Irreversibilityofabadstart:Earlyexposure Toosmoticstress Limitsgrowth Andadaptive Developmental

Irreversibilityofabadstart:earlyexposure toosmoticstress limitsgrowth andadaptive developmental plasticity

Chi-Shiun Wu• IvanGomez-Mestre •

Yeong-ChoyKam

AbstractHarshenvironments experiencedearlyin development haveimmediateeffectsandpotentiallylong- lastingconsequencesthroughoutontogeny.Weexamined howsalinityfluctuations affectedsurvival,growthand developmentofFejervaryalimnocharistadpoles.Specifi- cally,wetestedwhetherinitialsalinityeffectsongrowth andratesofdevelopmentwerereversibleandwhetherthey affectedthetadpoles’ abilitytoadaptivelyaccelerate developmentinresponseto deterioratingconditionslaterin development.Tadpoleswereinitiallyassignedtoeitherlow orhighsalinity,andthensomewereswitched between salinity levelsuponreachingeitherGosner stage30(early switch)or38(lateswitch). Alltadpoles initiallyexperi- encinglowsalinitysurvivedwhereas thoseinitiallyexpe- riencinghighsalinityhadpoorsurvival,evenifswitchedto lowsalinity.Growthanddevelopmental ratesoftadpoles initiallyassignedtohighsalinitydidnotincrease after osmotic stressrelease.Initiallowsalinityconditions allowedtadpolestoattainafastpaceofdevelopment even ifexposedtohighsalinityafterwards. Tadpolesexperi- encing high salinity only late in development

CommunicatedbyRaoulVanDamme. C.-S.Wu Y.-C.Kam()

DepartmentofLifeScience,TunghaiUniversity, Taichung40704,Taiwan

e-mail:

I.Gomez-Mestre

Estacio´nBiolo´gicadeDon˜ana,

ConsejoSuperiordeInvestigacionesCient´ıficas, Avda.Ame´ricoVespucios/n,Sevilla,Spain

I.Gomez-Mestre

ResearchGroupofBiodiversity(UO,CSIC,PA), c/Catedra´ticoRodrigoUr´ıas/n,33071Oviedo,Spain

metamorphosedfasterandatasmaller size,indicatingan adaptiveaccelerationofdevelopment toavoidosmotic stress.Nonetheless, earlyexposuretohighsalinitypre- cludedadaptiveaccelerationofdevelopment, always causingdelayedmetamorphosisrelativetothoseininitially lowsalinity.Ourresultsthusshowthatstressfulenviron- mentsexperiencedearlyindevelopment cancritically impactlifehistorytraits,havinglong-lastingor irreversible effects,andrestrictingtheirabilitytoproduce adaptive plasticresponses.

Keywords Developmentalplasticity

Growthcompensation Metamorphosis Salinity

Fejervaryalimnocharis

Introduction

Harshenvironmental conditionscanhavedifferentconse- quences forsurvivalandlifehistorytraitsoforganisms dependingontheontogenetic timepointatwhichtheyare exposedtothem.Thus,harshenvironmentsexperiencedat anearlydevelopmental stagenotonlyhaveimmediate effectsontheorganism butcanalsohavelonglasting consequences throughoutontogeny(MetcalfeandMona- ghan2001).Inorganisms withcomplexlifecycles,the consequences ofalteringgrowthanddevelopmentaltra- jectoriesearlyinontogeny canevenpersistaftermeta- morphosis, affectingtheirsize,morphologyanddegreeof maturationatlaterstages(Pechenik2006;Gomez-Mestre etal.2010).Exposuretostressfulconditions earlyin development canalsorestricttheabilitytoproduceadap- tiveplasticresponseslaterinlife(AlvarezandNicieza

2002;De Block and Stoks 2005;Kishidaet al. 2010), hence reducing their ability to cope with changing

environments. Somespecies,however,canattenuatethe negativeeffectsofpoorinitialconditions byachieving compensatory growthiftheenvironmentalconditions improve(MetcalfeandMonaghan2001;Alietal.2003).It isessentialtounderstand howorganismsbecomemoreor lesssensitiveorresponsivetoenvironmentalchanges over thecourseofontogeny ifwewanttopredicthowenvi- ronmentalfactorswillinfluencetheirsurvivalandevolu- tion(Hensley1993).

