Factors affecting variation in the reproductive investment of female treefrogs. Hyla intermedia

(1)

Other uses, including reproduction and distribution, or selling or

licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the

article (e.g. in Word or Tex form) to their personal website or

institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are

encouraged to visit:

(2)

ContentslistsavailableatSciVerseScienceDirect

Zoology

j o u rn a l h o m e pa g e:ww w . els e vi e r . co m / l o c a t e / z o o l

Factors

affecting

variation

in

the

reproductive

investment

of

female

treefrogs,

Hyla

intermedia

Giorgia

Cadeddu

∗

,

Sergio

Castellano

DipartimentodiBiologiaAnimaleedell’Uomo,UniversitàdiTorino,ViaAccademiaAlbertina,13,10123Torino,Italy

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received17January2012

Receivedinrevisedform20March2012 Accepted16April2012 Keywords: Hylaintermedia Eggsize Clutchsize Fecundityvariation Breedingseason

a

b

s

t

r

a

c

t

Inanurans,fecundity(clutchsize)isthemostimportantdeterminantoffemalereproductivesuccess. WeinvestigatedthreepossiblecausesresponsibleforfecundityvariationinfemaleItaliantreefrogs,Hyla intermedia,duringfourbreedingseasons:(i)variationinmorphological(bodysizeandcondition)and life-history(age)traits;(ii)variationinthetradeoffbetweenthenumberandthesizeofeggs;(iii)seasonal effectsandwithin-seasondifferencesinthetimingofdeposition.Atthepopulationlevel,wefoundno evidenceforatradeoffbetweenthenumberandthesizeofeggs,becausetheybothcorrelatedpositively withfemales’bodysize.Conversely,neitheragenorpost-spawningbodyconditionshowedanyeffect onfemalereproductiveinvestment.Independentofbodysize,wefoundnoevidenceforvariationin reproductiveeffortamongdifferentbreedingseasons,butstrongevidenceforadecreaseofclutchsize andanincreaseofeggsizewiththeadvancingofabreedingseason.Totestforthefunctionalsigniﬁcance oftheobservedtemporalvariationinallocationstrategy,wecarriedoutarearingexperimentin semi-naturalconditionsonarandomsampleoftenclutches.Theexperimentshowedanegativeeffectofclutch sizeandapositiveeffectofeggsizeonbothtadpolegrowthanddevelopmentalrates,suggestingthat reproductiveinvestment,althoughconstrainedbybodysize,canbeadjustedbyfemalestothetimeof depositiontoincreasethechancesofoffspringsurvival.

1. Introduction

Inspecieswithoutparentalcare,femalefitnesslargelydepends onfecundityandfactorsresponsibleforintra-specificvariationin fecundity areexpected tobeunder strongdirectionalselection (Stearns,1992;Roff,2002).Variationinfecundityarisesfromthe interactionof morpho-physiological constraints and life-history tradeoffs.Morpho-physiologicalconstraintsaffecttheamountof resourcesthatfemalescanacquirefromtheenvironment,store intheirbody,andinvestingameteproduction.Life-history trade-offs represent the way females cope withthese constraints in order to optimize their lifetime reproductive success (Stearns, 1992).Theunderstandingofwhatcausesintra-specificvariation infemalefecunditymaygivevaluableinsightsintotheselective forcesresponsiblefortheevolutionoffemalelife-historystrategies. Inanurans,aswellasinallanimalswithindeterminategrowth, bodysizeisoneofthemostimportantmorphologicalconstraints offecundity,becauseofitsallometricrelationshipwithovarymass (Wells, 2007).However,body sizeis alsotheexpressionof the tradeoffbetweengrowth,developmentandreproduction(Cadeddu etal., 2012).Femalesthat investlargeamountsof resourcesin

∗ Correspondingauthor.Tel.:+390116704559. E-mailaddress:giorgia.cadeddu@tiscali.it(G.Cadeddu).

growthwouldnecessarilyhavefewerresourcestoinvestin repro-ductionandtheywouldattainlargebodysize,usually,bydelaying sexualmaturity(MonnetandCherry,2002;Cadedduetal.,2012). Inthesespecies,size-dependentdifferencesinfecunditymaythus arise fromdifferencesinageatmaturity, thatis,fromdifferent tradeoffsbetweentheexpectednumberofreproductiveepisodes andthereproductivesuccessofeachepisode(Shine,1988). Fur-thermore, age may directly affect variation in fecundity if the proportionofresourcesallocatedtoreproduction,ratherthanto growth and self-maintenance, increases withage(Gibbons and McCarthy,1986;Berven,1988).

Variation in fecundity, however, may be independent of life-historytradeoffsandbetheconsequenceofamong-female dif-ferencesintheamountofavailableresources(Jörgensen,1992). Some femalesmay begenetically more efﬁcientthan others in convertingexternalresourcesintometabolicenergyortheymay experiencedifferentenvironmentalconditions,becauseofthe spa-tialorseasonalheterogeneityofpreyabundance(Kaplan,1987; Madsen and Shine, 1999; Reading, 2004; Santos et al., 2005). Whateverthecauses(genes, environmentorgene–environment interaction), if resource variation isresponsible for variation in fecundity,thenweshouldexpect,atthepopulationlevel,patterns ofco-variationlargelyinconsistentwiththosepredictedby life-history tradeoffs,because femalesthat have largerresourcesat theirdisposalmayinvestmorebothingrowthandinreproduction

(3)

thanfemaleswithfewresources(deJongandvanNoordwijk,1992; RoffandFairbairn,2007).

