.Conservationcover(327)forpollinators:UpperMidwestinstallationguideandjobsheet.TheXercesSocietyforInvertebrateConservation,Portland,OR.
Vaughan,M.,andM.Skinner.2008.Technicalnoteno.78.UsingFarmBillprogramsfor
pollinatorconservation.USDANRCS.
106
19
Volenberg,D.,andB.Jensen.2010.Cranberrygrowereducation.
www.uwex.edu/ces/ag/teams/fruit/.Wade,M.R.,G.M.Gurr,andS.D.Wratten.2008.Ecologicalrestorationoffarmland:
progressandprospects.PhilosophicalTransactionsoftheRoyalSocietyB‐BiologicalSciences363:831–847.
White,S.S.,andT.Selfa.2013.Shiftinglands:exploringKansasfarmerdecision‐makingin
aneraofclimatechangeandbiofuelsproduction.EnvironmentalManagement51:379–391.
Whitten,S.M.,A.Reeson,J.Windle,andJ.Rolfe.2012.Designingconservationtendersto
supportlandholderparticipation:Aframeworkandcasestudyassessment.EcosystemServices.
XercesSociety.2009.Invertebrateconservationfactsheet:2008FarmBill:Benefitstocrop
pollinators.
107
20
Figurecaptions
Figure1.Percentofgrowersresponding“very”or“extremely”tohowimportanteach
factorisintheirdecisionwhetherornottomanageforwild,nativebees(Q26).
Figure2.Aclassificationtreethatpredictswhethergrowersrespondedpositivelyto
managinghabitatforwildbeesontheirfarm.Cranberrygrowerswhomanagehabitatfor
beesarealsolikelytohavealteredtheirmanagementinsomeotherwaytoprotectbees
andmanagehabitatforwildlife.56%ofgrowerswhomanagehabitatforwildbeeshave
alsoalteredtheirmanagementinsomeotherwaytoprotectbeesasopposedto21%who
havenotalteredtheirmanagementpractices.51%ofgrowerswhomanagehabitatfor
wildbeesalsomanagehabitatforwildlifeasopposedto5%whodonot.
Figure3.Aclassificationtreethatpredictswhethergrowersrespondedpositivelyto
alteringtheirmanagementinsomewaytoprotectwildbeesontheirfarm.59%ofgrowers
whohavealteredtheirmanagementinsomewaytoprotectbeesalsomanagehabitatfor
wildbeesasopposedto41%whodonot.41%ofgrowerswhohavealteredtheir
managementtoprotectbeesalsomanagehabitatforwildlifeasopposedto19%whodo
not.
Figure4.Percentofgrowersresponding“very”or“extremely”tohowimportanteach
factorisintheirdecisionwhetherornottomanagehabitatforwildlife(Q32).
Figure5.Aclassificationtreethatpredictswhethergrowersrespondedpositivelyto
managinghabitatforwildlife.55%ofgrowerswhomanagehabitatforwildliferated
recreationsuchashuntingorhikingasveryorextremelyimportantintheirdecisionto
managehabitatforwildlife.45%ofgrowerswhomanagehabitatforwildliferated
108
21
recreationasonlyslightlyornotatallimportantintheirdecisiontomanagehabitatfor
wildlife.32%ofgrowerswhomanagehabitatforwildlifeanddidnotconsiderrecreation
asveryimportanthadbeengrowingcranberriesformorethan28.5years.Only13%of
growerswhomanagehabitatforwildlifeanddidnotraterecreationasimportanthave
beengrowincranberriesforlessthan28.5years.
Figure6.Percentofgrowersresponding“very”or“extremely”tohowimportanteach
factorisintheirdecisionwhetherornottoparticipateinaconservationincentives
program(Q38).
Figure7.Aclassificationtreethatpredictswhethergrowersareinterestedinparticipating
inacost‐shareprogramforwildbeesinthefuture.98%ofgrowerswhoareinterestedin
thecost‐shareprogramwerealsointerestedinmanaginghabitatforbeesinthefuture.
95%ofgrowerswhowereinterestedinabeehabitatcost‐shareprogramexpressedutility
inaninformationalpamphletaboutUSDAcost‐shareprogramsforwildbees.And95%of
growerswhowereinterestedincost‐shareprogramsforpollinatorsratedbeautyand
landscapingaslessthanextremelyimportantintheirdecisiontomanagehabitatforbees.
