Do container size and predator presence affect Culex (Diptera: Culicidae) oviposition preferences?

REVISTA

BRASILEIRA

DE

Entomologia

w w w . r b e n t o m o l o g i a . c o m

Biology,

Ecology

and

Diversity

Do

container

size

and

predator

presence

affect

Culex

(Diptera:

Culicidae)

oviposition

preferences?

José

Junior

dos

Santos

_,

_Nádia

_Kroth

_,

_Jennifer

_A.

_Breaux

_,

_Daniel

_{Albeny-Simões}

a_{UniversidadeComunitáriadaRegiãodeChapecó,UNOCHAPECO,CursodeGraduac¸ãoemCiênciasBiológicas,Chapecó,SC,Brazil}

b_{UniversidadeComunitáriadaRegiãodeChapecó,UNOCHAPECOProgramadePósGraduac¸ãoemCiênciasAmbientais,Chapecó,SC,Brazil}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received11May2017

Accepted6November2017

Availableonline16November2017

AssociateEditor:DianaGrisales

Keywords: Aquaticinsects Behavior Ecology Eggsrafts Mosquitooviposition

a

b

s

t

r

a

c

t

Organismswithcomplexlifecyclestypicallydonotexhibitparentalcare.Hence,theabilityofadult femalestochoosequalityovipositionsitesiscriticalforoffspringsuccess.Gravidfemalesofmanyinsect taxahavethecapabilitytodetectenvironmentalconditionsinwater-holdingcontainers(e.g.,resource level,presenceofcompetitorsorpredators)andtochoosethesitesthataremostsuitableforoffspring growthanddevelopment.Mosquitoesmayalsodetectphysicalcontainercharacteristicsrelatedtowater permanencesuchassurfacearea,volume,orcontainersize,andsomespeciessuchasthoseinthegenus Culexhavebeenshowntopreferlargercontainers.However,predatorsmayalsopreferentiallycolonize largercontainers;thus,ovipositingfemalesmayfacedecisionsbasedoncuesofsitequalitythatbalance thecostsandbenefitsforoffspring.Weusedafieldexperimenttoevaluatetheoviposition preferen-cesoftwoCulexspeciesinresponsetoexperimentalcontainersizeandpredatorabundanceswithin thecontainers.Wefoundthatbothspeciesavoidedovipositinginthelargestcontainers,whichhave highabundancesofChaoborussp.anddragonflylarvae(predators).However,thecontainersizemost commonlychosenforoviposition(15-Lbuckets)alsohadhighmeanabundanceperliterofdragonfly larvae.Theseresultssuggesteitherpreynaivetéorreducedvulnerabilityofthesespeciestodragonflies comparedtoChaoborussp.Otherpotentialmechanismsfortheobservedpatternsarediscussed.

©2017PublishedbyElsevierEditoraLtda.onbehalfofSociedadeBrasileiradeEntomologia.Thisisan openaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/

).

Introduction

Organisms withcomplex life cyclesrarely provision young;

thus,theabilityofadultfemalestochoosehigh-qualityoviposition

sitesiscrucialforoffspringsuccess.Femalesofmanytaxacandetect

cuesofsitequalitysuchasresourceavailability(Singer,1986),the

degreeofintra-or interspecificcompetition(Almohamadetal.,

2010),andthepresenceofpredators(Andradeetal.,2016),and

maychoosethesitesthatmaximizeoffspringsuccess(Sih,1986;

Blausteinetal., 2004;Andradeetal.,2016).Predation pressure

canstronglyinfluenceoviposition sitechoicebygravid females

(Sih,1986;KerfootandSih,1987),particularlyforaquaticinsects

thatoccupy‘container’habitats.Thesecontainerscanbenatural

(e.g.,treeholes, bromeliadtanks)or artificial(e.g.,tires,

ceme-teryvases) and are commonly occupiedby a variety of

micro-andmacroinvertebrates, includinginsects withaquatic juvenile

stages.Mosquitoescommonlycolonizethesehabitats,andsome

∗ Correspondingauthor.

E-mail:[email protected](D.Albeny-Simões).

speciesshowstrongpreferencesregardingovipositionsitechoice

(Blausteinetal.,1995).

Mosquitoovipositioninresponsetoapredationriskisthought

tobearesultofco-evolutionbetweenmosquitopreyand

preda-tors(KerfootandSih,1987).Differencesinevolutionaryexposureto

specificpredatorsamongmosquitospeciesmaycausedifferential

responsestocues(i.e.,chemicalorphysical)ofpredatorpresence

(Blausteinetal.,1995).Further,formanymosquitospecies,thereis

atradeoffbetweencompetitiveabilityandvulnerabilityto

preda-tionthatcanproducediverseovipositionresponsestothepresence

ofpredatorsrangingfromstrongavoidance(VoneshandBlaustein,

2010)toneutral(i.e.,lackofresponse)(Andradeetal.,2016);in

somecases,mosquitoesmayevenprefersitesholdingpredators

(Albeny-Simõesetal.,2014).