Variation in fecundity, finally,may arise fromdifferences in theallocation strategiesofreproductive resources.In fact, pro-videdthat a female hasa fixedamount of resources toinvest inreproduction,thenumberofeggsshecouldproducedepends onhow she distributes this amount within theclutch, that is, onhow shetradesoff eggnumberagainst eggsize.Theoretical modelspredictthatinaconstantenvironmentthereshouldexist an optimaltradeoff between thesize and the number of eggs (SmithandFretwell,1975)andthatmostofthevariationin repro-ductiveinvestmentshouldbeexplainedbyvariation innumber ratherthaninsize(WinklerandWallin,1987).Thesepredictions, however,do notnecessarilyholdin speciesthatliveinvariable andunpredictableenvironments(Dziminskietal.,2009).In anu-ransthatbreedintemporaryfreshwaterenvironments,atradeoff betweenthesizeandthenumberofeggshasoftenbeenobserved (Kuramoto,1978; Lips,2001;Lüddecke,2002).Inthesespecies, theamountof resourcesthat offspring receive mayhave long-termeffectson theirbody size: tadpolesfrom eggswith large yolkvolumeshatch at a larger size(Tejedo and Reques, 1992; Loman,2002),growfaster(Kaplan,1985),runalowerpredation risk(Travisetal.,1985;BanksandBeebee,1988;Semlitschand Gibbons,1988;RichardsandBull,1990)andmetamorphosizeat alargersize(Kaplan,1985;BervenandChadra, 1988)than tad-polesfromsmallereggs.Thebeneficialeffectsofalargeamountof offspringprovisioning,however,dependonthequalityofthe envi-ronment(Crump,1981),beingusuallymorerelevantinharshthan inbenignenvironmentalconditions.Forthisreason,variationinthe qualityoftheoffspringenvironmentmayfavorbothamong-female variationinsize-numbertradeoffsandwithin-femalevariationin eggsize(Dziminskietal.,2009).

Here, we investigatethecausesof intra-specificvariation in femalefecundityin a specieswithindeterminategrowth(thus, wherethereisalifelongtradeoffbetweengrowthand reproduc-tion)and withoutparentalcare(thus,wheretheonlymaternal investmentisyolkprovisioning). Weuseasa modeltheItalian treefrog,Hylaintermedia,ananuranshowingaprolongedbreeding seasonandanactivefemalematechoice(Castellanoetal.,2009). Specifically,thisstudyhastwomaingoals.First,wedescribethe patternofintra-andinter-annualvariationinclutchandeggsize andhowthesecomponentsofreproductiveinvestmentcovarywith femaleage,bodysizeandbodycondition.Second,weinvestigate theeffectofeggsizeandnumberontadpolegrowthand develop-mentalratestoprovideinsightsintothefitnessconsequencesof variationinfemalereproductiveinvestment.

2. Materialsandmethods

2.1. Studyspecies

Formerly attributed to the European species Hyla arborea, thetreefrogpopulationsoftheItalianPeninsulaandSicilywere recentlyassignedtoaseparatespecies, H.intermedia,based on genetic studies (Nascetti et al., 1995). Just like the European treefrogs(FriedlandKlump,2005),Italiantreefrogsareprolonged breeders(sensuWells,1977)withalekmatingsystem:males con-gregateinlargechoruses,femalesvisitthechorusesandactively choosetheirpartnersonthebasisofthequalityoftheir advertise-mentcalls(CastellanoandRosso,2006,2007;Rossoetal.,2006).

2.2. Studyarea

The studypopulationbredin a dismissedpaddyﬁeld inside the‘ParcodellaValledelTicinoPiemontese’(PVTP),nearCameri

(Novara, Italy, 45◦210N, 8◦660E; 178m above sea level). The breedingsitehasbeenusedbythePVTPauthorityinalong-term amphibianconservationprojectsince2001.Asystemofirrigation channelsconveyswatertothepaddyfieldfromthenearbyTicino river,thuspermittingdirectwater-levelcontrol.Inallfouryears ofstudy(2006,2007,2009and2010),thepaddyfieldwasflooded inthesecondhalfofAprilandthewaterlevelwaskept approxi-matelyconstant(withamaximumdepthofabout60cm)forthe entirebreedingseasonuntilautumn,whenthebreedingsitewas drained.

2.3. Samplingandmarkingtechniques

Breedingactivitystarted assoonasthesitewasﬂoodedand terminated at the end of May. Every night during this period (from 9:00 p.m. to 1:00 a.m.), we moved along the shoreline and caught all pairing males and females. Pairs were carried tothelaboratoryand left tospawnovernightinseparate plas-tic boxes (25cm×20cm×10cm) ﬁlled with tap water. Paired malesand femaleswereanesthetized ina 0.2%solutionof MS-222(Sandoz,Basel,Switzerland),weighed(±0.1g)withadigital scale(multifunctionpocketscaleMF-250;MTIWeight Systems, Inc.,Kingstown,RI, USA)andtheirsnout–ventlength(SVL)was measured(±0.01mm)withadigital calliper(MitutoyoCD-15C; MitutoyoInc.,Kawasaki,Japan).

In 2006 and 2007, treefrogs were individually marked by implanting a ﬂuorescent alphanumeric tag (VIAlpha Tags, size 1.0mm×2.5mm, 0.1mm thick; Northwest Marine Technology Inc.,ShawIsland,WA,USA)beneaththeskinoftheventralsideof therighthindlimbthigh(Castellanoetal.,2009).In2009and2010, females weretoe-clippedbycuttingoutthelasttwo phalanges ofthefourth toeoftheirrighthindlimb.These phalangeswere preservedina70%ethanolsolutionandsuccessivelyusedin skele-tochronological analysis.Bothmalesandfemales werereleased atthebreedingsitethemorningaftertheircaptureandafterfull recoveryfromanesthesia.

2.4. Clutchandeggsize

Assoonasthepairsterminatedspawning,wephotographedthe entireclutchwithadigitalcamera(CanonPowerShotA75;Canon Inc.,Tokyo,Japan).Fromthreeclustersofeachclutch,weselected 15eggs,whichwerephotographedwithadigitalcameraintegrated onastereomicroscope(LeicaEZ4D;LeicaMicrosystems,Wetzlar, Germany)with16×magniﬁcation.Topreventembryomortality, clutcheswerelefttodevelopundercontrolledlaboratory condi-tions.Oncehatchingwascompleted,thetadpoleswerereleasedat thebreedingsite.