Theenvironments experiencedbyamphibianlarvaeare usuallyheterogeneous andunpredictable,astheyareoften atriskfromponddrying,foodshortages, extremetem- peratures, predatorattacks,wateracidityorsalinity (Newman1992;Alford1999;Rose2005;WuandKam

2009).Development ofamphibianlarvaeisveryplastic, varyingtheirtimetoandsizeatmetamorphosis either adaptivelyinresponsetospecificenvironmentalcuesor non-adaptivelyduetoenvironmentalfluctuations them- selves (Newman 1992; Rose 2005; Rudolf and Ro¨del

2007).Tadpoles shouldgrowasmuchaspossiblewhile conditionsarebenignintheirlarvalenvironment(i.e.risk of mortality is low) but accelerate development and achieveanearlymetamorphosis iftheenvironment becomesinhospitable(Werner1986;Benard2004;Vonesh andWarkentin 2006).However, theeffectofanygiven environmental factorontadpoledevelopmentisoften highlydependent onthestageandcondition oftheanimal. Thus,thesameenvironmentalfactormaydecreasesurvi- vorship, growthanddevelopmental ratesifexperienced duringpremetamorphic stages,butacceleratedevelopment andinduceearlymetamorphosis ifexperiencedduring prometamorphicones(Denver1997;Denveretal.2002). Acceleratinglarvaldevelopmentmaycomeatthecostof smallerandlessmaturejuveniles(Newman 1992),butso doescontinueddevelopment understressfulconditions,in addition to reducing larval survival. This is relevant becausesmallsizeatmetamorphosisreduces post-meta- morphic survival (Smith1987;Berven1990;Gomez- MestreandTejedo2003;Chelgrenet al. 2006),delaystime tosexualmaturity(Semlitschetal.1988;Berven1990), andreducesthesizeatmaturity(Semlitsch etal.1988), overallreducingpopulationrecruitment(Smith1987).

Salinityisanimportantstressorformostamphibians duetotheirhighly permeableskinandinabilitytocon- centrateurine(Balinsky 1981;DuellmanandTrueb1994). Alltadpoles aresensitivetosalinity,andmoderatesalinity (2–9partsperthousand;ppt)decreases tadpolesurvival, growth,anddevelopment (ChristyandDickman2002; Gomez-Mestreetal.2004;Chinathambyetal.2006;Wu andKam 2009). Inmostsystems,pondsalinity islowestat thebeginning oftheamphibianreproductiveperiod,as pondsfillup,andgraduallyincreasesovertimeasponds evaporateandaremaximaljustpriortoponddesiccation.

Surprisingly,nostudiestodatehaveaddressedhowsalinity fluctuations canaffectamphibianlarvaldevelopment. However, afewspeciesbreedincoastalpoolswithstrong marineinfluence wheresalinitycanfluctuate considerably (WuandKam2009;SilleroandRibeiro2010).Tadpolesof theIndianricefrog(Fejervaryalimnocharis)livinginthe brackishwaterofcoastalrockpuddlesandpoolsformedby upliftedcoral onthetropicalGreenIslandofTaiwan pro- videanexcellentmodel systemtostudytheeffectsof salinity-induced stressoverlarvalontogenyongrowth, development andadaptivedevelopmental plasticity(Wu andKam2009).Weexaminedhowchanges insalinity affectedsurvival, growthanddevelopmentofF.limno- charistadpoles.Specifically, wetestedwhethersalinity effectsearlyindevelopmentwerereversibleandwhether theyaffectedtadpoles’ abilitytoadaptivelyaccelerate development.

Materialsandmethods

Studyanimals

Fejervarya limnocharis is a medium-sized frog (30–60 mm)distributedthroughouteast,southeast,andsouthAsia, including many small, isolated islands (Sumida et al.

2007).InTaiwan,F.limnocharisiswidelydistributedup toelevationsof1,000monthemainislandandonoff- shoreislands(Alexanderetal.1979;Yang1998).Fejer- varyalimnocharisusuallybreedsfromFebruary toSep- tember, with breedingcorrelated with rainfall and irrigationandrestrictedbylowtemperatures(Alexander etal.1979). Thisspecies usually breeds intemporary freshwater pools suchasricepoolsandroadside puddles, butsomepopulations breedinthebrackishwaterofrock poolsincoastalareas(WuandKam2009).Meanclutch sizevariesgreatly,witharangeofabout450–1,800,and thelarvalperiodislessthan2months(Alexanderetal.

