Color polarization vision mediates the strength of an ...degrees of polarization (d) depending on the darkness/brightness of the water body and on the angle of reflection (Horváth

Evolutionary Applications 201912175ndash186 emsp|emsp175wileyonlinelibrarycomjournaleva

Received1January2018emsp |emsp Revised19July2018emsp |emsp Accepted20July2018DOI 101111eva12690

O R I G I N A L A R T I C L E

Color polarization vision mediates the strength of an evolutionary trap

Bruce A Robertson1 emsp|emspGaacutebor Horvaacuteth2

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicensewhichpermitsusedistributionandreproductioninanymediumprovidedtheoriginalworkisproperlycitedcopy2018TheAuthorsEvolutionary ApplicationspublishedbyJohnWileyampSonsLtd

1DivisionofScienceMathematicsandComputingBardCollegeAnnandale-on-HudsonNewYork2EnvironmentalOpticsLaboratoryDepartmentofBiologicalPhysicsPhysicalInstituteELTEEoumltvoumlsLoraacutendUniversityBudapestHungary

CorrespondenceBruceARobertsonDivisionofScienceMathematicsandComputingBardCollegeAnnandale-on-HudsonNYEmailbrobertsbardedu

Funding information FundingforthisstudywasgenerouslyprovidedbyBardCollege

AbstractEvolutionarytrapsarescenariosinwhichanimalsarefooledbyrapidlychangingcon-ditionsintopreferringpoor-qualityresourcesoverthosethatbetterimprovesurvivalandreproductivesuccessThemaladaptiveattractionofaquaticinsectstoartificialsourcesofhorizontallypolarizedlight(egglassbuildingsasphaltroads)hasbecomeafirstmodelsystembywhichscientistscaninvestigatethebehavioralmechanismsthatcausetrapstooccurWeemploythisfield-basedsystemtoexperimentallyinves-tigate(a)inwhichportion(s)ofthespectrumarepolarizationallywater-imitatingre-flectorsattractivetonocturnalterrestrialandaquaticsinsectsand(b)whichmodernlamptypesresultingreaterattractioninthistypicalkindofnocturnalpolarizedlightpollutionWefound thatmostaquatic taxaexhibitedpreferences for lampsbasedupontheircolorspectramosthavinglowestpreferenceforlampsemittingblueandredlightYetdespitepreviouslyestablishedpreferenceforhigherdegreesofpolari-zationofreflectedlightmostaquaticinsectfamilieswereattractedtotrapsbasedupontheirunpolarizedspectrumChironomidmidgesaloneshowedapreferenceforthecoloroflamplightinboththehorizontallypolarizedandunpolarizedspectraindi-catingonlythisfamilyhasevolvedtouselightinthiscolorrangeasasourceofinfor-mationtoguideitsnocturnalhabitatselectionTheseresultsdemonstratethatthecolorofartificial lightingcanexacerbateorreduce itsattractivenesstoaquatic in-sectsbutthatthestrengthofattractivenessofnocturnalevolutionarytrapsandsotheirdemographicconsequences isprimarilydrivenbyunpolarized lightpollutionThisfocusesmanagementattentiononlimitingbroad-spectrumlightpollutionaswellasitsintentionaldeploymenttoattractinsectsbacktonaturalhabitats

K E Y W O R D S

aquaticinsectbehaviorcolorpreferencecolorsensitivityevolutionarytraplightpollutionmaladaptationpolarizedlightpollution

1emsp |emspINTRODUC TION

Animalsuselightasasourceofinformationtoguidebehaviorssuchasorientation(LorneampSalmon2007)reproduction(RandBridarolliDries amp Ryan 1997) and foraging (Dwyer Bearhop Campbell amp

Bryant2013)Someofthislightispolarizedmeaningthatthedirec-tionoftheelectricfieldvectoroflightisorientednonrandomlysuchthat itvibratespredominantly insomedirection (eghorizontally)Skypolarizationpatternsmaybeusedbyanimalsasacuefororien-tation (DackeNilsson Scholtz ByrneampWarrant 2003Muheim

176emsp |emsp emspensp ROBERTSON aNd HORVAacuteTH

Sjoumlberg amp Pinzon-Rodriguez 2016) but the predominant sourceof polarized light on the earthrsquos surface iswater bodies (HorvaacutethKriskaMalikampRobertson2009)Whensun-ormoonlightreflectsfromwatersurfacesittypicallybecomeshorizontallypolarizedwithdegreesofpolarization(d)dependingonthedarknessbrightnessofthewaterbodyandontheangleofreflection(Horvaacutethetal2009)Historicallywater-seekingaquaticinsectshavereliablyusedthehor-izontalpolarizationof lightreflectedwithahighenoughdegreeofpolarizationdtolocatewatersourcestheyrequiretofeedmateandlayeggs(reviewedinHorvaacutethampCsabai2014)Todayitiswell-knownthattheabundanceofsmoothdark-coloredman-madeobjects(egasphalt glass buildings black cars solar panels dasymp80ndash100 atand near the Brewsterrsquos angle θBrewster=arctan n measured fromthenormalvectorofthereflectingsurfacewithrefractiveindexn KriskaHorvaacutethampAndrikovics1998KriskaCsabaiBodaMalikampHorvaacuteth2006KriskaBernaacutethFarkasampHorvaacuteth2009Horvaacutethetal20092010)canpolarizelightmorestrongly(iewithhigherd)thananynaturalwaterbody(dasymp30ndash80FlamariqueampBrowman2001 Horvaacuteth 2014) andwater-seeking aquatic insects are pref-erentially attracted to oviposit upon these supernormally intensesources of horizontal polarization where their eggs fail to hatch(Horvaacutethetal2010Kriskaetal199820062009)Adults thendietypicallywithnosecondopportunitytoreproduceSuchcasesofanthropogenicchangewhereevolvedcue-resourcecorrelationsareuncoupledandanimalscometoprefer inferiorhabitatsareknownasecologicaltraps(DwernychukampBoag1972)whichareonetypeofthebroadercategoryofevolutionarytrapsinwhichanytypeofbehaviorissimilarlyaffected(SchlaepferRungeampSherman2002)Thesesevereformsofbehavioralmaladaptationarebeingdetectedat an increasing frequency (RobertsonampHutto 2006 RobertsonRehageampSih2013)areknowntoleadtosevereandrapidpopu-lationdeclines(DelibesFerrerasampGaona2001FletcherOrrockampRobertson2012KokkoampSutherland2001)andaffectabroadrange of insect taxa (Robertson etal 2013 Robertson Campbelletal2017Robertsonetal2018)

Whenplacedabovewaterbodiesorartificialpolarizers lampscreate both polarized (due to reflected light) and unpolarized (bydirectlight)photopollutionthatcanindividuallyorincombinationtriggermaladaptivebehaviorinadiverserangeofinsecttaxa(BodaHorvaacutethKriskaBlahoacuteampCsabai2014RobertsonCampbelletal2017Szaacutezetal2015)Yetspeciesdifferwidelyintheirsusceptibil-itytothesedifferentformsoflightpollution(RobertsonCampbelletal2017)andthespecificmechanismsbywhichtheseman-madelightsourcesinterfacewithevolvedbehavioralrulestotriggermal-adaptationremainpoorlyunderstood(Houmllkeretal2010Horvaacutethetal2009LongcoreampRich2004PerkinHoumllkerampTockner2014Robertson etal 2013) Artificial light sources (eg light-emittingdiodesmdashLEDshigh-pressure sodium [HPS] lamps)differ fromsun-lightandmoonlight in thatdifferent lamptypesemphasizeoren-tirelyomitparticularwavelengthsoflight(GastonVisserampHoumllker2015)Whenitreflectsfromaman-madeobjectlamplightwillnotbe polarized evenly across the visible spectrumWavelengths re-flectedwith highest intensity are polarized the least that iswith

lowestdegreeofpolarization (HegeduumlsampHorvaacuteth2004) Incon-trastapartfromthecasewhenthesunandmoonareclosetothehorizonsun-andmoonlightarerelativelyuniforminintensityacrossvisiblewavelengths(Gastonetal2015)andsopolarizationofnat-urallightviareflectionfromcolorless(ieblackgreyorwhite)waterbodieswillresultinasimilarlyalmostuniformdegreeofpolarizationacrossthevisiblespectrumCollectivelythismeansthatbyempha-sizingparticularwavelengthsoflightartificiallightpollutionhastheevolutionarilynoveleffectofdecouplingthehistoricallycorrelatedrelationship between the intensity of light and the degree of po-larizationofwater-reflectedlight Italsomeansthat ifnight-activewater-seekingflying insectsdisproportionatelyrelyupondifferenthorizontally polarized portions of the visible spectrum to locatewaterbodiesartificiallightinglocatednearartificialpolarizers(egasphaltglassbuildings)couldmakethoseobjectsmoreattractivetonocturnalovipositinginsectsthanactualwaterbodies