We usedImageJ version 1.44(National Institutes of Health, Bethesda,MD,USA)fortheanalysisofthedigitalpictures.Onthe entire-clutchpictures,wemanuallycountedthetotalnumberof eggslaid byfemales, using thecell-counterfunction. After set-tingascaleofabout300pixels/mminthesingle-eggpictures,we usedtheanalyze-particlefunctiontomeasureFeretdiameters(i.e., thelongestdistancebetweenanytwopointsalongtheselection boundary),fromwhichweestimatedeggvolumes.

2.5. Agedetermination

Forskeletochronologicalanalysis,wefollowed theprocedure described by Cadeddu et al. (2012; see also Smirina, 1972). Preservedphalanges werecleaned, decalciﬁed in 5% nitricacid for about 30min, and soaked in tap water overnight. Phalanx cross-sections(12␮m)wereobtainedwithafreezingmicrotome (LeicaCM1850;LeicaMicrosystems,Wetzlar,Germany),stained with hematoxylin for 25min and washed in water for 10min.

(4)

In the mid-diaphyseal bone region thereabsorption process is limited(GibbonsandMcCarthy,1983),sowechosethesections withthenarrowestmedullarcavity derived fromthis areaand mountedthemonglassslidesusingAqua-Mount(Thermo Scien-tiﬁc,Waltham,MA,USA).Twoobserversindependentlycounted thenumberoflinesofarrestedgrowthinthesectionsunderalight microscopeandlatercomparedresults.Ambiguoussectionswere discounted.Theoutermarginofthesectionswascountedasan additionallineofarrestedgrowth,becausespecimenswere col-lectedjustafteremergencefromhibernation(RogersandHarvey, 1994).

2.6. Rearingexperiment

In2010,weselected10clutches,ﬁvelaidon24thAprilandﬁve on4thMay.Clutcheswerelefttodevelopundercontrolled labo-ratoryconditionsatatemperatureof19◦C(±1◦_C)_until_hatching

wascompleted.Fromeachclutch,weselected40tadpolesatstage 25(Gosner,1960),whichwerefurthersplitintotwogroups(20 tadpoleseach).Allgroupswereplacedinseparateplasticcages (50cm×30cm×25cm)ataconcentrationofabout1tadpoleper 1.5l.Cageswerepartiallyimmersedintoawater-storagecement basin(3m×2m,2mdeep).Oneachsideofthecages,inlets cov-eredwithmosquitonettingpreventedtheentranceoflargeaquatic insects,butallowedforwaterexchangeandinﬂuxofsmall plank-tonicorganisms,whichweretheonlysourceoffoodduringthe rearingexperiment.

Onceaweek,weselectedarandomsampleof10tadpolesfrom eachcage.EachtadpolewasplacedonaPetridishwithagraph paperonthebottomandwasphotographedfromabovewitha digitalcamera(CanonPowerShotA75),whichwasmountedona photographystandataconstantdistancefromthetadpoles.After photographing,tadpoleswerereturnedtotheircages,whichwere randomlymovedwithinthecementbasintopreventanyposition effect.

Atthebeginningoftheexperiment,duetouncontrolled preda-toryevents,we observeda markedreductionin thenumberof tadpolesinﬁvecages.Sincelarvaldensityhasastrongeffecton growthrate,thesehigh-mortalityreplicateswereeliminatedfrom theexperiment.

2.7. Maternaleffectsonlarvaldevelopment

Therearingexperimentallowedustomeasure,foreachclutch, theaveragedevelopmentalrateandtheaveragegrowthrateof larvae.Developmentalratewastheinverseofthetimeat meta-morphosis(dayselapsedfromhatchingtofulltailresorption).To estimateaveragegrowthrate,wemeasuredthetadpoleSVLfrom thedigitalpictureand calculated themean larvalbody sizefor eachclutchandweek.Tostatisticallycontrolfordifferencesin lar-valdensityamonglow-mortalitycages,weestimatedgrowthrate froma generallinearmodel,whichincludedclutch-meanlarval bodysizeasthedependentvariableandclutch(randomfactor), numberofsurvivedtadpoles(covariate),age(dayselapsedsince hatching−covariate),andclutch–ageinteractionasindependent factors.Thisanalysiswascarriedoutonthedatasetoftheﬁrsttwo monthsoflarvaldevelopment,whenbodysizeincreasedlinearly withtime.Thesegrowthrateestimatesshowedastrongcorrelation withclutch-meanSVLsatmetamorphosis(R2₌_0.72;_P₌_0.002).

2.8. Statisticalanalyses

Toinvestigatethefactorsaffectingfemalefecundity,weuseda generallinearmodel(GLM),inwhichclutchsize(totalnumberof eggs)wasthedependentvariableand(i)bodysize,(ii)body condi-tion,(iii)depositionday,and(iv)yearweretheindependentfactors.

Bodysize(BS)wastheﬁrstprincipalcomponentcalculatedfrom thecorrelationmatrixoflog-transformedSVLsandpost-spawning bodyweights(canonicalloadings=0.975;percentageofvariance explained=94%). Bodycondition (BC) wasthe second principal component;itwaspositivelycorrelatedwithbodyweight (canon-ical loadings=0.26)and negatively withbody length(canonical loadings=−0.26).Femaleswithhighvaluesofthiscomponentwere consideredingoodbodyconditionbecausetheywereheavierthan expectedfromtheirbodylength.Depositiondaywasexpressedin relativeterms,asthenumberofdayselapsedsincethebeginning ofthebreedingseason.Theyearofstudywasenteredasarandom categoricalfactortoaccountforinter-annualdifferencesinfemale reproductivesuccess.Thismodelwasappliedtothefour-year sam-ple.Sincewemeasuredageonlyinthe2010breedingpopulation, wecouldnotuseageasacovariateinthefour-yearsample. How-ever,totestifagehadanysize-independenteffectsonclutchsize, we usedit inamultipleregression analysiswithbody sizeand condition.