109
Financial commitment
Time commitment
Space requirements
Availability of cost share programs
Availability of techincal support
Environmental stewardship
Encouragement of neighbor or friend
Knowledge about wild bees
Knowledge about pollinator habitat
Perceived weed problems from pollinator habitat
Perceived pest problems from pollinator habitat
Beauty or landscaping
Other
0 10 20 30 40 50 60 70 80 90
Importance of pollination for cranberries
Farmers who currently manage habitat for wild beesFarmers who do NOT manage habitat for wild bees
Percent responding very/extremely important
Figure 1110
Growers who actively manage habitat to encourage wild bees on their property (Q22)
YES(56%)
NO(21%)
Altered management for wild bees (e.g., spray at night)
Manages habitat for wildlife
YES(51%)
NO(5%)
Figure 2111
Growers who have altered their management in some way to protect wild bees (Q25)
YES(59%)
NO(41%)
Manage habitat for bees
Manages habitat for wildlife
YES(41%)
NO(19%)
Figure 3112
Financial commitment
Time commitment
Space requirements
Availability of cost share programs
Availability of techincal support
Environmental stewardship
Encouragement of neighbor or friend
Knowledge of wildlife
Knowledge about wildlife habitat
Perceived weed problems from wildlife habitat
Perceived pest problems from wildlife habitat
Beauty or landscaping
Recreation, such as hunting or hiking
Other
0 10 20 30 40 50 60
Farmers who manage habitat for wildlifeFarmers who do NOT manage habitat for wildlife
Percent responding very/extremely important
Figure 4113
Growers who manage habitat for wildlife (Q29)
Very/extremely(55%)
Not/slightly(45%)
How important the grower rates recreation such as hunting or hiking in their decision to manage habitat for wildlife
Years growing cranberry
>28.5 years(32%)
<28.5 years(13%)
Figure 5114
0 10 20 30 40 50 60
Financial commitment
Time commitment
Availability of cost share programs
Availability of techincal support
Environmental stewardship
Encouragement of neighbor or friend
Awareness of programs
General interest
Amount of paperwork
Government sponsorship of the program
Other
70
Percent responding very/extremely important
Farmers who have participated in USDA conservation incentives programsFarmers who have NOT participated in USDA conservation incentives programs
Figure 6115
Growers who expressed interest in participating in a cost-share program for pollinators
YES(98%)
NO(1%)
Interested in managing habitat for wild bees in the future
Informational pamphlet about USDA cost-sharing programs for wild bees would be useful
YES(95%)
NO(3%)
Not extremely(95%)
Extremely(0%)
Importance of beauty or landscaping in their decision to manage habitat for wild bees
Figure 7116
Cranberry Grower Survey
1. Are you currently acting as the primary farm operator for your cranberry farm? Acting as the primary farm operator would include handling day to day operations and making decisions regarding farming practices for cranberry production done on your farm.
No
Yes Go to question 2
We would appreciate your assistance in passing this survey on to the person who is the primary farm operator for your Cranberry farm.
If you are unable to pass the survey on to that person, please make note of that here and return the blank survey in the enclosed postage paid envelope.
Thank you!
This study is being conducted in order to understand how cranberry growers regard on-farm conservation programs in general and pollinator conservation programs in particular.
Your answers will be kept confidential, and not released in a way that would allow you to be identified.
117
2. Within the past 5 years, if you ever irrigated, in general, what time of day did you irrigate your cranberry beds? (Check all that apply)
Between 3:00 p.m. and 7:00 p.m.
Between 8:00 a.m. and 3:00 p.m.
Before 8:00 a.m.
After 7:00 p.m.
I never irrigated over the past 5 years
Before 8:00 a.m.
Between 8:00 a.m. and 3:00
p.m.
Between 3:00 p.m. and 7:00
p.m.After 7:00
p.m.
I did not apply in past 5 years
a. Insecticide
3. Within the past 5 years, if you applied any of the following substances, what time of day did you usually apply them to your cranberry beds? (Check all that apply)
b. Herbicide
c. Fungicide
4. Within the past 5 years, if you applied any of the following substances, how many applications, on average, did you spray per year of each of the following? If you did not spray at all, please write ‘0’ for that substance.
a. Insecticide
b. Herbicide
c. Fungicide
d. Other, please describe
Number of applications
per year
1
MANAGEMENT OF CRANBERRY MARSHES
This section is about current management practices and where you obtain information on management practices.