In this way lamps that emit different spectra of unpolarizedlighthave thecapabilityofexacerbating theseverityofecologicaltrapscausedbynocturnalpolarizedlightpollutionortoreducetheirdemographic impacts because stronger habitat preferences trig-ger ecological trapswithmore severedemographic consequences(Fletcheretal2012HaleampSwearer2016Robertsonetal2013)Global-scale shifts toward broader spectrum street lighting (egLEDs) for theirenergyefficiencyandsafetybenefitswill result inshiftsintheavailablewavelengthsofnocturnalpolarizedlightacrosslargelandscapeswithpotentialtoexacerbateormitigatemaladap-tivebehavioralresponsesinwater-seekinginsects

We designed a field experiment to ask (a) whether nocturnalwater-seekingaquatic insects relyupondifferent spectraofpolar-izedlighttolocatewaterbodiesand(b)whichlamptypes(andpor-tionsoftheirvisiblespectra) insectsarepreferentiallyattractedtoand sowhich sensory cuesmitigate or exacerbate the severity ofevolutionary traps triggered by nocturnal polarized light pollutionOur approach examines variation in riverside captures of sevenaquatic and six terrestrial night-active insect families in simulatedwater bodies (white and black oil-filled trays) illuminated by LEDmetal-halide(MH)orHPSlampsWeinterpretdifferentiallyhighercaptures among lamp types illuminating oil-filled horizontal whitetrays (full-spectrumreflectorsofunpolarized light)asanattractiontoaparticularspectrumofunpolarizedlightproducedbythatlampandreflectedbythetrayDifferentialattractiontoaparticularlamptreatment illuminatingblack trays (full-spectrumreflectorsofhori-zontallypolarizedlight)isinterpretedasevidenceofpreferentialat-tractiontoparticularspectraofhorizontalpolarizationInteractionsbetweentrayandlamptypessuggestpreferentialattractiontohori-zontallypolarizedlightofaspectrumthatdiffersfromthepreferredspectrumofunpolarizedlightandvice versaindicatingindependenceof detectionof thewavelength andpolarization of reflected lightTerrestrialtaxaarecommonlyattractedtobrightunpolarizedlights(Nowinszky2003NowinszkyKissSzentkiraacutelyiPuskaacutesampLadaacutenyi2012) andempirical studies suggest their ventral eye regiondoesnotperceivethepolarizationof lightcomingfrombelow(reviewedinHorvaacuteth2014RobertsonCampbelletal2017)butevidenceis

emspensp emsp | emsp177ROBERTSON aNd HORVAacuteTH

limitedWeexaminethebehavioralresponsesofterrestrialtaxatopolarizedlighttreatmentstoconfirmthesenegativeresponsesatourstudysitesandsocontextualizetheimpactsofpolarizedandunpo-larizedlighttreatmentsonaquaticinsects

2emsp |emspMATERIAL S AND METHODS

21emsp|emspStudy sites and experimental design

Weselected study siteson fivedifferent tributariesof theHudsonRiverinsouthernNewYorkStateUSA(Figure1a)Studysiteswereidentifiedinsparselypopulatedareasalongheavilyforestedrivercor-ridorsWechoseresidentialpropertiesthatmaintainedmownlawnsextendingfromthehighwaterlineinlandatleast60mtoensuresuffi-cientareaforourexperimentandsothatvegetationwouldnotimpede

insectsrsquolinesofsightormovementtowardexperimentaltestsurfacesIn2013wetrappedflyingemergentaquaticinsectstwotimesateachsiteVisit1)June17ndashJuly6andVisit2)July10ndash24Eachtrappingses-sionlastedfor120minbeginningexactly30minaftersunset

We used black and white trays (60times40cm) filled with trans-parent common salad oil to capture insects and assess their rela-tivepreferencefortestsurfacesvaryingintheira)illuminationtypeandb)degreeofpolarizationofreflectedlight(dFigures1band2)Weplaced threeofeachof the two traycolors (blackwhite) inarowparallel totheriverbankatadistanceof5metersspaced05mapartAboveeachofthetrayswesuspendeda180-wattarrayof LED a150-wattHPSbulbor a150-wattMHbulb (Figure1b)WeusedanOceanOpticsUSB2000+spectrometertomeasuretheabsoluteirradiance(μWcm2nm)ofeachlightsourceasafunctionofwavelengthWeselectedlampsofwattagesthatproducedsim-ilaroverall luminosity (11000ndash12000 lumens) among lamp typesthough at differentwavelengths (Figure3)We chose a relativelynarrowband(yellowgreen550ndash630nm)HPSsourceaLEDsourcethatpeaksattwocomplementaryrangesofwavelengths(blue420ndash480nmred620ndash680nm)withredwavelengthsbeingmorecom-monatdawnandduskandaMHlampthatproduceslightatnearlyallvisiblewavelengths(Figure3)Eachlightsourcewasplacedinsideadownwardfacing35times54cmaluminumreflectorpaintedblackontop tominimize itsvisual signatureandelevated30cmabove thetest surface In thisway therewere twotreatmentsofeach lamptypeilluminatingbothwhiteandblacktraysLampswerehunglowabove test surfaces (25cm) and shaded to illuminate the test sur-faces such that insect responseswould be based upon their per-ceptionofonlytray-reflected(notdirect)lightexceptatveryshortrange(afterRobertsonCampbelletal2017)

We exposed these six combinations of lamp spectra and traycolortreatmentstoaquaticinsectsduringeachofthetwovisitstoeachofthefivestudysitesAdjusted(seeStatisticalAnalysisbelow)capturesfromeachofthetwosamplingsessionsatastudysitewerecombinedpriortoanalysis leadingtoatotalsamplesizeofn=30forthestudyWerandomizedthelocationoftraysandlamptypesinourstudytoavoidbias incapturesassociatedwiththephysicalpositionoftreatmentsintheexperimentrelativetothesurroundingenvironmentandtheothertreatmentsThespatialarrangementoftreatmentcombinationswasrandomizedateachsitevisitwiththeexceptionthatlightsourcesofthesametypewereneverplacedim-mediatelynexttoeachotherThepositionsofthewhiteandblacktrays within a lighting treatment were exchanged every 20minAftereachsamplingsessionwepouredtraycontentsthroughfinecheeseclothtoseparateinsectswhichwerestoredin80ethanolforlateridentificationtothefamilylevel

22emsp|emspFocal taxa and experimental predictions

Wedefine aquatic insect taxa as thosewith a larval or adult life-historyphasedependentontheavailabilityofafreshwatersource(JohnsonampTriplehorn2004)Wefocusedouranalysisonaquaticinsect taxa known to exhibit stronger attraction to sources of

F IGURE 1emspa)LocationsoffieldexperimentsintheHudsonValleyregionofsouthernNewYorkStateStarsindicatelocationsofexperimentalsitesonfivetributariesoftheHudsonRiver(inwhite)locatedintwoofthesixcountiesoutlinedinblackb)Thespatialarrangementofoil-filledblackandwhitetraysilluminatedby1)high-pressuresodium(HPS)lamp2)light-emittingdiodes(LED)and3)metal-halide(MH)lampateachstudysiteOil-filledtraysoftwocolorsreflectlamplightandpolarizeittoagreater(black)orlesser(white)degreeandareplaced5metersfromtheedgeofstudystreams


horizontally polarized light with greater d-values Ephemeroptera(Horvaacuteth etal 2010 Kriska etal 1998 2006 Szaacutez etal 2015)Trichoptera(Horvaacutethetal2010KriskaMalikSzivaacutekampHorvaacuteth2008) Empididae (Robertson etal 2018) Simuliidae (Robertsonetal2018)andChironomidae(HorvaacutethMoacuteraBernaacutethampKriska2011Lerneretal20082011Robertsonetal2018)Whenseek-ingasuitableovipositionsitethesetaxatouchdownonthesurfaceofwaterbodies(Horvaacuteth2014Kriskaetal2006)butwillbecap-turedintheoiltrapsbecauseoilwetschitinpreventingtakeoff