Todescribevariationineggsize,wecarriedoutanestedANOVA, withyearasarandomfactor,femaleidentityastheﬁrstnested ran-domfactor,andclusterasthesecondnestedrandomfactor.Once ithadbeenshownthateggsizedifferedamongfemales(see Sec-tion3.3),weusedaGLMtoassesswhichfactorcouldhavebeen responsibleforthesedifferences.Themodelincludedfemalemean eggsizeasthedependentvariableandﬁveindependentfactors: bodysize,bodycondition,depositionday,yearandclutchsize.The lattervariablewasincludedtotestforatradeoffbetweenthetwo componentsoffemalereproductiveinvestment.Thisanalysiswas carriedoutonthe2009and2010samples.Toanalyzetheeffectsof ageoneggsize,wecarriedoutamultipleregressionanalysisonthe 2010sample,withage,bodysize,bodyconditionandclutchsizeas independentvariables.

AllstatisticalanalyseswerecarriedoutusingPASWvs.18(IBM Corp.,Somers,NY,USA)andR2.11.1(TheRFoundationfor Statis-ticalComputing,Vienna,Austria).

3. Results

3.1. Femalereproductiveecology

Duringthefourbreedingseasons,wecapturedandmeasured 169females:23in2006,89in2007,10in2009and47in2010. DescriptivestatisticsareshowninTable1.SVLandbodyweight dif-feredsigniﬁcantlyamongyears(P<0.001),asdidBS(F3,169=13.0,

P<0.001)andBC(F3,169=4.16,P=0.001).In2010,weanalyzedthe

ageof37outof47adultfemaleswhichwereonaverage2.32±0.85 yearsold(Fig.1).Aportionoffemales(16.2%)hadalreadyreached sexualmaturityatoneyearand8.2%offemalesshowedamaximum ageof4years.

In the ﬁrst two years of the study, 10 females (5 in 2006 and 5 in 2007) laid twoclutches in thesame breedingseason, withabetween-depositionmeandelayof15±1.2daysin2006 and 16±5.8 days in 2007. Independent of the year, the ﬁrst clutcheswerelargerthanthesecondones(pairedt-test:t=3,df=7, P=0.02).ControllingforBS,wefoundthatinbothyears(F1,98=3.08,

P=0.082) double-clutch females produced a larger number of eggsthansingle-clutchfemales(F3,169=14.82,P<0.001).Toavoid

pseudo-replication,weomittedfemaleswithdoubledepositionsin thesubsequentanalyses.

3.2. Clutchsize

Clutchsizewasahighlyvariabletrait(Table1),itscoefﬁcientof variation(CV)rangingfrom31%(in2009)to84%(in2006).TheGLM thatbestexplainedvariationinclutchsizeincluded,asindependent

(5)

Table1

Descriptivestatisticsoffemalebodysize(SVLandweight)andreproductivesuccess(clutchsizeandeggsize)infourbreedingseasonsofHylaintermedia.

Yearofstudy

2006 2007 2009 2010 ANOVA

N Mean±SD N Mean±SD N Mean±SD N Mean±SD F P SVL(mm) 22 36.08±3.47 89 33.45±2.31 10 34.43±3.15 47 36.29±3.41 12.17 <0.001 Weight(g) 23 3.37±1.03 88 2.91±0.62 10 3.22±0.83 44 3.81±0.94 13 <0.001 Clutchsize 20 456±384 82 440±284 10 422±130 44 616±327 3.49 0.017 Eggsize(mm3₎ ₁₀ _2.06_±_0.51 ₄₆ _1.92_±_0.25 _1.66 _0.203

factors, BS (b=100.05, se=20.76, P<0.001) and deposition day (b=−15.13,se=2.70,P<0.001):clutchsizeincreasedwithfemale bodysize(Fig.2A)anddecreasedfromthebeginningtotheend ofthebreedingseason(Fig.2B).Incontrast,nosigniﬁcant

differ-Fig.1.Age–frequencydistributionoffemalebreedingtree-frogs(Hylaintermedia). Asshownbythesolidline,femaleswereonaverage2yearsold,whereassome females(16.2%)reachedsexualmaturityatoneyear.

Fig.2. Theeffectof(A)bodysizeand(B)timeofdepositiononfemaleclutchsize inthe2006(circles),2007(squares),2009(triangles)and2010(upside-down trian-gles)breedingseasons.Bodysizeistheﬁrstprincipalcomponentofthecorrelation matrixofSVLandbodyweight,whereasthetimeofdepositionisdeﬁnedinterms ofthenumberofdayssincethestartofthebreedingseason.

enceswereobservedeitheramongbreedingseasons(F3,136=0.582,

P=0.628)oramongfemales ofdifferentBC(b=11.36,se=20.89, P=0.587).Whenwerestrictedtheanalysistothe2010sampleto includeageinthesetofindependentvariables,thepositive associ-ationbetweenclutchsizeandBSwasconﬁrmed,butnotbetween clutchsizeandage(b=36.51,se=65.61,P=0.582).

3.3. Eggsize

Wemeasuredthesize(thesphericalvolumeinmm3₎_of₈₀₀

eggsfrom56clutches:148eggsfrom10 clutchesin 2009,and 652eggsfrom46clutchesin2010.Themeanwithin-femaleCVsof eggsizewere12.6%in2009and12.7%in2010(range:5.6–28.7%). Themeanamong-femaleCVswere25%in2009and13%in2010. ThenestedANOVAshowedsigniﬁcantdifferencesamongclutches (F54,107=15.4, P<0.001) and among clustersof thesameclutch

(F108,636=1.34,P=0.02).TheGLMthatbestexplainedegg-size

vari-ationincludedBSanddepositiondayasindependentfactors.As observedforclutchsize,largefemalestendedtolaylargereggsthan smallerfemales (b=0.133,se=0.034,P<0.001)(Fig.3A).Unlike clutchsize,however,eggsizetendedtoincreasewiththeadvancing ofthebreedingseason(b=0.025,se=0.008,P=0.003)(Fig.3B). Nei-therBC(b=0.06,se=0.05,P=0.196),noryear(F1,45=1.33,P=0.255)

norclutchsize(b=−7.59×10−6,se=1.42×10−4,P=0.958)showed anysigniﬁcanteffectoneggsize.Asimilaranalysis,restrictedtothe

Fig.3.Theeffectof(A)bodysizeand(B)timeofdepositiononfemaleeggsizein the2009(triangles)and2010(upside-downtriangles)breedingseasons.Bodysize istheﬁrstprincipalcomponentofthecorrelationmatrixofSVLandbodyweight, whereasthetimeofdepositionisdeﬁnedintermsofthenumberofdayssincethe startofthebreedingseason.