In addition topredation, the physical container

characteris-ticsmayalsoinfluencemosquitoovipositionsitechoice.Container

size, surface area and depth may indicate water permanence,

wherebylargercontainershave lowerrisk ofdryingduringthe

larval development period (Reiskind and Zarrabi, 2012; Segev

etal.,2011;Burronietal.,2007).Largercontainersmayalsohave

higher resource capacity (Harrington et al., 2008), and several

https://doi.org/10.1016/j.rbe.2017.11.002

0085-5626/©2017PublishedbyElsevierEditoraLtda.onbehalfofSociedadeBrasileiradeEntomologia.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://

specieshavebeenshowntoprefercontainerswithcharacteristics

indicativeofgreater waterholdingcapacity(Wongetal.,2011;

Torrisi and Hoback,2013).However, largercontainers canalso

supporthigherabundancesofallorganisms(ConnorandMcCoy,

2001),includingcompetitorsandpredators(Sunaharaetal.,2002),

andpredatorstypicallyhavelongerdevelopmenttimesandmay

preferentially occupy larger containers (Sunahara et al., 2002).

Mosquitoesmayavoidovipositinginlargercontainerstoreduce

theriskofpredationforoffspring(ReiskindandZarrabi,2012).It

isalsoexpectedthatforspecies withrapidlarval development,

container size likely does not influence oviposition

preferen-ces tothe same extentas in species withlonger development

periods.

Weusedafieldexperimenttoevaluatetheoviposition

prefer-encesbasedoncontainersizeintwo,locallycommonCulexspecies

(CulexeduardoiandCulexsp.1).Wethenassessedtherelationships

betweencontainersizeandtheabundancesofcommonpredators.

Wehypothesizedthatpredatorabundanceswouldbehighestin

largercontainers,andthatCulexfemaleswouldavoidovipositing

inlargercontainersduetohighpredatorabundances.

Methods

Studyarea

This experiment was conducted at the Floresta National de

Chapecó (27◦0639S;52◦4511W)in Guatambú, SantaCatarina,

Brazil.Thevegetationiscomposedmainlyofmixedombrophilous

forestandlieswithintheAtlanticforestbiome,withatotalarea

of 1604.35ha.The areahashumid subtropical climatewithan

averageannualrainfallaround1600mmandaveragemonthly

tem-peratures of 22◦C in the summer(28.9◦C max) and 10.4◦C in

thewinter(4.4◦Cmin).Previousobservationsrevealedadiverse

micro-andmacroinvertebratecommunityoccupyingartificial

con-tainerinhabitants,withoverthirteentaxaandatleast15species.

The majority of mosquitoes that colonize containers belongto

the genus Culex, where the most dominant mosquito species

is C. eduardoi, followed by Culex sp. 1 (Albeny-Simões, pers.

obs.).

Experimentaldesign

Experimentalmicrocosms(blackplasticcontainers)were

estab-lished along four, 300-m individual transects. Each transect

contained1-L(n=8),15-L(n=4),100-L(n=2)and200-L(n=1)

con-tainersplaced20mapart.Becausethesmallestcontainers(1L)are

expectedtobemorevariableovertime,weincludedahigher

num-berofcontainerreplicatesforsmallercontainers(n=8pertransect,

24total).Wethencollected8,4and2sub-samplesfromthe200-L

(n=4),100-L(n=8)and15-L(n=16)containers,respectively.The

positionof containersalong thetransectswasdesignedtokeep

aminimaldistancebetweensame-sizedcontainersandtoensure

roughlyequalprobabliltyofadultfemalemosquitoencounterwith

thevariouscontainersizes.Thisresultedinadistanceof40,80,

160, and320mbetweenthe1-Lcups,15-Lbuckets, 100-Land

200-Lbarrels, respectively, along each pairof contiguous

tran-sects(Fig.1).Inordertoassurecolonizationbymicrobiota,0.3g/L

offield-collected,mixed-species,previouslydriedleaflitterwas

addedtoeachcontainer.Weusedtapwatertofillthecontainers

to70%maximumvolume.Thewaterlevelwasnotmanipulated

thereafter,andthecontainerswerepermittedtoundergonatural

colonizationbyinvertebrates.Aftera15-daycolonizationperiod,

thecontainersweresampledeverytwoweeksfor105days(n=7

samplingperiods). 1.0 0.8 0.6 0.4 0.2 0.0 15 30 45 75 90 105 120 Sampling day Mean cule x egg r afts (log+1) 1L - b 15L - a 100L - a 200L - ab b b b ab ab ab a

Fig.1. Numbersofeggsraftsnumberasafunctionofsamplingdaysandcontainer

size.The100and200-Lbarrelsarerepresentedbyclosedtrianglesandclosedcircles,

respectively.Theopencirclesrepresentthe15-Lbuckets,andasterisksrepresent

the1-Lcups.Thebarsrepresentstandarderror.Differentlettersindicatedifferent

meansamongthetreatments.