23emsp|emspPolarizing properties of traps

Wemeasuredthereflectionndashpolarizationcharacteristicsofoil-filledtraysusingimagingpolarimetry(HorvaacutethampVarjuacute2004)inthered(650plusmn40nm=wavelength of maximal sensitivityplusmnhalf band-width of the polarimeter detector) green (550plusmn40nm) and blue(450plusmn40nm)portionsof thespectrumWeperformedpolarimet-ricmeasurementsinadarkroomwiththeopticalaxisaimeddown-wardwiththereflector(oil-filledtrays)aimedattheBrewsterrsquosangleθBrewster=arctan(n =15)=563degfromtheverticalcalculatedfortherefractiveindexn=15ofvegetableoilExposureswerestandardizedfor130thofasecondat400ISOAttheBrewsterrsquosanglesurface-reflectedlightisperpendiculartotherefractedraypenetratingthe

oilresultinginthehighestpossibledegreeofpolarizationThreeim-agesweretakenofeachobjectwiththepolarizationfilteratdiffer-entanglesThoseimageswereprocessedviaPolarworkscopysoftwareinto images illustrating thedegree (d) andangle (α)ofpolarizationat each pixelWe estimated the maximum degree of polarizationassociatedwithtraysbyquantifyingthefractionofthed-patternscomposed of black pixels using ImageJcopy softwareWe calculatedawindowsizerepresenting110ofthetraylengthandheightandestimatedmaximumdasthewindowgivingthemaximumvalueoutof20nonoverlappingsamplelocationswithintheimageofeachtraynonrandomlyfocusedonportionsofthetraywiththehighestvisualestimatesofdontheblack-to-whitescale

24emsp|emspStatistical analyses

Weexaminedtheeffectof1)lamptypeand2)thed(degreeofpolar-izationofreflectedlamplight)onthenumberofcapturedindividualswithineachinsectfamilyFamily-specificanalysesservetoillustratethe behavioral preferences of each taxon For these analyseswecreatedastandardizedabundancemetricthatwouldbereflectiveoftherelativefractionofindividualsattractedtoeachtreatmentandthatwasinsensitivetobiasassociatedwithdisproportionatelyhighor low captures at one study site relative to othersWe adjusted

F IGURE 2emspReflectionndashpolarizationcharacteristics(angleαmeasuredclockwisefromtheverticalanddegreedoflinearpolarizationofreflectedlight)ofsalad-oil-filledblack(left)andwhite(right)traysilluminatedbyhigh-pressuresodium(HPS)(top)metal-halide(center)andlight-emittingdiode(bottom)lampusedinthechoiceexperimentsmeasuredwithimagingpolarimetryinthered(650nm)green(550nm)andblue(450nm)portionsofthespectrumWhitetraysreflectedlamplightwithmuchlowerdegreesofpolarizationthanblacktraysThemaximumdvalueisgivenforeachtrayHPSlightwasreflectedwiththehighestdinthebluereflectedLEDlightwasmoststronglypolarizedinthegreenandMHlamplightwaspolarizedtoahighandmoresimilardegreeacrosstheredgreenandbluespectraBlacktraysconsistentlyreflectedhorizontallypolarizedlightWhitetraysreflectedprimarilyverticallypolarizedlightwiththeexceptionofhorizontalpolarizationofthemirrorimageofthelightsource


captures of each family associatedwith each test surface to be afractionof the total numberof captures for that visit at that sitethenmultipliedeachfractionby100roundingtothenearestwholeindividualWemodeledstandardizedcapturesusinggeneralizedlin-earmodels and fitmodels to negativebinomial or Tweedie distri-butionswitha log-link functionTweedie isa familyofprobabilitydistributionswhichhavepositivemass at zero but areotherwisecontinuousandincludethePoissondistributionwhichistypicalofcountdataNegativebinomialdistributionsarealsocommonlyusedtomodelcountdataWemodeled lampandtraytypeascategori-calvariablesinpredictingcapturesforeachfamilyandselectedtheerrordistributionmodelthatminimizedoverdispersion(ĉ)

3emsp |emspRESULTS

31emsp|emspPolarization imagery

Black trays were more efficient polarizers of reflected light thanwhite ones (Figure2) The black test surfaces reflected lightwithvery high degrees of polarization (dblack=53ndash88) including

valueshigherthanwaterisgenerallycapableof(dwater=30ndash80Horvaacuteth amp Varjuacute 2004 Horvaacuteth 2014) White trays reflectedvery low degrees of polarization (dwhite=15ndash21) Black testsurfaces reflected alwayshorizontally polarized light (because thehorizontally polarized light reflected from the airndashwater interfacewas dominating)whilewhite test surfaces reflectedmainly verti-callypolarizedlight(becausetheverticallypolarizedlightreflectedfromthetrayrsquosbottomandrefractedatthewaterndashairinterfacewasdominating)Thesmallmirrorimageoflightsourcesreflectedbythewhitetrayswashorizontallypolarized(becausethereagainthehori-zontallypolarizedlightreflectedfromthewatersurfacedominated)withmaximumd-values(Figure2)

Blacktraysreflectedlightwithlowerandhigherdinthosespec-tralrangesinwhichtheintensityofilluminatinglamplightwashigherandlowerrespectively(Figures2and3)HPSlightreflectedbytheblacktestsurfaceswasmostpolarized inthebluerange(Figure2)whereitsunpolarizedintensityisminimal(Figure3)MHlampsemit-tedmaximalintensityinthegreenspectralrange(Figure3)inwhichthereforethedof lightreflectedfromtheblacktrayswasminimal(Figure2)ContrarytothelatterlamptypetheLEDsusedwereless

F IGURE 3emspSpectraoflightreflectedfromthewhite(toprow)andblack(bottomrow)traysilluminatedbythreedifferenttypesofnocturnalilluminationhigh-pressuresodiumlamplight-emittingdiodesandmetal-halidelampWemeasuredtherelativeirradianceoflightenteringthecosine-correctingsensorofaspectrometer4cmabovethegeometriccenterofthetrayandaimingdownwardat45ointhedirectionofthemirrorimageofthelampabove


intenseinthegreen(Figure3)thusdofLEDlightreflectedbytheblacktestsurfaceswasmaximalinthegreen(Figure2)

32emsp|emspBehavioral responses of arthropods

Wecapturedatotalof75992adultaquaticinsectsin27familiesWefocusedouranalysison13familiesforwhichwewereabletofitmod-elswithoutoverdispersion(ĉlt40alldf=30Table1)Theseincludedsevenfamiliesofaquaticinsects(CaenidaeChironomidaeEmpididaeGlossosomatidae Heptageniidae Hydropsychidae and Isonychiidae)and six families of terrestrial insects (Cecidomyiidae CercopidaeCicadellidaeScarabaeidaeTipulidaeandGeometridae)HydropsychidandisonychiddatawerefitusingmodelswithaTweediedistributionwhiletheremainingfamilieswerefitwithnegativebinomialdistribu-tionsFitofdatatomodelswasgood(ĉ=095ndash391)

Allsevenaquaticfamiliesexhibitedapreferencefor lampandortraytype(Figure4Table1)Isonychidmayflieswerecapturedinpro-gressivelygreaternumbersinLEDandMHlampsthansodiumlamps(Figure4a) while caenid mayflies exhibited a preferential attractiontoLEDsandtowhitetestsurfacesingeneral(Figure4b)Danceflies(Empididae)exhibitedpreferenceonlyforwhitetestsurfacesregard-less of lamp type (Figure4c) Flat-headed mayflies (Heptageniidae)were captured in lowest abundance in LED treatments (Figure4d)Glossomatidandhydropsychidmayflieswerealsocapturedinlowestnumbers in LED-illuminated traps but were also more attracted towhite test surfaces (Figures4ef) Nonbiting midges (Chironomidae)weredifferentially attracted toLEDsassociatedwithwhite test sur-faces while exhibiting lowest preference for themwhen they were

associatedwithblackones(Figure4g)Amongterrestrialinsectsonlyone familyexhibiteddifferential attraction toanexperimental treat-ment Geometrid moths were attracted to sodium and MH lamps relativetoLEDs(Figure4m)