1979). Studysite

Green Island(121.28°E,22.35°N)isasmalltropicalisland offthesoutheasterncoastofTaiwan,withatotalareaof

15km2.Theclimateonthisislandischaracterizedbyhigh

airtemperatureandabundant annualrainfall,withnodis- tinctdryseason(Ni2003).Typhoonsarefrequentbetween June and September,and usually bring abundantrains. Themonthly averageairtemperaturewas23.5°C andthe annualrainfallwas2,575mm(1998–2006;datafromthe CentralWeatherBureau,Taiwan). Duringthebreeding seasonofF.limnocharis(February–September),themost abundant rains occurred mainly between July and

September (averagemonthly rainfallabout290–320 mm). Fejervaryalimnocharisusuallybreeds inrockpools just above thehightideline.Poolswithtadpolesvaried greatly in size and shape (range of maximum diameter

11–534cm),andwererelatively shallowindepth(rangeof maximumdepth5–22cm) (WuandKam2009).These rockpoolsareuncommon breedingsitesforamphibians becausetheir smallsize,seawaterspillsandstrongrainfalls during the monsoon season cause frequent and large salinityfluctuations.Salinityinthesepoolsvariesbetween

0and23ppt,althoughnotadpoleshavebeenfoundin poolswithsalinityover12 ppt(WuandKam2009).Our fieldsurveysshowedtadpolescouldbeexposedtodifferent salinities(0–12ppt)duringontogeny,especiallyatanearly developmental stage (Gosner stage 26; Gosner 1960) (Fig.1).

Experimentaldesign

Wecollectedtadpolesof F.limnocharisatGosnerstage26 (Gosner1960)fromsixpools(salinitywas0,1,1,2,3and

3 ppt)onGreenIslandinJune2006.Tadpoles fromeach poolweremixedtogetherandthenrandomlyassigned to oneofsixtreatments.Tadpoleswereinitiallyassignedto eitherloworhigh salinities(3and9ppt,respectively).We established experimentalsalinityswitchesatspecific developmental stages,anearlysalinityswitchatGosner stage30,andalateswitch atGosner stage38.Thereason wechosestage30andstage38forswitchpointsis because theformeriswithinthepremetamorphic period,when tadpolesgrowwithfewmorphological changesandlow plasmathyroidhormone(TH),andthelatteriswithinthe

Withouttadpoles(n=14)

Withtadpoles(n=37)

6 stage26only

5stage27-30

4stage31-35

stage36-40

02469 11 13 15 17 19 21 23

Salinity (ppt)

Fig.1 SalinityofpoolswithandwithoutIndianricefrog(Fejer- varyalimnocharis)tadpoles. Tadpolesappearedonlyinpoolswith salinities under12ppt,andtheycouldexperiencedifferent salinities throughoutdevelopment. Wecategorizedfourdevelopmentalperiod oftadpolesaccordingtothe oldesttadpoleswesurveyedin eachpool. Wesurveyedatotalof51pools

prometamorphic period,whentadpolesgrowwithobvious morphologicalchangesandincreasingplasmaTH(Denver

2009).Hence,weestablishedatotalofsixtreatmentsas

follows: constant low salinity (LLL), low salinity with earlyswitchtohighsalinity(LHH), lowsalinitywithlate switchtohighsalinity(LLH),constanthighsalinity(HHH), highsalinitywithearlyswitch tolowsalinity(HLL), and highsalinity withlateswitch tolowsalinity (HHL). Because highersalinityslowsdowntherateoftadpole development(Gomez-Mestreet al. 2004;WuandKam

2009),tadpolesreachedthespecificdevelopmentalstages atdifferenttimeintervals.Becausethetimerequired for animalstodevelop fromonestagetothenextisnotuni- form,but dependsonthespecificstagesand environmental conditions,thetimingofsalinityswitches variedacross treatmentsandwere determinedbyaveragedevelopmental stage. Salinity switches occurred on days 6.8±0.2 (mean±SE, n=20), 20.3±0.2 (n=20), 12.2±0.6 (n=20),and27.3±2.7(n=9),forLHH,LLH,HLL andHHL,respectively.Likewise, theaveragetadpole weight at switch points was 114±2.5, 343.1±

6.8, 94.6±4.3, and 329.8±19.1mg, for LHH, LLH, HLLandHHL,respectively.Each treatment wasreplicated

20times,eachreplicateconsistingofindividualtadpoles raised in plastic containers (length9width9height=