Eggraftsampling

Eggraftsweresearchedforbydirectobservationofthewater

surfaceincontainers.Usingawhiteplasticspoon,allsurfaceegg

raftswerecollectedfromthewatersurfaceandstoredindividually

inEppendorftubes.Eachcontainerwasinspectedfor10min.The

eggraftswerethentransportedtotheEntomologyLaboratoryat

theCommunityUniversityoftheChapecóRegioninChapecó,SC,

Brazil.Theeggraftswereinspectedunderastereoscopic

micro-scope, and those already hatched were not considered in the

analysis.Forspeciesidentification,theraftswereplaced

individ-uallyin250-mLNalgene®plasticcontainersholding100mLoftap

water.Afterhatching,thelarvaewererearedtofourthinstarbya

singlefeedingwith0.02gofTetraFin® goldfishflakes.Forlarval

identification,thefourth-instarlarvaeweremountedonslidesand

identifiedfollowingthemethodsdescribedbyForattini(1965).

Predatorsampling

Thepredators inthecontainers weresampledby sievingall

water contentfromhalf ofthe1-Lcups everysamplingperiod

(i.e.,destructivesampling)andbysub-samplingtheothercontainer

sizes.The15-Lbucketsweresampledusingacoffeefiltermadewith

cloth,andthe100and200-Lbarrelsweresampledusingaconical,

52cmheight×18cmradiusplanktonnet.Thepredatorscollected

inthesampleswereplacedina500-mLplasticcontainerandfixed

in70%ethanolforlateridentificationtothelowestpossible

taxo-nomiclevel.

Dataanalysis

Weperformarepeated-measuresANOVAtoevaluatetheeffects

ofcontainersize,samplingdayandtheinteractiononthe

num-berofCulexeggrafts(bothspecies)andthoseofC.eduardoiand

Culexsp.1individually.Wealsotestedforeffectsofpredator

abun-dancesofthetwomostcommontaxa(dragonflyandChaoborussp.

larvae)combined,thenfortheeffectsofthesetaxaindividually.

WethenusedAIC-basedmodelselection((Table3)todetermine

theimportanceofthemaintreatmenteffects(containersizeand

sampleperiod)andimportantpredatorytaxa(dargonflylarvaeand

Chaoborussp.)bothcombinedandseparatedonthenumbersofegg

raftsofbothmosquitospecies(combinedandseparately).Foreach

analysis,theresponse variablewastestedfor normalityusinga

Shapiro–Wilktest.Non-normaldatawerelog-transformedinorder

toadjustthedatadistribution.Allstatisticalanalyseswerecarried

Table1

Repeated-measuresANOVAindicatingtheeffectsofsite,timeandsite:time

inter-actionsontheabundanceofCulexeggrafts,thetwomainpredatorscombined,and

individualeffectsofdragonfliesandChaoborussp.

Abundance Size (df=1,3) Sampleday (df=1,6) Size:day (df=3,6) F-value p-value F-value p-value F-value p-value Culexeggrafts 16.95 <0.001 4.44 <0.001 3.57 <0.001 C.eduardoi 21.45 <0.001 1.18 0.31 – – Culexsp.1 4.13 <0.05 1.98 0.06 – – Predators(Total) 16.43 <0.001 9.49 <0.001 – – Dragonflies 15.46 <0.001 6.97 <0.001 – – Midges 29.64 <0.001 0.98 0.43 – – Results

Atotalof78eggraftsweresampledduringtheentire

exper-imental period.The sampled eggrafts yielded onaverage10.3

mosquitolarvae.Twenty-fourpercentwasidentifiedasCulexsp.1

(21eggrafts)andtheremaining76%belongedtoC.eduardoi(57egg

rafts).Wesampled775predatorsdividedamong5taxa.Chaoborus

sp.anddragonflies(Odonata:Libellulidae)werethemostabundant

taxawith224and542individuals,respectively.Theremaining9

individualsincludedDysticidae,Hydrophilidae(bothColeoptera)

andToxorhynchitessp.(Diptera:Culicidae).