4emsp |emspDISCUSSION

Thereducedintensityofphotonsavailableatnighthasbeenthoughttomaketheprocessofcolorvisiondifficultandsoselectagainstitsuse(WarrantampDacke2011)Yetthespectralcompositionofarti-ficiallightisknowntoaffectdegreeofattractionbybothterrestrialandaquaticinsects(vanGrunsvenetal2014Longcoreetal2015PawsonampBader2014)andwefoundcolorpreferencewasubiqui-tousamongthenight-activeaquaticinsectswestudiedThisstudyrepresentsthefirstexperimentalinvestigationoftheroleofspec-tralpreferenceintriggeringnocturnalevolutionarytrapsinducedbypolarized-light-pollutingartificialreflectorsWefoundthataquaticinsect families exhibited distinct andmaladaptive preferences forspecificwavelengths of both unpolarized light and polarized lightreflectedfromblackorwhitetestsurfacesthathaveimplicationsfortheirconservationandthemanagementoflightpollution

Colorvisionandpreferenceareknownamongdiurnal(HobackSvatos SpomerampHigley 1999 RadwellampCamp 2009) and cre-puscular(Schwind19911995)aquaticinsectsWefoundspectrum-dependentphototaxis innocturnalaquatic insects to fall into fourcategories(a)AttractiontoMHlamps(Isonychiidae) (b)attractionto LED lamps (Caenidae) (c) preference for MH and HPS lamps

TABLE 1emspAbundancemodeloutputinformationrelativetothesevenmostabundantfamiliesofemergentaquaticinsectsandsixfamiliesofterrestrialinsectswithnoaquaticlife-historystagecapturedinthefieldexperiments

Family Common nameNo of individuals

of captures Model

χ2 values

Lamp Tray Interaction

Aquatic

Caenidae Squaregillmayflies 889 11 NB 695 407

Chironomidae Nonbitingmidges 14465 184 NB 1078 5864 5469

Empididae Danceflies 3622 46 NB 265 935

Glossosomatidae Saddle-casemayflies 4955 63 NB 913 480

Heptageniidae Flat-headedmayflies 28375 362 T 722 008

Hydropsychidae Net-spinningcaddisflies 16010 204 T 2310 1033

Isonychiidae Brush-leggedmayflies 1187 15 T 854 091

Terrestrial

Cecidomyiidae Gallmidges 5279 67 NB 058 280

Cercopidae Froghoppers 652 08 NB 495 126

Cicadellidae Leafhoppers 2624 33 NB 052 281

Scarabaeidae Scarabbeetles 222 03 NB 090 261

Tipulidae Craneflies 84 01 NB 399 360

Geometridae Inchwormmoths 85 01 NB 2870 078

NotesCapturesof individuals inthesegroupsweremodeledasfunctionsof lampandtraytypeDatawerefittoeithernegativebinomial (NB)orTweedie(T)distributionsWaldchi-squaredvaluesassociatedwitheachindependentvariablearegivenalongwiththeirstatisticalsignificance(plt005plt001plt0001)


relativetoLEDs(HeptageniidaeGlossosomatidaeHydropsychidae)and(d)polarization-dependentpreferencerelativetoblue-redLEDs(Chironomidae)Broad-spectrumMHlampswerethemostattractive

lamptypeforfiveofthesevenfamiliesThespectraofthespecial-izedblue-redLEDsweused in this studywere foundconsistentlymostattractiveonlybyafamilyoftinymayflies(Caenidae)known

F IGURE 4emspStandardizedabundanceofinsectfamiliescapturedbythedifferentlyilluminatedoiltrapsasafunctionofthetypeofillumination(Lamp)andthecolor(blackorwhite)oftheinsecttrap(tray)Blackandwhiteoil-filledtrayswereplacedbelowdownward-facinglampsofthreetypeshigh-pressuresodiumlight-emittingdiodeandmetal-halide(MseeFigure1)BlacktraysstronglypolarizethevisiblewavelengthsaswoulddarklakesandotherdarkwaterbodiesStandardizedabundanceistherelativenumberofcapturespertreatment(plusmnstandarderror)foreachfamilyofaquatic(a-g)andentirelyterrestrial(h-m)insectsStatisticalsignificanceofindependentvariablesisshownineachfigureplt005plt001plt0001(seeTable1)Lettersabandcaboveeachtreatmentrepresentresultsofpairwisecontrastposthoctestssuchthatcolumnsthatdonotsharethesameletterrepresentmeansthataredifferentatthep=005levelImagesaretypicalmorphologicalrepresentationsofeachtaxonomicgroup


frompreviousstudiestofindthesetwocolorsrelativelyunattractive(RadwellampCamp2009)OthermayfliesinthisstudypreferredMHandHPS lamplight over blue-red LEDswhich is themore typicalbehaviorofdiurnalEphemeroptera(Hobacketal1999RadwellampCamp2009)FiveofthesixterrestrialinsectfamiliesweexaminedexhibitednoclearpreferencewithrespecttospectraExceptionallygeometridmothswereequallyandmoreattractedtotheHPSandMHlampsrelativetoLEDsThisgroupisknownforitsaffinityforbroad-spectrumlightingandlampsemphasizingbluelight(BarghiniampdeMedeiros2012Batesetal2013LampharampKocifaj2013vanLangeveldeEttemaDonnersWallis-DeVriesampGroenendijk2011Somers-YeatesHodgsonMcGregorSpaldingampffrench-Constant2013)butinthisstudywereleastattractedtoLEDswhichemittedthewidest diversity of bluewavelengths and at highest intensityWhether the spectral preferenceswe identifiedoriginally evolvedtoguideanimalsduringanydiurnalorcrepuscularactivityperiodsisunclearItisclearthatthesecolorpreferencesarecurrentlyguid-ingnocturnalbehaviorandcouldrepresentadaptationsdesignedtoguidenocturnalhabitatselectionnavigationormateselectionandsomaybe adaptive non-adaptive ormaladaptive indifferentbe-havioralcontext

Onlyoneofthesevenaquaticandnoneoftheterrestrialinsectfamilies in this study exhibitedpatterns of capture indicating dif-ferentialattractiontoreflectedpolarizedlightspectraIfinsectsin

thisstudywereexhibitingattractionforhighdegreesofpolarizationof reflected light independent of its color they should have beencaptured more frequently in our water-imitating traps associatedwithHPSandLEDlampswhichproducedthehighestdegreesofre-flectedpolarizationbutpatternsofcapturesuggestedthatinsectswerenotusingpolarizedlighttoguidetheirhabitatselectiondeci-sions Insteadcaptureswerecommonly (fiveof thesevenaquaticfamilies)higherintreatmentswithnonpolarizingwhitetestsurfacesindicatingthatphototaxiswasdominantoverpolarotaxisNotablytheseaquaticinsectgroupshavebeenexperimentallydemonstratedtoexhibitpreferenceforwater-reflectedlightwithhigherdegreesofpolarization(EphemeropteraKriskaetal2009trichopteraKriskaetal 2008 Lerner etal 2008 Chironomidae Robertson etal2018) yet ignore thesehabitat selectionpreferenceswhen in thepresence of unpolarized lamp light which simulates their primarynavigationalcuethemoon(Bodaetal2014RobertsonCampbelletal 2017 Szaacutez etal 2015) Robertson Campbell etal (2017)concludedthatthisresultindicatesthatabehavioralcueevolvedtoguidebehaviorinonecontextcanbeexploitedtotriggeranevolu-tionarytrapinanotherAmoreproximateexplanationmaybethatpolarizationsensitivityishousedintheommatidialcellsthatdetectbrightness in insecteyes (eg thebrightnessndashpolarizationhypoth-esisLabhart2016)muchthesamewaythatPapiliobutterflyom-matidiacontainbothpolarizationandcolorsensitivityleadingtoan