10.597.594.5cm) holding100mlofwater.Containers

werekeptcovered withatransparentperforated lidto reduceevaporation. Weobtainedthedifferentsaline solutionsbydissolvingCoralifescientific-grademarinesalt (Energy Savers,Carson, CA,USA)indistilledwater,and thelevelofsalinity waschecked at26°Cwithasalinity refractometer. We fed the tadpoles boiled vegetables adlibitumandchangedthewatereverythirdday.Tadpoles werekeptinincubators at26°Cundera12:12hlight:dark cycle.Wemonitored tadpolesurvivalandwatersalinity daily,adjustingsalinitywhennecessary. Wemeasuredthe growthanddevelopment oftadpolesonceaweekuntil metamorphosisbystagingandweighingeachtadpoleto thenearest0.0001g.Beforeweighing,wegentlyblotted eachtadpoletoremoveexcess water.Tadpoleswerecon- sideredtohavereachedmetamorphosisatGosnerstage42, uponforelimbemergence(Gosner 1960).Atthistime,we recordedthedateandmeasuredtadpolebodysize.

Dataanalyses

WeusedtheKaplan–Meiersurvivalanalysistocompare thecurvesoftadpolessurviving tometamorphosisindif- ferenttreatments(KaplanandMeier1958). Allindividuals wereweighed andstagedweekly todeterminegrowth trajectoriesineachtreatment.Wetestedfordifferences in growthanddevelopment amongtreatmentsbyusinga repeated measures ANOVA (Potvin et al. 1990). This

analysis couldonlybeconductedoverfourtimepoints because after 3weeks tadpolesin sometreatmentshad metamorphosed.Wetestedtheassumption ofTypeH covariance(HuynhandFeldt1970)inourrepeatedmea- suresANOVAs,applyingsphericitytests(Anderson1958). Asthedatadidnotsatisfysuchan assumption,weadjusted the significancelevels for the F tests according to an epsiloncorrectionfactoroftheirdegreesoffreedom(Huynh andFeldt1976).

Larvalperiodandsizeatmetamorphosis wereheteros- cedastic,andthereforeweranked thesevariablespriorto conductingANOVAtotestfordifferences amongtreat- ments,followed byFisher LSDposthoctests.Dataanal- ysiswasconductedwithSPSS 11.0(SPSS,Chicago,IL, USA)andSAS9.2(SASInstitute,Cary,NC,USA).

Results

Survivalandmetamorphosis

Alltadpolesinitiallyassigned tolowsalinity(LLL,LLH, andLHHtreatments)survivedtometamorphosis(Fig. 2). Survivorshipwasdecreasedintadpolesinitiallyexposedto

anymorphologicalabnormalitiesduringontogenyintad- polesindifferentsalinities.

Growthanddevelopment

Growthanddevelopment oftadpoles differed significantly duringthefirst 3weeksamongtreatments(repeatedmea- sures ANOVA, treatment9time: F15,318=30.63, P\

0.0001 and F15,315=18.14, P\0.0001, respectively;

Fig.3).Earlyexposure tohighsalinityhadastrongeffect ontadpolegrowthanddevelopment.Tadpolesexposed to highsalinityearlyindevelopment grewanddeveloped moreslowlythanthosewhoseearlydevelopment occurred inlowsalinity,regardlessofwhethertheywereexposedto lowsalinitylaterinlifeornot(i.e.Lxxvs.Hxx,treat- ment9time: F3,330=72.75, P\0.0001 and F3,327=

68.30,P\0.0001,respectively;Fig.3).Onaverage,tad-

polesininitiallyhighsalinity (Hxx)gained 41%asmuch mass during the first3weeks as those in initially low salinity (Lxx). Afterthose3weeks,tadpoles intreatments withearlyexposure tohighsalinitywereinstages34–35, beforecompletetoeseparation, whereastadpolesintreat- mentsexposedtolowsalinityearlyin developmentwerein

salinity, survivalbeing95,45and30%inHLL,HHLand HHH,respectively.Thesurvivaloftadpolessignificantly differedamong theinitiallyhighsalinity treatments(log ranktest,v2 =11.78,P=0.0028,df=2).Tadpolesur- vivalinHLLwassignificantlyhigherthanthatateither constant high salinity (HHH) (v2 =8.77, P=0.0030, df=1) or HHL (v2 =11.18, P=0.0010, df=1), whereasHHHandHHLdidnotdifferfromoneanother (v2 =1.71,P=0.6790,df=1;Fig.2).Wedidnotfind