EffectsofcontainersizeandsamplingdayonCulexoviposition

Container size, sampling day and the size*day interaction

significantly affected thetotal number of ovipositedeggs rafts

(Table 1).Eggraft abundancepeaksonthe75thsamplingday,

mostlyin100-Lbarrelsand15-Lbuckets(Fig.1).Onthe15th,45th,

90thand120thsamplingdays,the15-Lbucketscontainedthe

high-estnumbersofeggrafts(Fig.1).Wethenanalyzedtheeffectsof

themodelvariablesonC.eduardoiandCulexsp.1separately,and

foundaninteractionbetweencontainersizeandsamplingdayforC.

eduardoi.However,Culexsp.1wasaffectedonlybycontainersize

(Table1)andwasneverfoundinthe200-Lbarrelsand1-Lcups

(Fig.2).

Effectsofcontainersizeandsamplingdayonpredatorabundance

Thetotalpredatorabundanceperliterwasstronglyaffectedby

sizeandday(Table1).Thetwomainpredatorsweremore

abun-dantandstatisticallyequallyabundantinthe15-Lbuckets,100and

200-Lcontainersbeginningonthe45thsamplingday,andslightly

increaseduntilthelastsamplingday.Incontrast,the1-Lcupshad

lowpredatorabundanceperliterduringtheentireexperimental

period(Fig.3).Thetwomostabundantpredatorytaxa(dragonflies

andChaoborussp.)whenanalyzedseparatedfromthemaindataset

werestronglyandoppositelyaffectedbysizeandthesize:day

inter-action(Table1).Whiledragonflyabundancewashighestinthe15-L

bucketsfollowedby100-Lbarrels,Chaoborussp.larvae

preferen-tiallycolonized200-Lbarrelsandwererarelyfoundin1-Lcontainer

and15-Lbuckets,andshowedlowabundancein100-Lbuckets

(Fig.4).

EffectsofthemainpredatorsonCulexoviposition

Thetotalnumbersofpredators(dragonflies+Chaoborussp.),did

notsignificantlyaffectC.eduardoiandCulexsp1.ovipositioninthe

experimentalcontainers(Table2).However,whenweanalyzedthe

presence/absenceofdragonflyandChaoborussp.larvaeseparately,

wefoundaneffectofdragonfliesonoverallnumbersofCulexegg

rafts(Table2).Noeggswerefoundinthe1-Lcups. Thebarrels

(100and200L),whichheldhighabundancesofChaoborussp.

lar-vae,receivedstatisticallyequivalentnumbersofeggraftsasthe

predator-freebarrels.Interestingly,15-Lbucketsholding

dragon-flylarvaereceivestatisticallymoreeggraftsthanthepredator-free

containers(Fig.5).

Discussion

OurresultsdidnotsupportourhypothesisthatCulexwouldlay

moreeggsinthesmallestcontainerstoavoidpredation(Fig.1).

However,theresultsshowthatfewereggraftswerelaidinthe

200-Lbarrels and 1-Lcups compared to the100-Lbarrels and

15-Lbuckets.Althoughtheydidnotstatisticallydifferfrom

100-Lbarrels,the15-Lcapacitybucketscontainedthehighestnumbers

of egg raftsonfour out of sevensampling days. Chaoborussp.

larvaehadlowabundanceperliterinthe15-Lbucketsandthe

1-Lcupsandshowedsignificantlyhigherabundanceperliterin

the200-Lbarrelscomparedtotheothercontainersizes(Fig.4).

Culexsp.1eggraftswerenotfoundinthe200-Lbarrels.

Preda-toravoidancebehavior hasbeencommonlydescribed forCulex

species(Vonesh and Blaustein,2010;Andradeet al.,2016).For

example,predatoravoidanceovipositionbehaviorhasbeen

doc-umentedinCulexmollis(anAtlanticforestspecies)inresponseto

chemicalcuesofToxorhynchitesthobaldi(Diptera:Culicidae)

preda-torylarvae(Andradeetal.,2016).ThissuggeststhatadultCulex

sp.1females maybeabletodetectenvironmentalindicatorsof

predatorpresenceinsidetheovipositioncontainers(Bentleyand

Day,1989).Ourexperimentaldesigndidnotruleoutthe

possibil-itythateggpredationimpactedourcountsofeggraftsoverthe

1.0 0.8 0.6 0.4 0.2 0.0 1.0 0.8 0.6 0.4 0.2 0.0 15 30 45 60 75 90 105 15 30 45 60 75 90 105 Days Days Cule x sp1 egg r afts (log+1) C . eduardoi egg r afts (log+1) 100L 15L 1L 200L 100L 15L 1L 200L

Fig.2. NumbersofCulexsp.1(left)andC.eduardoi(right)eggraftsasafunctionofsamplingdayandcontainersize.The100and200-Lbarrelsarerepresentedbyclosed

trianglesandclosedcircles,respectively.Theopencirclesrepresentthe15-Lbuckets,andasterisksrepresentthe1-Lcups.Thebarsrepresentstandarderror.Differentletters

2.0 1.5 1.0 0.5 0.0 15 30 45 75 90 105 120 Sampling day 1L - b 15L - a 100L - a 200L - ab C C bc ab b ab a Mean predator ab undance/L (log+1)

Fig.3. Predatorabundanceperliterasafunctionofsamplingdayandcontainersize.