F I G U R E 4emspContinued


inabilitytodistinguishbetweenthesetwomodalitiesoflightunderartificial laboratoryconditions (HegeduumlsampHorvaacuteth2004Kelber1999KelberThunellampArikawa2001)

Onlychironomidmidgesexhibitedpatternsofcapturesconsis-tentwithwavelength-specificpreferenceforhorizontallypolarizedlightatnightBecauseunpolarizedwavelengthsproducedathigh-estintensitybylampsarepolarizedtheleastbyreflection(Hegeduumlsamp Horvaacuteth 2004 Figures2 and 3) inference about evidence forcolor-dependentpolarizationpreferencemustalsofollowthisrulePreviousstudiesindicatedthatday-activechironomidsaredifferen-tiallyattractedtounpolarizedgreenbluelight(Burkett1998)butexperimental studiesof colorpreference in thisgroupare rare Inour studymidges homed to the blue andor red unpolarized LEDlightbuttheirreducedattractiontothehorizontallypolarizedlighttreatmentof this lamp type is actually consistentwithpreferencefor polarized light in this same spectrum In accordance with theinverse relationship between intensity and degree of reflectionndashpolarizationredblue lightwaspolarizedthe leastandgreenlightthemostbyourLEDlamppairedwiththeblackpolarizingreflector(Figure2)Andbecausepatternsofchironomidcaptureareinconsis-tentwithpreferencefortreatmentforthehighestdegreeofpolar-izationindependentofcolorwecaninferthatchironomidsexhibitcolor-dependent preference for red or blue horizontally polarizedlightThisresultissurprisingbecauseSchwindrsquos(19911995)stud-iesshowthatall18speciesofday-activeaquaticinsectshestudiedusedspecificspectraofpolarizedlighttoguidetheirhabitatselec-tionbehavior

Whymightsuchwavelength-specificpolarizationsensitivityorpreference evolveThe first possibility is that lakes and rivers ortheirbottomsdifferintheirabsorbancespectruminwaysthataretied to thesurvivalofeggsandyoungShallowwaterbodieswithabundantgreenplantoralgalsubstrates forexample reflect lightwiththehighestdegreesofhorizontalpolarizationintheultravioletandbluerangesofthespectrum(HorvaacutethampVarjuacute2004Schwind1995) Because blue light can penetratemore deeply into naturalclearwaterthanUVandred lightpreferencescouldreflectadap-tationtothedepthatwhichaquaticlife-stageslive(Schwind1995)Alternativelypreferencesforunpolarizedlightofdifferentcolorinnocturnal insectsmaybe aby-productof the evolutionary tuningof light sensitivity to spectra associated with food (HillWells ampWells1997)refugiafrompredationorconspecifics(EllersBoggsampMallet2003)

Nocturnalpolarizedlightpollution(Horvaacutethetal2009)differsfundamentallyfromitsdiurnalcounterpartinthatincontrasttothefullspectrumofsunlightlampspectracanbeaconstrainedsubsetofwavelengthsbut thosethatareproduceddifferdramatically intheintensityatwhichtheyareproduced(Gastonetal2015)Thusthedegreeofpolarizationofareflector(naturalorartificial)canbedecoupled from thewavelength it has been associatedwith overevolutionaryhistoryandtheavailabilityofwavelengthsof lighttobepolarizedcanvarydramaticallydependingonthelampitselfThisrepresentsanovelmechanismfortheformationofanevolutionarytrapFinallybycapturinginsectsimmediatelyupontouchingtheoil

surfaceourtrappingmethodpreventedinsectsfromusingolfactorycues toavoidunsuitablehabitatswhichmaybemorecommon inlongerlivedinsects(Lerneretal2011)

Collectivelyourresultsindicatethatnocturnalaquaticinsectsare capable of color-dependent polarization sensitivity that canexacerbateormitigatetheirresponsestolightpollutiondepend-ing on the spectrum of light pollution they are exposed to andthequalitiesofthereflectingsurfaces(egbuildingsasphalt) inthe vicinity Our results confirm that of previous studies (Bodaetal 2014 Robertson Campbell etal 2017 Szaacutez etal 2015)that nocturnal polarized light pollution is less important in trig-geringevolutionarytrapsthanunpolarizedlamplightandindicateapartialtrade-offbetweenspectrumintensityininducingevolu-tionary trapsviaunpolarizedvspolarized lightBroad-spectrumlamps thatemphasizeallwavelengthsat relativelyhigh intensity(egsomecommercialLEDs)mayhaveveryweakpolarizedlightsignaturesbutattractaquatic insects to theirunpolarizedspec-traBecausereducing theattractivenessofevolutionary traps isthemostimportantmethodtoreducetheirdemographicimpacts(Fletcheretal2012HaleampSwearer2016RobertsonOstfeldampKeesing2017RobertsonCampbelletal2017)ourresultssug-gest thatselectionofmorenarrowband (especiallyblue-redem-phasizing)lampsbyindividualsandgovernmentswillhavegreaterconservationbenefitsformoretaxaofnocturnalaquatic insectsthaneffortstoreducetheuseofpolarizingbuildingandroadwaymaterialsnearlakesandrivers

Giventhestrongnegativefitnessconsequencesthatevolution-arytrapshaveforindividualsnaturalselectionagainstmaladaptivebehaviors shouldbe strong and so traps shouldbeephemeral intime (Schlaepfer Sherman Blossey amp Runge 2005 Delibes etal2001) Indeeda recentstudy illustrates theevolutionof reducedsensitivitytounpolarizedlightpollutionbyurbanmoths(AltermattampEbert2016)Yet theymaypersisteven in the faceofpopula-tioncollapseifuseofpolarizedlighttolocatewaterorunpolarizedlighttonavigateiscanalized(Fletcheretal2012)Theintentionalconstructionofevolutionary traps tocontrolpestspecieshas re-cently been highlighted as a powerful new wildlife managementtool(HorvaacutethBlahoacuteEgriampLerner2014RobertsonOstfeldetal2017)theprinciplesofwhichcanbeadaptedtomanipulatenativespeciesexposed toevolutionary traps intomakingadaptivedeci-sions (egHorvaacutethetal2010)Forexampleaquatic insectsareattractedupwardtothesurfacesofbridgesbyunpolarizedbridge-lightingwheretheythenovipositonhorizontallypolarizingasphaltsurfaces(Maacutelnaacutesetal2011)butdownstream-facingLEDsplacedat thewater surface canpreferentially attract insects toovipositonthewater instead(Egrietal2017)Becauseunpolarizedlightisprimarilyresponsiblefordrawingaquaticinsectsawayfromsuit-ablehabitatsasimilarstrategyofplacingabrighterormorebroad-spectrum lamp above water can be employed where residencesor businesses expose lakes and streams to artificial light Globaltrends toward the wholesale adoption of broad-spectrum lampsfor nocturnal illumination portend the expansion of evolutionarytrapsforbothterrestrialandaquaticinsectsacrosstheglobeata


timewheninsectpopulationsarealreadyundercollapse(Hallmannetal 2017) and when continent-scale evolutionary traps havebeen implicatedascausal (monarchbutterfliesFaldynHunterampElderd 2018 European honey bee Kessler etal 2015 Rundloumlfetal2015)Anestimateoftheeffectsoftheevolutionarytrapef-fectassociatedwithunpolarizedstreetlampsinGermanyindicatedthat the light couldwipe outmore than 60 billion insects over asingle summer (Eisenbeis 2006) Selectionof narrow-wavelengthlamp types where safety andor cost is allowed will lead to themostsignificantreductionsintheattractionofaquatic(thisstudyRobertsonCampbelletal2017RobertsonOstfeldetal2017)andterrestrial(PawsonandBader2014Longcoreetal2015)in-sectstononhabitatswhilemorecarefulselectionofpotentialar-tificialpolarizersnearwaterbodies(egHorvaacutethetal2010)willhaveanadditionalthoughminorbenefit