400

300

200

100

LLLHHH LHHHLL

LLHHHL

100

LLL39

LHH

LLH

HHH33

HLL

HHL

010203040506070

Time(days)

Fig.2 Survival(%)ofF.limnocharistadpolesrearedindifferent salinityregimes.Thelastdatumpointforeachtreatmentreflectsthe timeatwhichthelasttadpolemetamorphosed

07142128

Time (days)

Fig.3 Growth(a)anddevelopment(b)ofF.limnocharistadpolesin different salinity treatments.Thelastdatumpointforeachline representsthetimeatwhich thefirsttadpolemetamorphosed.Sample sizesduringthefirstthreeweekswere20pertreatment,exceptfor HHHandHHL(n=18andn=14,respectively).Samplesizein the fourthweekwas12, 13,and9forHHH,HLLandHHL,respectively. Bars±SE

stages 38–39, much closer to metamorphosis (Gosner

1960).Nonetheless, switchingtohighsalinitysignificantly reducedgrowthandsloweddowndevelopment duringthe first3weeksintadpolesinitiallyinlowsalinity(among Lxx treatments, treatment9time: F6,171=28.75, P\

0.0001 and F6,171=12.10, P\0.0001, respectively).

Extended exposure to salinity further contributed to reducedgrowth(amongHxx treatments, treatment9time: F6,147=5.42,P=0.004)butnottoinfluencedevelopment (treatment9time:F6,144=1.98,P=0.1240)intadpoles initiallyinhighsalinity.

Metamorphictraits

Thetimetoandsizeatmetamorphosis oftadpolessignif- icantly differed among treatments (ANOVA, F5,88=

14.61,P\0.0001andF5,88=11.30,P\0.0001,respec-

tively).Ininitiallylowsalinity treatments(LLL,LHH, and LLH),posthoccomparisons showedthattimetometa- morphosis oftadpolesswitchedtohighsalinityatGosner stage30(LHH)wasstatisticallysimilartothatataconstant lowsalinity(LLL),whereas timetometamorphosisof tadpolesswitchedtohighsalinityatGosnerstage38(LLH) wassignificantly shorterthanthatataconstantlowsalin- ity,indicatingadevelopmentalaccelerationinresponse to lateincreaseinsalinity(Fig.4a).Theaveragedevelop- mental acceleration was 2.4days, approximately a 9% shorterlarvalperiodthantadpoles inconstantlowsalinity. Sizeatmetamorphosisofjuveniles fromLLLwason average32%greaterthanthatofLHHindividuals,and

30%greaterthanthoseinLLH(Fig.4b),whereasLHH

andLLHdidnotdifferfromoneanother(Fig.4b).

Timetometamorphosisdidnotvaryamong treatments withinitialhighsalinity(HHH,HLL,andHHL;Fig. 4a). SizeatmetamorphosisofjuvenilesfromHHH,however, wasonaverage46%lowerthanjuvenilesfromHLL,and

31%smallerthanjuvenilesfromHHL(Fig.4b).

Discussion

Asexpected,higher salinity reduced survival,growthand development oftadpoleswhenexposureoccurredearlyin development(premetamorphosis). Somespeciesseemto avoidbreedinginpoolswithhighsalinity(Gordonetal.

1961;Haramura 2008), butevenifF.limnochariscould discriminate,theiroffspringwouldstillbeathighriskof experiencingharmfulsalinitiesasaresultofhighevapo- rationratesinthesmallcoastalrockypools inwhich they breed.Salinity stressexperiencedearlyindevelopment markedly reducedgrowthanddevelopmentofF.limno- charistadpoles,seemingly incapableofcompensatory growthuponrelease from salinitystress.Consequently,

a45

40c

35c

30aa

b 500

400

bbb

300

200

LLLLHHLLHHHHHLLHHL

Treatment

Fig.4 Timeto(a)andsizeatmetamorphosis (b)forF.limnocharis tadpolesindifferentsalinityregimes. Identicallettersabovebars indicatenosignificantdifferencebetweentreatments.Samplesizes were20pertreatment,exceptfortreatmentsHHH,HLL,andHHL (n=6,n=19andn=9,respectively).Bars±SE

evenifF.limnocharistadpolessurviveinitialexposuresto osmoticstress,theyarelikelytosufferincreased vulnera- bilitytootherrisks,asreduced growthandextended developmentexposesthemtoaquaticpredatorsforalonger time,putsthematriskofponddesiccation,reducestheir odds of post-metamorphic juvenile survival (Newman

1992; Alford 1999), and impacts population dynamics

(Karrakeretal.2008).