(6)

2010sample,providednoevidenceforaneffectofageoneggsize (b=0.056,se=0.05,P=0.238).

3.4. Maternaleffectsonlarvaldevelopment

TheGLMthat bestexplained variation in larvalgrowth rate includedinthesetofindependentvariablesboththetotalnumber andthemeanvolumeofeggsperclutch:thenumberofeggs nega-tivelyaffectedgrowthrate(N=10;b=−6.3×10−5,se=1.7×10−5; P=0.008),whereaseggvolumehadapositive,butstatistically non-signiﬁcant,effect(N=10;b=4.9;se=2.4;P=0.076).Developmental rate correlated positively with growth rate (N=10; R=0.896; P<0.001)andwasaffectednegativelybythenumberofeggsina clutch(N=10;b=−3.5×10−5;se=1.2×10−5;P=0.022),and posi-tivelybythemeaneggsize(N=10;b=0.44;se=0.17;P=0.032).

4. Discussion

Inanuranswithoutparentalcare,femalereproductive invest-ment is mainly explained by the number and the size of eggs produced.Inthislong-termstudy,weinvestigatedfactors respon-siblefortheobservedvariationinthesetwocomponentsoffemale reproductiveeffort.Thisstudyprovidesthreemainresults.First, itshowsthatboththenumberandthesizeofeggsarepositively correlatedwithfemales’bodysize,butnotwiththeiragenorwith theirpost-spawningbodycondition.Second,thereisnoevidence forvariationinreproductiveeffortbetweenbreedingseasons,but strongevidencefor variationwithinbreedingseasons:withthe advancingof the breedingseason, females tend to lay smaller clutcheswithlargereggs.Third,thisstudyprovidesevidencefor anoppositeeffectofclutchsizeandeggsizeontadpolegrowth anddevelopmentalrates:independentofeggsize,thelargerthe clutches theslower thetadpole growth; independentof clutch size,thelargertheeggsthequickerthegrowth.In thelightof theseresults,wesubsequentlydiscussfactorsaffectingvariationin reproductiveinvestmentandtheirroleintheevolutionofoptimal reproductivestrategiesoffemaletreefrogs.

4.1. Bodysize,ageandreproductiveinvestment

InfemaleItaliantreefrogs,aswellasinmanyother amphib-ianspecies(Wells,2007),bodysizeispositivelyassociatedwith clutchsize.Sinceinanuranswithindeterminategrowthbodysize oftenpositivelycorrelateswithage(HallidayandVerrell,1988),its effectonclutchsizemaybeeitherdirectorindirect(orboth).Itis directifitdependsontheallometricrelationshipthatallowslarger femalestodeveloplargerovariesthansmallerfemales(Castellano etal.,2004).Itisindirect(orspurious)ifolder(andconsequently larger)femalesinvestproportionallymoreineggproductionthan younger(andconsequentlysmaller)females.Sinceourresultsdo notprovideevidenceforage-dependentdifferencesinallocation strategies,theyareconsistentwiththehypothesisthatbodysize directlyaffectsclutchsizevariation.

In many anurans, morphological constraints on fertility are viewedasone of thecausesof sexualbimaturity (Monnetand Cherry,2002;Kupfer,2007).Whenthecostsoffemale reproduc-tiveinvestmentarehigh(Elmberg,1991;Cox,2006)andfemale reproductivesuccessstronglydepends onbodysize,youngand hencesmallfemalesareexpectedtopostponereproductionand torenounceimmediatebeneﬁts inorder togrowlargerand to increasefuturefecundity(Roff,2002).Consistentwiththis predic-tion,Cadedduetal.(2012)foundthatfemaleSardiniantreefrogs, Hylasarda,reachsexualmaturitynotearlierthanattwoyearsofage andoneyearlaterthanmales.Inourpopulation,however,some femalesbredatoneyearofageandnobetween-sexdifferences

in agestructure wereobserved (Cadeddu,pers.obs.). This sug-geststhatthecostsandbeneﬁtsofpostponingreproductiondiffer betweenItalianandSardiniantreefrogfemales,possiblybecause ofdifferencesinsurvivalprobability.

Largefemalesproducenotonlyagreaternumberofeggs,but alsoeggsoflargersizethansmallerfemales.Incontrast,wefound noevidence for a direct effectof age oneggsize. The positive associationbetweenbodysizeandeggsizeiscommonlyobserved among anurans(reviewinKusanoand Hayashi,2002),whereas therelationshipbetweenageandeggsizeshowsalessconsistent pattern.Insomespecies,eggsizeincreaseswithage(Gibbonsand McCarthy,1986;Berven,1988),whereasinothers,norelationships areobservedaftercontrollingforbodysizevariation(Tejedo,1992; KusanoandHayashi,2002;thepresentstudy).

Overall, these results suggest that fast-growing females can investmoreinreproductionthanslow-growingfemalesand,thus, that, atthepopulationlevel,there isnoevidenceforatradeoff betweensomatic growthand reproduction.Since theresources that afemalecan allocatetoeithergrowthor reproductionare limited, these tradeoffs necessarilyexist at theindividual level (Stearns, 1992; Roff, 2002) and they disappear at the popu-lation level because females differ in the amount of available resources(deJongandvanNoordwijk,1992;RoffandFairbairn, 2007): females with large resources can invest more both in the current reproduction and in growth, survival and future reproductiveevents(LardnerandLoman,2003;Castellanoetal., 2004).