ACKNOWLEDG EMENTS

We thank the logistical support of the Estrato Research andDevelopment Ltd (Dr Andraacutes Barta and Dr Ramoacuten HegeduumlsBudapestHungary)

CONFLIC T OF INTERE S T

Wedeclarenocompetinginterests

DATA ARCHIVING S TATEMENT

Data are available from theDryadDigital Repository httpsdoiorg105061dryad7m5cv66

ORCID

Bruce A Robertson httporcidorg0000-0002-6101-1569

R E FE R E N C E S

Altermatt F amp Ebert D (2016) Reduced flight-to-light behaviour ofmothpopulationsexposedtolong-termurbanlightpollutionBiology Letters1220160111httpsdoiorg101098rsbl20160111

Barghini A amp de Medeiros B A S (2012) UV radiation as an at-tractor for insects Leukos 9 47ndash56 httpsdoiorg101582LEUKOS20120901003

BatesAJSadlerJPEverettGGrundyDLoweNDavisGhellipYoungH(2013)AssessingthevalueoftheGardenMothSchemecitizen science dataset how does light trap type affect catchEntomologia Experimentalis et Applicata 146 386ndash397 httpsdoiorg101111eea12038

BodaPHorvaacutethGKriskaGBlahoacuteMampCsabaiZ(2014)PhototaxisandpolarotaxishandinhandNightdispersalflightofaquaticinsectsdistracted synergistically by light intensity and reflectionpolariza-tion Naturwissenschaften 5 385ndash395 httpsdoiorg101007s00114-014-1166-2

Burkett D A (1998) Light Color Attraction and Dietary SugarComposition for Several Mosquito (Diptera Culicidae) Species

FoundinNorthCentralFloridaDissertationUniversityofFloridaGainesville159pp

Dacke M Nilsson D E Scholtz C H Byrne M ampWarrant E J(2003) Insect orientation to polarizedmoonlightNature424 33httpsdoiorg101038424033a

DelibesMFerrerasPampGaonaP(2001)AttractivesinksorhowindividualbehaviouraldecisionsdeterminesourcendashsinkdynamicsEcology Letters4401ndash403httpsdoiorg101046j1461-0248200100254x

DwernychukLWampBoagDA (1972)Ducksnesting inassociationwithgulls-anecological trapCanadian Journal of Zoology50559ndash563httpsdoiorg101139z72-076

Dwyer R G Bearhop S Campbell H A amp Bryant D M (2013)Shedding lighton lightBenefitsofanthropogenic illuminationtoanocturnally foraging shorebird Journal of Animal Ecology82 478ndash485httpsdoiorg1011111365-265612012

EisenbeisG (2006)Artificial night lighting and insectsAttraction ofinsectstostreetlampsinaruralsettinginGermanyInCRichampTLongcore(Eds)Ecological consequences of artificial night lighting(pp281ndash304)WashingtonDCIslandPress

EgriAacute SzaacutezD FarkasA PereszleacutenyiAacuteHorvaacutethGampKriskaG(2017)Methodtoimprovethesurvivalofnight-swarmingmayfliesnearbridgesinareasofdistractinglightpollutionRoyal Society Open Science4171166httpsdoiorg101098rsos171166

EllersJBoggsCLampMalletJ (2003)TheevolutionofwingcolorMalematechoiceopposesadaptivewingcolordivergenceincoliasbutterfliesEvolution571100ndash1106

FaldynMJHunterMDampElderdBD(2018)ClimatechangeandaninvasivetropicalmilkweedAnecologicaltrapformonarchbut-terfliesEcologyl991031ndash1038httpsdoiorg101002ecy2198

FlamariqueINampBrowmanHI(2001)Foragingandprey-searchbe-haviourofsmalljuvenilerainbowtrout(Oncorhynchus mykiss)underpolarizedlightJournal of Experimental Biology2042415ndash2422

FletcherRJOrrockJLampRobertsonBA(2012)Howthetypeofanthropogenicchangealters theconsequencesofecological trapsProceedings of the Royal Society of London B2792546ndash2552httpsdoiorg101098rspb20120139

GastonK JVisserMEampHoumllkerF (2015)Thebiological impactsof artificial light at night The research challenge Philosophical Transactions of the Royal Society of London B37020140133httpsdoiorg101098rstb20140133

vanGrunsvenRHDonnersMKoekeeKTichelaarIvanGeffenKGroenendijkDhellipVeenendaalE (2014)Spectralcompositionoflightsourcesandinsectphototaxiswithanevaluationofexistingspectral responsemodels Journal of Insect Conservation18 225ndash231httpsdoiorg101007s10841-014-9633-9

HaleRampSwearerSE(2016)EcologicaltrapsCurrentevidenceandfuturedirectionsProceedings of the Royal Society of London B28320152647httpsdoiorg101098rspb20152647

Hallmann C A Sorg M Jongejans E Siepel H Hofland NSchwan H hellip de Kroon H (2017) More than 75 percent de-cline over 27 years in total flying insect biomass in protectedareasPLoS ONE12 e0185809 httpsdoiorg101371journalpone0185809

Hegeduumls R amp Horvaacuteth G (2004) Polarizational colours could helppolarization-dependent colour vision systems to discriminate be-tween shiny and matt surfaces but cannot unambiguously codesurface orientation Vision Research 44 2337ndash2348 httpsdoiorg101016jvisres200405004

HillPSMWellsPHampWellsH (1997)Spontaneousflowercon-stancy and learning in honey bees as a function of colourAnimal Behaviour54615ndash627httpsdoiorg101006anbe19960467

HobackWWSvatosTMSpomerSMampHigleyLG(1999)Trapcolorandplacementaffectsestimatesofinsectfamily-levelabundanceanddiversityinaNebraskasaltmarshEntomologia Experimentalis et Appli cata91393ndash402httpsdoiorg101046j1570-7458199900507x


HoumllkerFMossTGriefahnBKloasWVoigtCCHenckelDhellipTocknerK(2010)ThedarksideoflightAtransdisciplinaryresearchagendaforlightpollutionpolicyEcology and Society1513[online]URLhttpwwwecologyandsocietyorgvol15iss4art13

HorvaacutethG (Ed) (2014)Polarized light and polarization vision in animal sciencesBerlinHeidelbergSpringer

Horvaacuteth G BlahoacuteM Egri A Kriska G Seres I amp Robertson B(2010)Reducing themaladaptive attractivenessof solarpanels topolarotactic insectsConservation Biology 24 1644ndash1653 httpsdoiorg101111j1523-1739201001518x

Horvaacuteth G Blahoacute M Egri A amp Lerner A (2014) Applying polar-ization-based traps to insect control InGHorvaacuteth (Ed)Polarized Light and Polarization Vision in Animal Sciences(pp561ndash584)BerlinHeidelbergSpringer

HorvaacutethGampCsabaiZ (2014)Polarizationvisionofaquatic insectsInGHorvaacuteth(Ed)Polarized light and polarization vision in animal sci-ences(pp113ndash145)BerlinHeidelbergSpringer

Horvaacuteth G Kriska G Malik P amp Robertson B (2009) Polarizedlight pollution A new kind of ecological photopollutionFrontiers in Ecology and the Environment 7 317ndash325 httpsdoiorg101890080129

Horvaacuteth GMoacutera A Bernaacuteth B amp Kriska G (2011) Polarotaxis innon-bitingmidgesFemalechironomidsareattractedtohorizontallypolarizedlightPhysiology and Behavior1041010ndash1015httpsdoiorg101016jphysbeh201106022

Horvaacuteth G amp Varjuacute D (2004) Polarized light in animal vision BerlinHeidelbergSpringerhttpsdoiorg101007978-3-662-09387-0

Johnson N F amp Triplehorn C A (2004) Borror and DeLongrsquos intro-duction to the study of insects 7th ed SilverwaterNSWAustraliaCengageLearning

KelberA(1999)WhylsquofalsersquocoloursareseenbybutterfliesNature402251httpsdoiorg10103846204

KelberAThunellCampArikawaK(2001)Polarisation-dependentco-lourvisioninPapiliobutterfliesJournal of Experimental Biology2042469ndash2480

Kessler S C Tiedeken E J Simcock K LDerveau SMitchell JSoftleyShellipWrightGA(2015)Beespreferfoodscontainingne-onicotinoidpesticidesNature521(7550)74