118
5. Within the past 5 years, did you ever use the following types of insecticide products on your cranberry beds?
a. Insect growth regulators such as Confirm or Intrepid
b. Neonicotinoids such as Assail
c. Organophosphates such as Diazinon or Lorsban
d. Some other products, please describe
Yes No
6. Is your farm certified organic?
No
Yes
7. Do you hire a scout to monitor for insect pests?
No
Yes
8. Within the past 5 years, which of the following methods did you use to determine when to spray for insect pests? Did you use…
a. ...a calendar schedule?
b. ...scout reports?
c. ...some other method? please describe
Yes No
9. Within the past 5 years, what insect pest management strategies did you use in addition to or instead of insecticides? Did you use…
a. ...flooding?
b. ...biological control?
c. ...sanding?
d. ...some other strategy? please describe
Yes No
2
119
10. Below is a list of sources of information about pest, nutrient, water, frost and other management practices. For each source of information, please indicate if you have used that source in the past, if you are using it now, if you want to use it in the future, or if you have NEVER used that source of information, and would not want to in the future. (Check all that apply)
c. Natural Resources Conservation Service (NRCS) Staff
Used it in the past
Use it now
Want to use it in the future
Never used it and will not in the
futurea. Crop Scout
b. University Extension Agent
d. Neighbor or friend
e. Cranberry School
f. Information hotline (phone)
g. Newsletter (paper)
h. Email listserv
i. Text message (automatic alerts)
j. Social networking site (Facebook, Twitter)
13. Have you ever rented honey bees for pollination in the past?
No
Yes
12. How many hives per acre?
hives
11. Do you currently rent honey bees for pollination?
No
Yes
Go to question 13
k. Website (please list below)
l. Other (please list below)
3
This section is about your current pollination practices and your awareness about wild bees and pollinator habitat.
POLLINATORS - WILD AND MANAGED
120
15. How many hives per acre?
hives
14. Do you use commercial bumble bees for pollination?
No
Yes
Go to question 16
16. Have you noticed a change in the availability of honey bee hives in the past 5 years?
No
Yes Please explain
17. Have you observed a change in the quality of rented honey bee hives in the past 5 years?
No
Yes Please explain
18. How familiar are you with Colony Collapse Disorder (CCD)?
Somewhat familiar
Slightly familliar
Not at all familiar
Very familiar
Extremely familiar
19. Can you distinguish between wild bees and honey bees?
No
Yes
20. How important do you think honey bees are for cranberry pollination?
Somewhat important
Slightly important
Not at all important
Very important
Extremely important
21. Are you aware of any wild bees on your property?
No
Yes
4
121
22. Do you actively manage habitat to encourage wild bees on your property?
No
Yes Please explain
23. If you do manage habitat for wild bees, did you receive cost-share funding to do this?
No
Yes Through whom?
24. Are you aware of cost-share programs for establishing pollinator habitat?
No
Yes
25. Have you altered your management in any way to encourage wild bees on your property?
No
Yes Please explain
Not at all Slightly Somewhat Very Extremelya. Financial commitment
Not applicable
b. Time commitment
c. Space requirements
d. Availability of cost share programs
26. How important are each of the following in your decision about whether or not to manage for wild bees?
e. Availability of technical support
f. Environmental stewardship
g. Encouragement of neighbor or friendh. Knowledge about wild bees
i. Knowledge about pollinator habitat
j. Perceived weed problems from pollinator habitat
k. Perceived pest problems from pollinator habitat
l. Beauty or landscaping
m. Importance of pollination for cranberries
n. Other (please explain)
5
Go to question 24
122
27. If you do not currently manage for wild bees on your property would you be interested in managing for wild bees in the future?
No
Yes
28. Which of the following would be useful to you?
UsefulNot
usefula. Field day demonstration of a pollinator habitat project
b. Informational pamphlet about wild bees
c. Field guide to wild bees
d. Informational pamphlet about USDA cost-sharing programs for wild bees
e. Website about wild bees
f. Website about pollinator habitat
g. Website about USDA cost-sharing programs for wild bees
h. Other (please explain)
29. Do you currently manage habitat (prairie plantings, woodlots, wetland, etc.) for the specific goal of protecting wildlife (birds, beneficial insects, mammals, etc.)?
No
Yes
Go to question 32
30. If yes, do you receive cost-share funding to support these conservation management activities?
No
Yes
31. Through whom do you receive cost-share funding to support these conservation management activities?
6
The next questions are about on-farm conservation programs in general.
Go to question 32
ON-FARM CONSERVATION PROGRAMS
123
32. How important are the following in your decision whether or not to manage habitat for wildlife such as birds, beneficial insects, and mammals?
a. Financial commitment
b. Time commitment
c. Space requirements
d. Availability of cost share programs
e. Availability of technical support
f. Environmental stewardship
g. Encouragement of neighbor or friendh. Knowledge of wildlife
i. Knowledge of wildlife habitat
j. Perceived weed problems from wildlife habitat
k. Perceived pest problems from wildlife habitat
l. Beauty or landscaping
m. Recreation, such as hunting or hiking
Not at all Slightly Somewhat Very ExtremelyNot
applicable
n. Other (please explain)
33. Do you or have you ever participated in any of the following United States Department of Agriculture (USDA)-sponsored conservation incentive programs? (Check all that apply.)