The100and200-Lbarrelsarerepresentedbytheclosedtrianglesandclosedcircles,

respectively.Theopencirclesrepresentthe15-Lbuckets,andasterisksrepresent

the1-Lcups.Thebarsrepresentstandarderror.Differentlettersindicatedifferent

meansamongthetreatments.

2.0 1.5 1.0 0.5 0.0 2.0 1.5 1.0 0.5 0.0 15 30 45 75 90 105 120 15 30 45 75 90 105 120 Mean dr agonfly ab undance/L (log+1) Mean midge ab undance/L (log+1) Sampling day Sampling day 1L - b 15L - b 100L - b 200L - a 1L - c 15L - a 100L - ab 200L - bc C C bc b b ab a

Fig.4. Theabundanceofdragonflylarvaeperliter(top)andmidge(Chaoborussp.)

larvaeperliter(bottom)asafunctionofsamplingdayandcontainersize.The100

and200-Lbarrelsarerepresentedbytheclosedtrianglesandclosedcircles,

respec-tively.Theopencirclesrepresentthe15-Lbuckets,andasterisksrepresentthe1-L

cups.Thebarsrepresentstandarderror.Differentlettersindicatedifferentmeans

amongthetreatments.

experimentalperiod;however,dragonflylarvaearewell-known

toassociatewiththebottomportionandsidesofwater-holding

containers,occupyinga space thatis farfromsurface eggrafts

(Mooreetal.,1994);anddragonflylarvaehavenotbeenreportedas

mosquitoeggpredatorsintheliterature.Thereisalsonoindication

thatChaoborussp.consumemosquitoeggs.Instead,thepreferred

foodsourcesofChaoboridae,inadditiontolarvaeofmosquitoes

andotherorganisms(e.g.,copepods),arezooplanktonand

phyto-plankton.Thedescribedfeedingbehaviorsofthesetaxatherefore

suggestthattheabundanceofCulexeggraftsreflectsadultfemale

ovipositionpreferencesinsteadofpost-colonizationprocessessuch

aseggpredation.

Themeanpredatorabundanceperliterwasaffectedby

samp-lingdayandcontainersize,withnostatisticaldifferenceamong

the15-Lbucketsand100and200-Lbarrels(Fig.3).Thenumberof

C.eduardoieggraftsdidnotstatisticallydifferamong15-Lbuckets

andthe100and200-Lbarrels(Fig.2).ThissuggeststhatC.eduardoi

femalesdonotrespondtothehigheroccurrenceofpredatorsin

Table2

Repeated-measuresANOVAindicatingtheeffectsofthepresence/absenceofthe

twomainpredatorscombined,anddragonflylarvaeandChaoborussp.individually

onnumbersofCulexeggrafts(combinedandindividually).

Numbersofeggrafts Totalpredators Dragonflies Phantommidge F-value p-value F-value p-value F-value p-value TotalCulex 3.33 0.06 5.42 0.02 0.91 0.33 Culexeduardoi 1.62 0.20 3.01 0.08 0.16 0.68 Culexsp.1 1.45 0.22 1.69 0.19 0.97 0.32

Table3

AIC-basedmodelselectioncriteria.Themodel(withorwithouttheinteraction)was chosenbasedonthelowestAICvalue,andthebestmodelsforeachoftheresponse variablesarerepresentedinbold.

Responsevariable Nointeraction model(AIC)

Interactionmodel (AIC)

Culexeggrafts 180.53 169.23

C.eduardoieggrafts 98.51 90.00

Culexsp.1 −112.46 −81.38

Predators(total) 681.38 658.73

Dragonflies 619.23 622.33

Midges 41.77 67.39

Akaikeinformationcriterion(AIC)isanestimatoroftherelativequalityofstatistical modelsforagivensetofdata.

0.5 0.6 0.4 0.3 0.2 0.1 0.0 1L 15L 100L 200L Mean cule x egg r afts (log+1) No Yes

Dragonfly larvae presence

a

b

Fig.5.TotalnumberofCulexeggraftsnumberasafunctionofdragonflylarvae (predator)presence/absenceintheexperimentalcontainers.The100and200-L barrelsarerepresentedbytheclosedtrianglesandclosedcircles,respectively.The opencirclesrepresentthe15-Lbuckets,andasterisksrepresentthe1-Lcups.The barsrepresentstandarderror.Differentlettersindicatedifferentmeansamongthe treatments.

bucketsandbarrels(Fig.3)whenchoosingovipositionsites(Fig.2).