KokkoHampSutherlandWJ(2001)Ecologicaltrapsinchangingenviron-mentsEcologicalandevolutionaryconsequencesofabehaviourallymediatedAlleeeffectEvolutionary Ecology Research3537ndash551

Kriska G Bernaacuteth B Farkas R amp Horvaacuteth G (2009) Degrees ofpolarization of reflected light eliciting polarotaxis in dragonflies(Odonata)mayflies (Ephemeroptera) and tabanid flies (Tabanidae)Journal of Insect Physiology551167ndash1173httpsdoiorg101016jjinsphys200908013

KriskaGCsabaiZBodaPMalikPampHorvaacutethG(2006)Whydored and dark-coloured cars lure aquatic insects The attraction ofwater insectstocarpaintworkexplainedbyreflection-polarizationsignalsProceedings of the Royal Society of London B2731667ndash1671httpsdoiorg101098rspb20063500

KriskaGHorvaacutethGampAndrikovics S (1998)Why domayflies laytheir eggs en masse on dry asphalt roads Water-imitating polar-izedlightreflectedfromasphaltattractsEphemeropteraJournal of Experimental Biology2012273ndash2286

KriskaGMalikPSzivaacutekIampHorvaacutethG(2008)Glassbuildingsonriver banks as lsquopolarized light trapsrsquo for mass-swarming polarot-actic caddis flies Naturwissenschaften 95 461ndash467 httpsdoiorg101007s00114-008-0345-4[accessed2012Jun8]

Labhart T (2016) Can invertebrates see the e-vector of polarizationasaseparatemodalityoflightJournal of Experimental Biology2193844ndash3856httpsdoiorg101242jeb139899

LampharHASampKocifajM(2013)LightpollutioninultravioletandvisiblespectrumEffectondifferentvisualperceptionsPLoS ONE8e56563httpsdoiorg101371journalpone0056563

vanLangeveldeFEttemaJADonnersMWallis-DeVriesMFampGroenendijk D (2011) Effect of spectral composition of artificiallightontheattractionofmothsBiological Conservation1442274ndash2281httpsdoiorg101016jbiocon201106004

Lerner A Meltser N Sapir N Erlick C Shashar N amp Broza M(2008)Reflectedpolarizationguideschironomidfemalestoovipo-sitionsitesJournal of Experimental Biology2113536ndash3543httpsdoiorg101242jeb022277

Lerner A Sapir N Erlick C Meltser N Broza M amp Shashar N(2011)Habitatavailabilitymediateschironomiddensity-dependentoviposition Oecologia 165 905ndash914 httpsdoiorg101007s00442-010-1893-9

LongcoreTAldernHLEggersJFFloresSFrancoLHirshfield-YamanishiEhellipBarrosoAM(2015)Tuningthewhitelightspec-trumoflightemittingdiodelampstoreduceattractionofnocturnalarthropodsPhilosophical Transactions of the Royal Society of London B37020140125httpsdoiorg101098rstb20140125

Longcore T amp Rich C (2004) Ecological light pollution Frontiers in Ecology and the Environment 2 191ndash198 httpsdoiorg1018901540-9295(2004)002[0191ELP]20CO2

LorneJKampSalmonM(2007)EffectsofexposuretoartificiallightingonorientationofhatchlingseaturtlesonthebeachandintheoceanEndangered Species Research 3 23ndash30 httpsdoiorg103354esr003023

Maacutelnaacutes K Polyaacutek L Prill E Hegeduumls R Kriska G Deacutevai G hellipLengyelS(2011)Bridgesasopticalbarriersandpopulationdisrup-tors for themayflyPalingenia longicauda An overlooked threat tofreshwaterbiodiversityJournal of Insect Conservation15823ndash832httpsdoiorg101007s10841-011-9380-0

MuheimR SjoumlbergSampPinzon-RodriguezA (2016)Polarized lightmodulates light-dependentmagnetic compass orientation in birdsProceedings of the National Academy of Sciences 113 1654ndash1659httpsdoiorg101073pnas1513391113

Nowinszky L (2003) The Handbook of Light Trapping (p 276)SzombathelyHungarySavariaUniversityPress

NowinszkyLKissOSzentkiraacutelyiFPuskaacutesJampLadaacutenyiM(2012)Influenceofilluminationandpolarizedmoonlightonlight-trapcatchofcaddisflies(Trichoptera)Research Journal of Biology279ndash90

PawsonSMampBaderMKF(2014)LEDlightingincreasestheecologicalimpactoflightpollutionirrespectiveofcolortemperatureEcological Applications241561ndash1568httpsdoiorg10189014-04681

Perkin E K Houmllker F amp Tockner K (2014) The effects of artificiallightingonadult aquaticand terrestrial insectsFreshwater Biology59368ndash377httpsdoiorg101111fwb12270

RadwellAJampCampNB(2009)ComparingchemiluminescentandLEDlightfortrappingwatermitesandaquaticinsectsSoutheastern Naturalist8733ndash738httpsdoiorg1016560580080414

RandASBridarolliMEDriesLampRyanMJ(1997)LightlevelsinfluencefemalechoiceinTungarafrogspredationriskassessmentCopeia1997447ndash450httpsdoiorg1023071447770

RundloumlfMAnderssonGKBommarcoRFries IHederstroumlmVHerbertssonLhellipSmithHG(2015)Seedcoatingwithaneonico-tinoidinsecticidenegativelyaffectswildbeesNature521(7550)77

RobertsonBACampbellDRDurovichCHetterich ILesJampHorvaacutethG (2017)The interfaceofecologicalnoveltyandbehav-ioralcontextintheformationofecologicaltrapsBehavioral Ecology281166ndash1175httpsdoiorg101093behecoarx081

RobertsonBAampHuttoRL(2006)Aframeworkforunderstandingecological traps and an evaluation of existing evidenceEcology871075ndash1085

RobertsonBAKeddy-Hector IAShresthaSDSilverbergLYWoolner C E Hetterich I amp Horvaacuteth G (2018) Susceptibilityto ecological traps is similar among closely related taxa but sensi-tive to spatial isolationAnimal Behaviour 135 77ndash84 httpsdoiorg101016janbehav201710023


RobertsonBAOstfeldRSampKeesingF(2017)TrojanfemalesandJudas goats Evolutionary traps as tools in wildlifemanagementBioScience67983ndash994httpsdoiorg101093bioscibix116

RobertsonBARehageJSampSihA(2013)Ecologicalnoveltyandtheemergenceofevolutionary trapsTrends in Ecology amp Evolution28552ndash560httpsdoiorg101016jtree201304004

Schlaepfer M A Runge M C amp Sherman PW (2002) EcologicalandevolutionarytrapsTrends in Ecology and Evolution17474ndash480httpsdoiorg101016S0169-5347(02)02580-6

SchlaepferMA ShermanPWBlosseyBampRungeMC (2005)IntroducedspeciesasevolutionarytrapsEcology Letters8241ndash246

SchwindR (1991)Polarizationvision inwater insectsand insects liv-ing on amoist substrate Journal of Comparative Physiology A169531ndash540

Schwind R (1995) Spectral regions in which aquatic insects see re-flected polarized light Journal of Comparative Physiology A 177439ndash448

Somers-Yeates R Hodgson D McGregor P K Spalding A ampffrench-Constant R H (2013) Shedding light on moths Shorter

wavelengthsattractnoctuidsmorethangeometridsBiology Letters920130376httpsdoiorg101098rsbl20130376

Szaacutez D Horvaacuteth G Barta A Robertson B A Farkas A Egri Aacutehellip Kriska G (2015) Lamp-lit bridges as dual light-traps for thenight-swarmingmayflyEphoron virgo Interaction of polarized andunpolarized light pollution PLoS ONE 10 e0121194 httpsdoiorg101371journalpone0121194

WarrantEampDackeM(2011)Visionandvisualnavigationinnoctur-nal insectsAnnual Review of Entomology56 239ndash254 httpsdoiorg101146annurev-ento-120709-144852