Environmental Quality Incentives Program (EQIP)
Conservation Reserve Enhancement Program (CREP)
Conservation Reserve Program (CRP)
Wildlife Habitat Incentives Program (WHIP)
Other, please explain
34. Do you currently participate in the Whole Farm Planning Incentives Program (WFPIP)?
No
Yes
7
124
35. Do you currently participate in any programs that address the following environmental/conservation issues? (Check all that apply.)
Water quality and conservation
Pest management
Nutrient management
Soil conservation
36. In the future, how likely are you to participate in a USDA-sponsored conservation incentive program?
Somewhat unlikely
Very unlikely
Extremely unlikely
Not sure
Somewhat likely
Very likely
Extremely likely
37. In the future, how likely are you to participate in a non-USDA-sponsored conservation incentive program?
Somewhat unlikely
Very unlikely
Extremely unlikely
Not sure
Somewhat likely
Very likely
Extremely likely
8
125
Very ExtremelyNot
applicable
38. How important are the following in your decision whether or not to participate in a conservation incentive program?
a. Financial commitment
b. Time commitment
c. Availability of cost share programs
d. Availability of technical support
e. Environmental stewardship
f. Encouragement of neighbor or friend
g. Awareness of programs
h. General interest
i. Amount of paperwork
Not at all Slightly Somewhat
k. Other (please explain)
j. Government sponsorship of program
39. Would you be interested in participating in a cost-share conservation program to enhance wild bees through the planting of pollinator habitat on your property?
No
Yes
40. Do you know who your NRCS officer or representative is?
No
Yes
41. Are you comfortable working with or getting information from your NRCS officer or representative?
No
Yes
9
126
42. What is your age?
35-44
25-34
Less than 25 years old
45-54
55-64
65 or older
43. What is your gender?
Female
Male
44. In which county (or counties) do you grow cranberries?
45. How many acres of cranberries do you manage?
150 – 300 acres
50 – 149 acres
Less than 50 acres
Over 300 acres
46. How many total acres do you own or manage, including all cranberry and non-cranberry?
acres
47. How long have YOU been growing cranberries?
years
10
These last questions are about you.
YOUR BACKGROUND 127
50. Is growing cranberries your primary source of income?
No
Yes
THANK YOU for your participation!
If you have any questions or concerns, please contact Hannah Gaines by phone (774-392-0498) or email ([email protected]).
If there is anything else you would like us to know, please use the space below to do so.
48. How long has YOUR FAMILY been growing cranberries?
years
49. Do you regularly attend Cranberry School?
No
Yes
11
128
1
THESISCONCLUSION
Theresearchpresentedinmydissertationcontributestoourunderstandingofhow
localandlandscapefactorsinfluencebeesandcranberrypollination.Fromthescaleofa
singlecranberryflowertothelandscapeofcentralWisconsin,theresultsofmy
dissertationprovidenewevidenceregardingthepollinationrequirementsofcranberry,the
contributionofnativeandmanagedbeestocranberryyield,andtheinteractingeffectof
landscapeonhivestockingdensities.Thisresearchalsoprovidespracticalinformation
thatcanbeappliedtofuturemanagementandconservationpractices.
1. NativebeesintheWisconsincranberryagroecosystemrespondpositivelytoa
gradientofincreasingwoodlandinthesurroundinglandscape.Similartothe
findingsofpreviousresearchinotheragroecosystems,Ifoundthatnativebee
abundanceandspeciesrichnessincreaseastheamountofwoodedhabitatinthe
surroundinglandscapeincreases.Sinceagriculturallandscapesarehighlymanaged
andprovideforagingresourcesforonlyashorttimeduringthegrowingseason,close
proximitytonon‐crophabitatswouldallowbeestouseresourcesinboththecropand
non‐crophabitats.