4.2. Intra-andinter-annualvariationinreproductiveinvestment

Althoughbodysizeissignificantlycorrelatedwithclutchsize andeggsize,itexplainsarelativelysmallportionoftheirvariance (17.5%and18.6%,respectively).Sincethereisnoevidencethata largereproductiveinvestmentreducespost-spawningbody con-dition,thesize-andage-independentvariationinclutchandegg sizeis furtherevidencefor thelargeamong-femalevariation in availableresources.Independentoftheirbody size,high-quality femalesmayaffordlargerinvestmentsthanlow-qualityfemales becausetheyaremoreefficientatconvertingexternalresources intointernalmetabolicenergy(andultimatelyintofitness).

Variationinreproductiveinvestment,however,mayalsobea matterofquantityratherthanefﬁciencyandmaydependon inter-and intra-annualvariationintheamountofavailableresources. Inter-annual variation in climaticconditions, for example, may directly affectfemale feedingconditions and,indirectly, female reproductive investment.Thisis what is hypothesizedtooccur inSardinian treefrogs,wherelargesize-independentdifferences in femalefecundity areobservedamongdifferentbreeding sea-sons(Cadedduetal.,2012).Intra-annualvariationmayariseby differencesinthetimingofdeposition.Inspecieswithprolonged breedingseasons,femalescanincreasetheamountofresources investedineggsbydelayingspawning.Inthisway,femalescan have more time to accumulate resources and to use them to completethedevelopmentofalargercomplementoffull-grown vitellogenicoocytes.

As concerns the pattern of inter-annual variation, we did observeamong-seasonvariationinfemalebodysize,butsincewe analyzedtheageoffemalesonlyinthe2010breedingpopulation, we do notknow whetherthis variation wasdue todifferences ingrowthrateorinpopulationagestructure(FriedlandKlump, 1997;Driscoll,1999).UnlikeinSardiniantreefrogs(Cadedduetal., 2012),however,wefoundnodifferencesin thelinear relation-shipsbetweenbodysizeandclutchsizeand,thus,noevidencefor inter-annual,size-independentdifferencesinfemalereproductive investment.Suchalackofdifferencesmaybeduetoconstant cli-maticandecologicalconditionsduringthestudyperiodortothe

(7)

maskingeffectofthelargeintra-annualvariationinreproductive investmentthatwedidobserve.

As concerns thepattern of intra-annualvariation, ourstudy showsaneffectofthetimingofdepositiononfemalereproductive investment,butthetemporalpatternobservedwasnotconsistent withthehypothesisthatfemalesdelayreproductiontoincrease reproductiveinvestment.Accordingtothishypothesis,the repro-ductiveinvestment is expectedto increase withthe advancing of the breeding season, whereas we observed that clutch size decreasedandeggsizeincreased.Adecreaseinclutch sizewas alsoobservedinBufocalamita(Tejedo,1992)but,differentlyfrom ourstudy,itwaslargelythesideeffectofsmallerfemales spawn-inglaterintheseason.Ourresultsseemtosupportthealternative hypothesisthatvariationinthetimingofdeposition,onceagain, maydependonfemaleenergetic budget.Femalesingood feed-ingconditionbeforehibernationmighthavebeenabletorecruit alargernumberofoocytesandcompletevitellogenesissoonafter thewinterrecovery,whereasfemalesinpoorerconditionrecruited asmallercomplementofoocytes,whichunderwentamuchlonger vitellogenesis.

4.3. Reproductiveinvestment,timingofdepositionandlarval development

Several lines of evidence show that female anurans bene-fitfrombreedingearly(MatsushimaandKawata,2005;Loman, 2009).Loman(2009)presentsanumberofplausibleexplanations: early-hatchingtadpolesmaysufferlessintenseintra-and inter-specificcompetition(Loman,2001;MatsushimaandKawata,2005) and less severepredation risksthan late-hatchingtadpoles; by metamorphosingearlier,early-hatchingtadpolemayreach win-terhibernationatalargersize(AltweggandReyer,2003;Loman, 2009)andreachsexualmaturityearlier(notethat,inour popula-tion,somefemalesbredatoneyearofage);finally,early-breeding femalesmaybeabletolayasecondclutchinthesamebreeding sea-son(SilverinandAndren,1992;DentonandBeebee,1996),thus increasingthenumberofoffspringproduced,asobservedinour study.

Whetherornotthedecreaseinclutchsizewiththeadvancing ofthebreedingseasonistheeffectofenergeticconstraints,the increaseineggsizemayarisefromastrategicchoice.Infact,in someanurans,eggsizeisfoundtohavelong-termeffectsonlarval growthanddevelopmentalrates(Kaplan,1987;BervenandChadra, 1988;Loman,2002;Cherdantsevaetal.,2007)andthereissome evidencethatfemalesadjustoffspringprovisioningtomaximize theirﬁtness.Forexample,inthequackingfrog,Criniageorgiana,

Dziminskietal.(2009) foundthat eggsizeaffects offspring ﬁt-nessindependentofmaternalphenotype,butdependentonthe qualityofthedevelopingenvironment.Inthisspecies,femalesare knowntoproduceclutcheswithhighlyvariableeggsizes,probably asabet-hedgingstrategyevolvedunderunpredictable environ-ments(Dziminskietal.,2009).Consistentwiththeseresults,we observedsigniﬁcantwithin-clutch(among-cluster)differencesin egg size and our rearing experiment conducted on a random sampleof ten clutches shows thatclutch and egg size interact andactadditivelyontadpolegrowthanddevelopment:tadpoles from small clutches of large eggs tend to grow faster and to reachmetamorphosisearlier thanthose fromlargerclutches of smallereggs.Theseresultssuggestthatfemalesadjusttheiregg provisioning plastically: at the beginning of the breeding sea-son,females investmore inthequantitythan inthequality of their offspring, because tadpole density and the risk of pond desiccationare low;asthebreedingseasonadvances,maternal investmentpersingleeggincreasesandthetotalnumberofeggs decreases.