Lackofcompensatorygrowthtoovercomeearly osmoticstress

Thelongertadpoles stayedinhighsalinity, thestronger weretheadverse effects,butwithastrongpreeminenceof theconditionsexperiencedearlyindevelopment.Evenif released fromosmoticstress,tadpoles exposed tohigh salinityearlyindevelopmentcouldnotincreasegrowthand

developmental ratesenoughtomatchLLL,andmetamor- phosed laterand oftenatasmallersize(Fig.3a,b).How- ever,tadpolesreleasedfromosmotic stressearly(HLL) managedtoreachsimilarsizeatmetamorphosis thantad- polesmaintainedinlowsalinitythroughout development, althoughittooklongertoreachmetamorphosis, indicating their ability forpartialgrowthcompensation(Aliet al.

2003;Squiresetal.2010).Tadpolesexposedtohighsalinity earlyindevelopment(Hxx),orforalongperiodoftime (LHH), mayhavebeenforcedtoallocatemoreenergy to osmoregulation(Gomez-Mestre etal.2004),andconse- quentlyitmayhavetakenlongertooffsetthisincreased energyconsumption andattaintheminimumsizefor metamorphosis. WilburandCollins(1973)hypothesized thattadpolesmustattaintheminimum(threshold) size beforemetamorphosis.We hypothesizethat tadpolesof F.limnocharisatGosner stage38haveattainedthe threshold sizeformetamorphosis. MoreyandReznick (2000)demonstratedthatthethresholdsizeexistsandthatit differsamongspecies, witharangebetweenGosnerstages

35and39,andthedifferenceisprobably relatedtoadap- tationtodifferent regimesofhabitatephemerality.During the periodof Gosnerstages36–41(prometamorphosis),the endocrinesystemoftadpolesissufficientlydevelopedto adjustthedevelopmentalrateinresponsetoenvironmental variation(reviewed byDenver2009).Incontrast,tadpoles atGosnerstage30(premetamorphosis)couldnotinitiate metamorphosisafterswitching fromlowtohighsalinity (LHH),andshowed decreased growthanddevelopment rates,consistentwithpreviousstudies(Gomez-Mestreetal.

2004;Chinathamby etal.2006;WuandKam2009).

Previousstudieshaveshowncompensatory growthin tadpolesfollowingreleasefromfooddeprivation(Alford andHarris1988;Beachyetal.1999;CapellanandNicieza

2007),ponddesiccation(Denveretal.1998)andsalinity

(Squires et al. 2010). The capacity for compensatory growth mayvaryamong species ordepend onthetypeof environmentalstressencountered(CapellanandNicieza

2007).Theresponsetowaterrefillingafterhabitatdesic- cationintadpolesofScaphiopushammondii ispartially reversible,restoring thebodysizeandretardingmeta- morphosis, butthereversibilitydependsonthedevelop- mentalstageoftadpoles(Denveretal.1998).

Severallinesofevidencesuggestthatthedevelopmentof internalgillsiscriticalinenhancedsalinitytolerance.Tad- polesexperiencinglowsalinityinearlydevelopment(before Gosnerstage38)hadmuchhighersurvivalrates(100%) thantadpolesexposedtohighsalinity during thesameper- iod,eveniftheywerelaterswitchedtolowsalinity.Gillsare the main organsresponsiblefor ion and water balance intadpoles(DietzandAlvarado 1974;Uchiyamaand Yoshizawa1992;Ultschetal.1999),andtadpolesmaylack sufficientsalt-excretingabilitytocopewithhighosmotic

stressuntilinternalgillsdevelop(Chinathambyetal.2006). UchiyamaandYoshizawa(1992)suggested thattheeury- halinetadpolesofFejervaryacancrivora(Ranacancrivora) toleratedhighersalinityoncetheyhadinternalgillsthan thosestillrelying onexternalonesbecausetheformerare moreabundantinmitochondria-rich cells(MRcells), thoughtkeyinionexcretion.Concordantly, MRcellsonly appearedinRanadalmatinatadpolesonceinternalgillshad developed, butwereabsentintheirexternalgills(Brunelli etal. 2004).Shiftstohighsalinitydidnotinfluencetadpole survivalinitiallyassignedtolowsalinity,maybebecause theirgillswerealreadywelldeveloped.Therefore,thetim- ingofinternalgilldevelopmentrelativetosalinityexposure maydeterminetadpolesurvival.