How to cite this articleRobertsonBAHorvaacutethGColorpolarizationvisionmediatesthestrengthofanevolutionarytrapEvol Appl 201912175ndash186 httpsdoiorg101111

eva12690


Sjoumlberg amp Pinzon-Rodriguez 2016) but the predominant sourceof polarized light on the earthrsquos surface iswater bodies (HorvaacutethKriskaMalikampRobertson2009)Whensun-ormoonlightreflectsfromwatersurfacesittypicallybecomeshorizontallypolarizedwithdegreesofpolarization(d)dependingonthedarknessbrightnessofthewaterbodyandontheangleofreflection(Horvaacutethetal2009)Historicallywater-seekingaquaticinsectshavereliablyusedthehor-izontalpolarizationof lightreflectedwithahighenoughdegreeofpolarizationdtolocatewatersourcestheyrequiretofeedmateandlayeggs(reviewedinHorvaacutethampCsabai2014)Todayitiswell-knownthattheabundanceofsmoothdark-coloredman-madeobjects(egasphalt glass buildings black cars solar panels dasymp80ndash100 atand near the Brewsterrsquos angle θBrewster=arctan n measured fromthenormalvectorofthereflectingsurfacewithrefractiveindexn KriskaHorvaacutethampAndrikovics1998KriskaCsabaiBodaMalikampHorvaacuteth2006KriskaBernaacutethFarkasampHorvaacuteth2009Horvaacutethetal20092010)canpolarizelightmorestrongly(iewithhigherd)thananynaturalwaterbody(dasymp30ndash80FlamariqueampBrowman2001 Horvaacuteth 2014) andwater-seeking aquatic insects are pref-erentially attracted to oviposit upon these supernormally intensesources of horizontal polarization where their eggs fail to hatch(Horvaacutethetal2010Kriskaetal199820062009)Adults thendietypicallywithnosecondopportunitytoreproduceSuchcasesofanthropogenicchangewhereevolvedcue-resourcecorrelationsareuncoupledandanimalscometoprefer inferiorhabitatsareknownasecologicaltraps(DwernychukampBoag1972)whichareonetypeofthebroadercategoryofevolutionarytrapsinwhichanytypeofbehaviorissimilarlyaffected(SchlaepferRungeampSherman2002)Thesesevereformsofbehavioralmaladaptationarebeingdetectedat an increasing frequency (RobertsonampHutto 2006 RobertsonRehageampSih2013)areknowntoleadtosevereandrapidpopu-lationdeclines(DelibesFerrerasampGaona2001FletcherOrrockampRobertson2012KokkoampSutherland2001)andaffectabroadrange of insect taxa (Robertson etal 2013 Robertson Campbelletal2017Robertsonetal2018)

Whenplacedabovewaterbodiesorartificialpolarizers lampscreate both polarized (due to reflected light) and unpolarized (bydirectlight)photopollutionthatcanindividuallyorincombinationtriggermaladaptivebehaviorinadiverserangeofinsecttaxa(BodaHorvaacutethKriskaBlahoacuteampCsabai2014RobertsonCampbelletal2017Szaacutezetal2015)Yetspeciesdifferwidelyintheirsusceptibil-itytothesedifferentformsoflightpollution(RobertsonCampbelletal2017)andthespecificmechanismsbywhichtheseman-madelightsourcesinterfacewithevolvedbehavioralrulestotriggermal-adaptationremainpoorlyunderstood(Houmllkeretal2010Horvaacutethetal2009LongcoreampRich2004PerkinHoumllkerampTockner2014Robertson etal 2013) Artificial light sources (eg light-emittingdiodesmdashLEDshigh-pressure sodium [HPS] lamps)differ fromsun-lightandmoonlight in thatdifferent lamptypesemphasizeoren-tirelyomitparticularwavelengthsoflight(GastonVisserampHoumllker2015)Whenitreflectsfromaman-madeobjectlamplightwillnotbe polarized evenly across the visible spectrumWavelengths re-flectedwith highest intensity are polarized the least that iswith

lowestdegreeofpolarization (HegeduumlsampHorvaacuteth2004) Incon-trastapartfromthecasewhenthesunandmoonareclosetothehorizonsun-andmoonlightarerelativelyuniforminintensityacrossvisiblewavelengths(Gastonetal2015)andsopolarizationofnat-urallightviareflectionfromcolorless(ieblackgreyorwhite)waterbodieswillresultinasimilarlyalmostuniformdegreeofpolarizationacrossthevisiblespectrumCollectivelythismeansthatbyempha-sizingparticularwavelengthsoflightartificiallightpollutionhastheevolutionarilynoveleffectofdecouplingthehistoricallycorrelatedrelationship between the intensity of light and the degree of po-larizationofwater-reflectedlight Italsomeansthat ifnight-activewater-seekingflying insectsdisproportionatelyrelyupondifferenthorizontally polarized portions of the visible spectrum to locatewaterbodiesartificiallightinglocatednearartificialpolarizers(egasphaltglassbuildings)couldmakethoseobjectsmoreattractivetonocturnalovipositinginsectsthanactualwaterbodies

In this way lamps that emit different spectra of unpolarizedlighthave thecapabilityofexacerbating theseverityofecologicaltrapscausedbynocturnalpolarizedlightpollutionortoreducetheirdemographic impacts because stronger habitat preferences trig-ger ecological trapswithmore severedemographic consequences(Fletcheretal2012HaleampSwearer2016Robertsonetal2013)Global-scale shifts toward broader spectrum street lighting (egLEDs) for theirenergyefficiencyandsafetybenefitswill result inshiftsintheavailablewavelengthsofnocturnalpolarizedlightacrosslargelandscapeswithpotentialtoexacerbateormitigatemaladap-tivebehavioralresponsesinwater-seekinginsects

We designed a field experiment to ask (a) whether nocturnalwater-seekingaquatic insects relyupondifferent spectraofpolar-izedlighttolocatewaterbodiesand(b)whichlamptypes(andpor-tionsoftheirvisiblespectra) insectsarepreferentiallyattractedtoand sowhich sensory cuesmitigate or exacerbate the severity ofevolutionary traps triggered by nocturnal polarized light pollutionOur approach examines variation in riverside captures of sevenaquatic and six terrestrial night-active insect families in simulatedwater bodies (white and black oil-filled trays) illuminated by LEDmetal-halide(MH)orHPSlampsWeinterpretdifferentiallyhighercaptures among lamp types illuminating oil-filled horizontal whitetrays (full-spectrumreflectorsofunpolarized light)asanattractiontoaparticularspectrumofunpolarizedlightproducedbythatlampandreflectedbythetrayDifferentialattractiontoaparticularlamptreatment illuminatingblack trays (full-spectrumreflectorsofhori-zontallypolarizedlight)isinterpretedasevidenceofpreferentialat-tractiontoparticularspectraofhorizontalpolarizationInteractionsbetweentrayandlamptypessuggestpreferentialattractiontohori-zontallypolarizedlightofaspectrumthatdiffersfromthepreferredspectrumofunpolarizedlightandvice versaindicatingindependenceof detectionof thewavelength andpolarization of reflected lightTerrestrialtaxaarecommonlyattractedtobrightunpolarizedlights(Nowinszky2003NowinszkyKissSzentkiraacutelyiPuskaacutesampLadaacutenyi2012) andempirical studies suggest their ventral eye regiondoesnotperceivethepolarizationof lightcomingfrombelow(reviewedinHorvaacuteth2014RobertsonCampbelletal2017)butevidenceis






















3emsp |emspRESULTS

















of captures Model

χ2 values


Aquatic








Terrestrial



























ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690






















3emsp |emspRESULTS

















of captures Model

χ2 values


Aquatic








Terrestrial



























ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690











3emsp |emspRESULTS

















of captures Model

χ2 values


Aquatic








Terrestrial



























ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690



3emsp |emspRESULTS

















of captures Model

χ2 values


Aquatic








Terrestrial



























ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690












of captures Model

χ2 values


Aquatic








Terrestrial



























ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690




















ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690
















ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690











ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690



ACKNOWLEDG EMENTS






ORCID


R E FE R E N C E S










































































eva12690



















































eva12690













eva12690

Documents

Color polarization vision mediates the strength of an ...degrees of polarization (d) depending on the darkness/brightness of the water body and on the angle of reflection (Horváth