5. Conclusion

Inconclusion,wefoundthatfemalesvaryagreatdealintheir reproductiveinvestmentand herewehighlighttwo main com-ponentsofthislargevariation.Theﬁrstcomponentisbody-size dependentandweprovideevidencethat,ratherthanthe expres-sionofgrowth–reproductionlife-historytradeoffs,itarisesfrom among-femaledifferencesintheefﬁciencyofconvertingexternal resourcesintointernalmetabolicenergy.Thesecondcomponentis body-sizeindependentandarisesfromtheplasticallocation strat-egythatallowsfemalestoadjusttheirmaternalinvestmenttothe timeofdeposition,inordertoincreasethechancesofoffspring survival.

Acknowledgments

Animalsamplingwascarriedoutundertheauthorizationofthe “ParcoDellaValledelTicinoPiemontese”thatalsoprovided logis-ticsupport.V.Marconi,V.Zanollo,G.Berto,G.BonadonnaandE. Cannarsahelpedinﬁeldwork,andE.Fiorettihelpedinlaboratory analyses.ManythankstoP.PerettoandC.Giacomaforlogistical supportduringskeletochronologicalanalysesandtoU.Sinschand A.Kupferfortheirconstructivecriticismonanearlyversionofthe manuscript.

References

Altwegg,R.,Reyer,H.U.,2003.Patternsofnaturalselectiononsizeatmetamorphosis inwaterfrogs.Evolution57,872–882.

Banks,B.,Beebee,T.J.C.,1988.ReproductivesuccessofnatterjacktoadsBufocalamita intwocontrastinghabitats.J.Anim.Ecol.57,475–492.

Berven,K.A.,1988.Factorsaffectingvariationinreproductivetraitswithina popu-lationofwoodfrogs(Ranasylvatica).Copeia1988,605–615.

Berven,K.A.,Chadra,B.G.,1988.Therelationshipamongeggsize,densityandfood levelonlarvaldevelopmentinthewoodfrog(Ranasylvatica).Oecologia75, 67–72.

Cadeddu,G.,Giacoma,C.,Castellano,S.,2012.Sexualsizedimorphisminthe Tyrrhe-niantreefrog:alife-historyperspective.J.Zool.(Lond.)286,285–292. Castellano,S.,Rosso,A.,2006.Variationincalltemporalpropertiesandfemale

pre-ferencesinHylaintermedia.Behaviour143,405–424.

Castellano,S.,Rosso,A.,2007.Femalepreferencesformultipleattributesinthe acousticsignalsoftheItaliantreefrog,Hylaintermedia.Behav.Ecol.Sociobiol. 61,1293–1302.

Castellano,S.,Cucco,M.,Giacoma,C.,2004.Reproductiveinvestmentoffemalegreen toads(Bufoviridis).Copeia2004,659–664.

Castellano,S.,Zanollo,V.,Marconi,V.,Berto,G.,2009.Themechanismsofsexual selectioninalek-breedinganuran,Hylaintermedia.Anim.Behav.77,213–224. Cherdantseva,E.M.,Cherdantsev,V.G.,Lyapkov,S.M.,2007.Theinﬂuenceofeggsize ontheintensityanddurationofRanaarvalismetamorphdevelopmentinan experimentperformedinaspawningwaterbody.Zool.Zh.86,329–339. Cox,R.M.,2006.Atestofthereproductivecosthypothesisforsexualsizedimorphism

inYarrow’sspinylizardSceloporusjarrovii.J.Anim.Ecol.75,1361–1369. Crump,M.L.,1981.Variationinpropagulesizeasafunctionofenvironmental

uncer-taintyfortreefrogs.Am.Nat.117,724–737.

deJong,G.,vanNoordwijk,A.J.,1992.Acquisitionandallocationofresources:genetic (co)variates,selection,andlifehistories.Am.Nat.139,749–770.

Denton,J.S.,Beebee,T.J.C.,1996.Double-clutchingbynatterjacktoadsBufocalamita atasiteinSouthernEngland.Amphibia–Reptilia17,159–167.

Driscoll,D.A.,1999.Skeletochronologicalassessmentofagestructureand popula-tionstabilityfortwothreatenedfrogspecies.Aust.J.Ecol.24,182–189. Dziminski,M.A.,Vercoe,P.E.,Roberts,J.D.,2009.Variableoffspringprovisioningand

ﬁtness:adirecttestintheﬁeld.Funct.Ecol.23,164–171.

Elmberg,J.,1991.OvariancyclicityandfecundityinborealcommonfrogsRana temporariaL.alongaclimaticgradient.Funct.Ecol.5,340–350.

Friedl,T.W.P.,Klump,G.M.,1997.Someaspectsofpopulationbiologyinthe Euro-peantreefrog,Hylaarborea.Herpetologica53,321–330.

Friedl,T.W.P.,Klump,G.M.,2005.Sexualselectioninthelek-breedingEuropean treefrog:bodysize,chorusattendance,randommatingandgoodgenes.Anim. Behav.70,1141–1154.

Gibbons, M.M., McCarthy, T.K., 1983. Age determination of frogs and toads (Amphibia,Anura)fromnorth-westernEurope.Zool.Scr.12,145–151. Gibbons,M.M.,McCarthy,T.K.,1986.ThereproductiveoutputoffrogsRana

tem-poraria(L.)withparticularreferencetobodysizeandage.J.Zool.(Lond.)209, 579–593.

Gosner,K.L.,1960.Asimpliﬁedtableforstaginganuranembryosandlarvaewith notesonidentiﬁcation.Herpetologica16,183–190.

Halliday,T.R.,Verrell,P.A.,1988.Bodysizeandageinamphibiansandreptiles.J. Herpetol.22,253–265.

(8)

Jörgensen,C.B.,1992.Growthandreproduction.In:Feder,M.E.,Burggren,W.W. (Eds.),EnvironmentalPhysiologyoftheAmphibians.TheUniversityofChicago Press,Chicago,pp.439–466.