Theabilityofapreyspeciestorecognizepredatorsisaproduct

ofco-evolution(Sih,1986;KerfootandSih,1987).Predator–prey

coexistenceisthuscontingentupontheabilitytosharethesame

environmentwithnaturalenemies(KingandHastings,2003),and

this is oftenfacilitated by behavioral changes in preyafter the

detectionofpredationrisk(Blausteinetal.,1995).

C.eduardoifemalesovipositedintocontainerswherethe

abun-danceofdragonfliesperliterwashigh.Dragonflylarvaedisplaya

“sitandpursue”predatorybehaviorinwhichpreyarecapturedby

chanceduetoproximity(Pierce,1988;Gavinetal.,2007).Culex

lar-vaeandpupaeareknowntobefastswimmersandtospendmore

timerestingatthesurfaceportionofthewatercolumn(Andrade

etal.,2016)comparedtoothergenera(Roberts,2017;Brackenbury,

2001),andthesebehaviorsmayresultinlowervulnerabilityto

pre-dation.Forthisreason,dragonflylarvaemightnotposeathreatto

C.eduardoi,asthisspeciesmaybemoreabletoescapepredation

comparedtoothertaxa.Thismayfacilitateco-existenceandlead

tolackofovipositionavoidancebyCulexfemalesintocontainers

Onmostsamplingdays,Chaoborussp.larvaeabundanceperliter

washighestinthe200-Lbarrels(Fig.4)andwereconcentrated

onthemedianandsurfacepartofthewatercolumn(pers.obs.).

Amongthelifehistorystrategies ofmosquitoes toenhance

off-springsuccess,gravidfemalesmayavoidlayingeggsinharsh(e.g.,

predator-rich)environments(Blausteinetal.,1995;Eitametal.,

2002;SilberbushandBlaustein,2008;Andradeetal.,2016),or

off-springmayreducepredationriskbyalteringtheirbehaviorwhen

exposed topredators or cuesof predator presence (Kesavaraju

and Juliano, 2004; Kesavaraju et al., 2011; Andrade et al., 2016).

TheabsenceorreducednumberofChaoborussp.larvaeperliter

insmaller-sized containers(1-Lcups and 15-Lbuckets) (Fig.3)

may be explained by the behavior of adult females choosing

largercontainers,whichtheoreticallyhavegreaterfoodavailability

(othermosquitospeciesorotheranimaltaxa)(ConnorandMcCoy,

2001),orduetodragonflypredatorsoccupyingthesecontainers

anddepletingChaoboruslarvalpopulations.Competitionbetween

predatorsatupperandintermediatetrophiclevelscanalsobenefit

preypopulationsbydecreasingpredationpressureonpreyatlower

levels(Strong,1992;CostaandVonesh,2013;Albeny-Simõesetal.,

2015).

OurresultscollectivelysuggeststhatCulexsp.1avoid

oviposi-tioninthelargestcontainers,andthatC.eduardoispreadtheiregg

raftsamongintermediateandlargecontainers.Aswepredicted,

containersizewaspositivelycorrelatedwithpredatorabundance

(perliter)forChaoborussp.larvae;however,dragonflylarvaein

thisstudyoccupiedallcontainers,withthelowestabundancesand

mostfrequentabsenceinthesmallestcontainers.Theabundance

perliterofbothCulexspecieswasaffectedbythesamplingdate,

andintermediate-sizedcontainers(15and100-L)werepreferred

bydragonflies.

Thereareseveralwaysinwhichcontainersizemayaffectadult

mosquitoovipositionchoices.First,largercontainerscanbe

indica-tiveorgreater resource availability (Connor andMcCoy, 2001),

which should positively influence larval development. Second,

largerwaterbodiesmayindicategreaterwaterpermanenceand

reducedrisk ofhabitatdrying.Mosquitoes mayperceive

physi-calcharacteristicsofahabitatthroughtactile,olfactoryorvisual

cues(BentleyandDay,1989;Beehleretal.,1993)relatedto

sur-faceareaor water depth(Reiskind andZarrabi, 2012),color or

material (Bartlett-Healy et al., 2012), water surface reflectance

(Harringtonetal.,2008)orcontainerheight(Obenaueretal.,2009),

amongothercharacteristics.Differentmosquitospeciesshow

dif-ferentovipositionpreferencesbasedoncontainercharacteristics

thatareindicativeofpredationrisk(Sunaharaetal.,2002),which

ispositivelycorrelatedwithhabitatsize(Juliano,2009).Our

exper-imentaldesigndoesnotallowustodiscernwhichcuesrelated

tocontainersizemayhave influencedtheobserved patternsof

oviposition;however,multiplecontainercharacteristicsarelikely

usedbymosquitoeswhenprocessinginformationaboutpotential

breedingsites(BentleyandDay,1989).