2. Thecontributionofnativebeestocranberryyieldishighlyvariablefromyearto
yearandonlybecomesevidentwithyearsofdatafrommanysites.WhileIwas
unabletofindevidenceofaneffectofnativebeestocranberryyieldinmyfieldstudies
(chp.3),whenIanalyzedoveradecadeworthofhistoricaldatafromthegrowers,I
foundevidencethat,intheabsenceofhoneybees,cranberrymarshesinhighwooded
129
2
landscapesachievedamarginallyhigheryieldthanmarshesinlowwoodland
landscapes(chp.2).Sincenativebeeabundanceandspeciesrichnessincreasedwith
theamountofwoodlandinthesurroundinglandscape,itseemslogicalthatthis
marginalincreasemaybeduetopollinationbynativebees.Additionally,studiesfrom
othersystemshavefoundthatcropyieldincreaseswithbeediversity,further
supportingthehypothesisthattheyieldincreaseisduetonativebees.Inashortterm
fieldstudy,anyeffectofnativebeesonyieldislikelymaskedbythestronginfluenceof
localmanagementpracticesincludingtheuseofmanagedhoneybees.Myresults
suggestthattheuseoflargehistoricaldatasetscouldbeavaluablesourceof
informationregardingthecontributionofnativebeestocropyieldforadiversityof
croppingsystemsaroundtheworld.Furthermore,thisdatashouldbeavailablefor
manycropsasgrowersoftenkeepcarefulrecordsoftheiryieldandmanagement
practices.
3. Non‐bioticfactorscontributesignificantlytocranberrypollination.Theresultsof
myfieldandgreenhouseexperimentsclearlydemonstratethat,contrarytoprevious
research,intheabsenceofbees,cranberryisstillabletoproduceviablefruit(chp.1).
Atthesametime,mydataalsosupportpreviousresearchconfirmingthatcranberryis
unlikelytoself‐pollinate.Previousstudiesweredesignedtocompareself‐pollination
andbioticpollinationonly,ignoringthepossibilityofalternativemechanisms.Isuggest
thatpollinationwithoutbeesisduetothephysicalagitationoftheplants(e.g.,bywind)
causingflowerstobumpintoeachother,resultinginthetransferofpollenbetween
flowers.Thismechanismisfacilitatedbyhorticulturalpracticesincommercial
130
3
cranberryproductionwhereplantsgrowdenselyandproduceasuperfluousnumberof
flowers.Muchofthecranberrypollinationliteraturewasdonedecadesagowhen
yieldsweremuchlower.Advancesintechnologyandmanagementpracticeshaveledto
higheryieldsandmoredenselyplantedcranberryuprights,suggestingthat
horticulturaldifferencesmaybepartiallyresponsiblefordifferencesbetweenmy
resultsandolderstudies.Innaturalbogcommunities,wherecranberrygrowssparsely,
thesefindingsmaynotbeapplicable.
4. Theeffectivenessofhoneybeesascranberrypollinatorsislandscapedependent.
Inlowwoodlandlandscapes,yieldisstronglycorrelatedwithhivedensity,butinhigh
woodlandlandscapes,thereisnoevidencethatincreasinghivedensityhasanyeffecton
yield(chp.2).Thissuggeststhatonsomemarshes,bringinginhoneybeesfor
pollinationhasabigeffect,whileonothers,honeybeesprovidelittleornoadded
benefit.Oneexplanationforthispatternisthathoneybeesmaybedrawnawayfrom
themarshestoforageonnon‐cropfloralresources.Apreviousstudyincranberry
foundthathoneybeesreturningtothehivecarriedverylittlecranberrypollen,
supportingtheideathatthebeesareforagingelsewhere.Thishasimplicationsfor
managementdecisionsincludingwhethertousehoneybeesforpollinationand
whetherprovidingsupplementalforagingresourceswithinagriculturalareasmay
enhancethecroppollinationactivityofhoneybees.
5. Cranberryyieldrespondsmoststronglytolocalfactorsincludinghoneybeehive
stockingdensityandfloweringuprightdensity.Thissuggeststhatgrowershave
quiteabitofcontroloveryieldthroughtheirmanagementdecisions.Therewasalsoan
131
4
interactingeffectoffloraldensityandhivestockingdensity(chp.3)indicatingthatif
growerscanincreasetheirfloraldensity,theywillalsobeincreasingtheamountof
pollinationservicescontributedbybees.Alternatively,iffloraldensityislow,
increasingthenumberofbeeswillnotleadtoahigheryield.
6. Interestamongcranberrygrowersregardingon‐farmpollinatorhabitat
conservationishigh,butalackoftechnicalsupportandbureaucratichurdles
createbarrierstotheirparticipationinfederalcost‐shareprograms.Theresults
ofmyresearchsuggestthatparticipationcouldbeincreasedthroughoutreachactivities
thatpromotethecost‐shareprograms,areductionofbureaucratichurdlesto
participate,andincreasingtheavailabilityoftechnicalsupporttogrowersonhowto
managehabitatfornativebees(chp.4).Ialsosuggestthatapeer‐mentorprogrammay
beusefulasmanyofthegrowersreportedgettingmanagementinformationfromtheir
friendsandneighbors.
132