Kaplan,R.H.,1985.MaternalinﬂuencesonoffspringdevelopmentintheCalifornia newt,Tarichatorosa.Copeia1985,1028–1035.

Kaplan,R.H.,1987.Developmentalplasticityandmaternaleffectsofreproductive characteristicsinthefrog,Bombinaorientalis.Oecologia71,273–279. Kupfer,A.,2007.Sexualsizedimorphisminamphibians:anoverview.In:Fairbairn,

D.J.,Blanckenhorn,W.U.,Székely,T.(Eds.),Sex,SizeandGenderRoles: Evo-lutionaryStudiesofSexualSizeDimorphism.OxfordUniversityPress,Oxford, pp.50–59.

Kuramoto,M.,1978.Correlationsofquantitativeparametersoffecundityin amphib-ians.Evolution32,287–296.

Kusano, T., Hayashi, T., 2002. Female size-speciﬁc clutch parameters of two closely related stream breeding frogs, Rana sakuraii and R. tagoi tagoi: femalesize-independentandsize-dependent eggsizes.Curr. Herpetol.21, 75–86.

Lardner,B.,Loman,J.,2003.Growthorreproduction?Resourceallocationbyfemale frogsRanatemporaria.Oecologia137,541–546.

Lips,K.R.,2001.Reproductivetrade-offsandbet-hedginginHylacalypsa,a neotrop-icaltreefrog.Oecologia128,509–518.

Loman,J.,2001.IntraspeciﬁccompetitionintadpolesofRanaarvalis:doesitmatter innature?Aﬁeldexperiment.Popul.Ecol.43,253–263.

Loman,J.,2002.MicroevolutionandmaternaleffectsontadpoleRanatemporaria growthanddevelopmentrate.J.Zool.(Lond.)257,93–99.

Loman,J.,2009.Primaryandsecondaryphenology.Doesitpayafrogtospawnearly? J.Zool.(Lond.)279,64–70.

Lüddecke,H.,2002.Variationandtrade-offinreproductiveoutputoftheAndean frogHylalabialis.Oecologia130,403–410.

Madsen,T.,Shine,R.,1999.Theadjustmentofreproductivethresholdtoprey abun-danceinacapitalbreeder.J.Anim.Ecol.68,571–580.

Matsushima,N.,Kawata,M.,2005.Thechoiceofovipositionsiteandtheeffectsof densityandovipositiontimingonsurvivorshipinRanajaponica.Ecol.Res.20, 81–86.

Monnet,J.M.,Cherry,M.I.,2002.Sexualsizedimorphisminanurans.Proc.R.Soc. Lond.B269,2301–2307.

Nascetti,G.,Lanza,B.,Bullini,L.,1995.Geneticdatasupportthespeciﬁcstatus oftheItaliantreefrog(Amphibia:Anura:Hylidae).Amphibia–Reptilia168, 215–227.

Reading,C.J.,2004.Theinﬂuenceofbodyconditionandpreyavailabilityonfemale breedingsuccessinthesmoothsnake(CoronellaaustriacaLaurenti).J.Zool. (Lond.)264,61–67.

Richards,S.J.,Bull,C.M.,1990.Size-limitedpredationontadpolesofthreeAustralian frogs.Copeia1990,1041–1046.

Roff,D.A.,2002.LifeHistoryEvolution.SinauerAssociates,Sunderland.

Roff,D.A.,Fairbairn,D.J.,2007.Theevolutionoftrade-offs:wherearewe?J.Evol. Biol.20,433–447.

Rogers,K.L.,Harvey,L.,1994.Askeletochronologicalassessmentoffossilandrecent Bufocognatusfromsouth-centralColorado.J.Herpetol.28,133–140. Rosso,A.,Castellano,S.,Giacoma,C.,2006.Preferencesforcallspectralpropertiesin

Hylaintermedia.Ethology112,599–607.

Santos,X.,Llorente,G.A.,Feriche,M.,Pleguezuelos,J.M.,Casals,F.,deSostoa,A., 2005.Foodavailabilityinducesgeographicvariationinreproductivetimingof anaquaticoviparoussnake(Natrixmaura).Amphibia–Reptilia26,183–191. Semlitsch,R.D.,Gibbons,J.W.,1988.Fishpredationinsize-structuredpopulations

oftreefrogtadpoles.Oecologia75,321–326.

Shine,R.,1988.Theevolutionoflargebodysizeinfemales:acritiqueofDarwin’s ‘fecundityadvantage’model.Am.Nat.131,124–131.

Silverin,B.,Andren,C.,1992.Theovariancycleinthenatterjacktoad,Bufocalamita, anditsrelationtobreedingbehaviour.Amphibia–Reptilia13,177–192. Smirina, E.M.,1972.Annual layersinbones ofRana temporaria.Zool. Zh.51,

1529–1534.

Smith,C.C.,Fretwell,S.D.,1975.Theoptimalbalancebetweensizeandnumberof offspring.Am.Nat.108,499–506.

Stearns,S.C.,1992.TheEvolutionofLifeHistories.OxfordUniversityPress,Oxford. Tejedo,M.,1992.Effectsofbodysizeandtimingofreproductiononreproductive successinfemalenatterjacktoads(Bufocalamita).J.Zool.(Lond.)228,545–555. Tejedo,M.,Reques,R.,1992.Effectsofeggsizeanddensityonmetamorphictraits

intadpolesofthenatterjacktoad(Bufocalamita).J.Herpetol.26,146–152. Travis, J., Keen, W.H.,Juilianna, J.,1985. The role ofrelative body sizein a

predator–preyrelationshipbetweendragonﬂynaiadsandlarvalanurans.Oikos 45,59–65.

Wells,K.D.,1977.Thesocialbehaviourofanuranamphibians.Anim.Behav.25, 666–693.

Wells,K.D.,2007.TheEcologyandBehaviorofAmphibians.TheUniversityofChicago Press,Chicago.

Winkler,W.,Wallin,K.,1987.Offspringsizeandnumber:alifehistorymodellinking effortperoffspringandtotaleffort.Am.Nat.129,708–720.