Theresultsfromthis study contributetoourunderstanding

ofhow predators affectmosquito colonizationprocesses inthe

natural environment. Understanding the relationshipsbetween

physical and biological conditions in aquatic habitats and the

resultingmosquitoovipositionsitechoiceisoffundamental

impor-tanceforourgeneralunderstandingofhowhabitatcharacteristics

influence mosquito colonizaiton and spatial and temporal

dis-tributions under natural conditions. Our results suggest that

Culexovipositionpreferencesaremediatedbycontainersizeand

samplingperiod,perhapsthroughtheinfluenceofthesefactors

on the presence and abundance of predators. Further studies

aimedare neededto evaluatehow aquatic habitat

characteris-tics (e.g.,predators, nutrient levels, physical aspects) influence

mosquitoovipositionchoiceandtobettercharacterizehowthe

environmentaland physical characteristicsof breedinghabitats

influencemosquitooccurrencepatternsandresultinglarval

abun-dances.

Conflictsofinterest

Theauthorsdeclarenoconflictsofinterest.

Acknowledgements

WethanktoCNPqforthefinancialsupport.

References

Albeny-Simões,D.,Murrell,E.G.,Elliot,S.L.,Andrade,M.R.,Lima,E.,Juliano,S.A.,

Vilela,E.F.,2014.Attractedtotheenemy:Aedesaegyptiprefersovipositionsites

withpredator-killedconspecifics.Oecologia175,481–492.

Albeny-Simões,D.,Murrell,E.G.,Vilela,E.F.,Juliano,S.A.,2015. Amultifaceted

trophiccascadeinadetritus-basedsystem:density-,trait-,or processing-chain-mediatedeffects?Ecosphere6,art32.

Almohamad,R.,Verheggen,F.J.,Francis,F.,Lognay,G.,Haubruge,E.,2010.

Assess-mentof oviposition site qualityby aphidophagous hoverflies:reactionto conspecificlarvae.Anim.Behav.79,589–594.

Andrade, M.R., Albeny-Simões, D., Breaux, J., Juliano, S., Lima, E., 2016. Are

behaviouralresponsestopredationcueslinkedacrosslifecyclestages?Ecol. Entomol.41,1–7.

Bartlett-Healy,K.,etal.,2012.Larvalmosquitohabitatutilizationandcommunity

dynamicsofAedesalbopictusandAedesjaponicus(Diptera:Culicidae).J.Med. Entomol.49(4),813–824.

Beehler,J.W.,Millar,J.G.,Mulla,M.S.,1993.Synergismbetweenchemicalattractants

andvisualcuesinfluencingovipositionofthemosquito,Culexquinquefasciatus (Diptera:Culicidae).J.Chem.Ecol.19(4),635–644.

Bentley,M.D.,Day,J.F.,1989.Chemicalecologyandbehavioralaspectsofmosquito

oviposition.Annu.Rev.Entomol.34,401–421.

Blaustein,L.,Kiflawi,M.,Eitam,A.,Mangel,M.,Cohen,J.,2004.Ovipositionhabitat

selectioninresponsetoriskofpredationintemporarypools:modeofdetection andconsistencyacrossexperimentalvenue.Oecologia138,300–305.

Blaustein,L.,Kotler,B.,Ward,D.,1995.Directandindirecteffectsofapredatory

backswimmerNotonectamaculataoncommunitystructureofdeserttemporary pools.Ecol.Entomol.20,311–318.

Brackenbury,J.,2001.Thevortexwakeofthefree-swimminglarvaandpupaofCulex

pipiens(Diptera).J.Exp.Biol.204.11,1855–1867.

Burroni,N.,etal.,2007.NewrecordandlarvalhabitatsofCulexeduardoi(Diptera:

Culicidae)inanirrigatedareaofPatagonia,ChubutProvince,Argentina.Mem. Inst.OswaldoCruz102(2),237–240.

Connor,E.F.,McCoy,E.D.,2001.Species-arearelationships.Encycl.Biodivers.5,

397–411.

Costa,Z.J.,Vonesh,J.R.,2013.Preysubsidyorpredatorcue?Directandindirect

effectsofcagedpredatorsonaquaticconsumersandresources.Oecologia173, 1481–1490.

Eitam,A.,Blaustein,L.,Mangel,M.,2002.EffectsofAnisopssardea(Hemiptera:

Notonectidae)onovipositionhabitatselectionbymosquitoesandother dipter-ansandoncommunitystructureinartificialpools.Hydrobiologia485,183–189.

Forattini,O.P.,1965.EntomologiaMédicaCulicini:Culex,AedesePsorophora.Editora

daUniversidadedeSãoPaulo,pp.2.

Gavin,F.,Volker,H.,Rudolf,W.,2007.Responsesoflarvaldragonfliestoconspecific

andheterospecificpredatorcues.Ecol.Entomol.32(3),283–288.

Harrington,L.C.,etal.,2008.Influenceofcontainersize,location,andtimeofday

onovipositionpatternsofthedenguevector,Aedesaegypti,inThailand. Vector-BorneZoonoticDis.8.3,415–424.

Juliano, S.A., 2009. Species interactions among larval mosquitoes: context

dependence across habitat gradients. Ann. Rev. Entomol. 54, 37–56,

http://dx.doi.org/10.1146/annurev.ento.54.110807.090611.

Kerfoot,W.C.,Sih,A.(Eds.),1987.Predation:DirectandIndirectImpactsonAquatic

Communities.Univ.PressofNewEngland,Hanover,NH,pp.203–224(Chapter: Predatorandpreylifestyles:anevolutionaryandecologicaloverview).

Kesavaraju,B.,Juliano,S.A.,2004.Differentialbehavioralresponsestowater-borne

cuestopredationintwocontainer-dwellingmosquitoes.Ann.Entomol.Soc.Am. 97,194–201.

Kesavaraju, B.,Khan, D.,Gaugler, R., 2011. Behavioraldifferencesof invasive

container-dwellingmosquitoes to anative predator.J. Med. Entomol.48, 526–532.

King,A.A.,Hastings,A.,2003.Spatialmechanismsforcoexistenceofspeciessharing

acommonnaturalenemy.Theor.Popul.Biol.64,431–438.

Moore,M.,Yan,N.D.,Pawson,T.,1994.Omnivoryofthelarvalphantommidge

(Chaoborusspp.)anditspotentialsignificanceforfreshwaterplanktonicfood webs.Can.J.Zool.

Obenauer,P.J.,etal.,2009.Infusion-baitedovitrapstosurveyovipositionalheight

preferencesofcontainer-inhabitingmosquitoesintwoFloridahabitats.J.Med. Entomol.46(6),1507–1513.

Pierce,C.L.,1988.Predatoravoidance,microhabitatshift,andrisk-sensitiveforaging

RDevelopmentCoreTeam, 2014.R:ALanguageand Environmentfor Statis-ticalComputing. R Foundation for StatisticalComputing, Vienna, Austria, pp.2012.

Reiskind,M.H.,Zarrabi,A.A.,2012.Watersurfaceareaanddepthdetermine

ovipo-sitionchoicein Aedesalbopictus(Diptera: Culicidae).J. Med.Entomol.49, 71–76.

Roberts,D.,2017.Mosquitolarvaecandetectwatervibrationpatternsfromanearby

predator.Bull.Entomol.Res.,1–7.

Segev,O.,etal.,2011.Spatiotemporalreproductivestrategiesinthefiresalamander:

amodelandempiricaltest.Behav.Ecol.22(3),670–678.

Sih,A.,1986.Antipredatorresponsesandtheperceptionofdangerbymosquito

larvae.Ecology67,434–441.

Silberbush,A.,Blaustein,L.,2008.Scientificnoteovipositionhabitatselectionbya

mosquitoinresponsetoapredator:arepredator-releasedkairomonesair-borne cues?J.VectorEcol.33,208–211.

Singer,M.C.,1986.Thedefinitionandmeasurementofovipositionpreferencein

plant-feedinginsects.In:MILLER,J.R.,MILLER,T.A.(Eds.),Insect–Plant Interac-tions.Springer-Verlag,NewYork,pp.66–94.

Sunahara,T.,Ishizaka,K.,Mogi,M.,2002.Habitatsize:afactordeterminingthe

opportunityforencountersbetweenmosquitolarvaeandaquaticpredators.J. VectorEcol.27,8–20.

Strong,D.,1992.Aretrophiccascadesallwet?Differentiationanddonor-controlin

speciesecosystems.Ecology73,747–754.

Torrisi,G.J.,Hoback,W.W.,2013.Colorandcontainersizeaffectmosquito(Aedes

triseriatus)oviposition.NortheasternNat.20(2),363–371.

Vonesh,J.R.,Blaustein,L.,2010.Predator-inducedshiftsinmosquitoovipositionsite

selection:ameta-analysisandimplicationsforvectorcontrol.Isr.J.Ecol.Evol. 56,263–279.

Wong,J.,etal.,2011.OvipositionsiteselectionbythedenguevectorAedesaegypti