Earlyexposuretosalinityprecludedadaptive developmentalplasticity

Inoursystem,initiallowsalinityconditions seemedto allowtadpolesto buildenoughcapacitytomaintainnormal development evenifexposed tohighsalinitylaterin development, althoughthismayhavebeenenergetically verycostlyasitcameattheexpense ofreducedsizeat metamorphosis (LLL vs. LHH, Fig.4). Priming fast development attheexpenseofreducedsizeattransforma- tionisconcordant withanadaptiveaccelerationofdevel- opmenttoavoidstressfulconditions(Newman 1992; Denver1997).Suchdevelopmental accelerationwaseven moremarked intheLLHtreatment,inwhichtadpoles rearedinlowsalinitythroughoutmuchoftheirdevelopment wereexposed toosmotic stressatalatestage.Tadpoles in LLHshowed theshortest timetometamorphosis,even shorterthanthosecontinuously inlowsalinity,evidencing developmentalaccelerationagainatthecostofreducedsize atmetamorphosis.Tadpolesgrowinginlowsalinityseemed tohaveattainedthethreshold sizeformetamorphosisat Gosner38,andprecipitatedmetamorphosis eitherasa neuroendocrine directresponsetoincreasedsalinity,or indirectlywhengrowthwasadversely impactedbyhigh salinity(Werner 1986). Suchshiftsinmetamorphictiming mayprovideameansfortadpolestoadapttotheunpre- dictablenatureofrockpools,whichfrequently experience largeandfastfluctuationsinsalinity,duetowavespray, evaporation, orrainfall(WuandKam2009). Nonetheless, earlyexposuretohighsalinityprecluded adaptiveacceler- ationofdevelopment, alwayscausingdelayedmetamor- phosisrelativetothoseinlowsalinityearlyindevelopment. Thisresult highlightsthefactthatstress experiencedearly inontogeny maycondition growthanddevelopmental tra- jectoriestothepoint oflimitingthepotentialforadaptive plasticresponsesinlaterstages.

In conclusion, our results indicate that the salinity experiencedintheearlystageswasimportantforsurvival,

growthanddevelopmentofF.limnocharistadpoles.Fur- thermore,theeffects ofosmoticstressongrowth were partiallyreversible butthoseondevelopmentwereirre- versibleevenafterreleasefromstress.However, F.lim- nocharistadpolescanacceleratedevelopmentinresponse toincreasedwatersalinityifexposed duringprometamor- phosis.Iflowsalinityallowsadequate growthearlyin development, tadpoles canquickly attainaminimumsize formetamorphosis,whichwould enablethemtoaccelerate developmentifconditions deteriorated(e.g.,salinity increased).Conversely, earlyexposuretoosmoticstress reducesgrowth anddevelopmentalratessothattadpoles cannolongerbuildupenergeticreserves required to acceleratemetamorphosis.Wehypothesizethatosmoreg- ulationunderosmotic stressmaytakeanimportant meta- bolic toll that impairs the accumulation of body fat requiredtoundergoacceleratedmetamorphosis.Thisstudy highlightstheimportance oftheenvironmental conditions experiencedearlyindevelopment, andhowtheycancrit- icallyimpactlifehistorytraits.

Acknowledgments ThisstudywassupportedbyaNationalScience CouncilGrant(NSC95-2311-B-029-006-MY3)toYCK.Wethank theStationoftheEastCoastNationalScenicAreaAdministrationof GreenIslandforlogisticsupport, andC.S.Wu,T.L.LinandR.F. Chaoforadministrativeandfieldassistance;andJamesR.Voneshfor commentsonanearlydraftofthispaper.

References

Alexander PS, Alcala AC, Wu DY (1979) Annual reproductive patterninthericefrogRanal.limnocharisinTaiwan.JAsian Ecol1:68–78

AlfordRA(1999)Ecology:resourceuse,competion,andpredation.

In: McDiarmid RW, Altig R (eds) Tadpole, the biology of anuranlarvae.UniversityofChicagoPress,Chicago,pp240–

278

Alford RA,HarrisRN(1988) Effectsoflarvalgrowth historyon anuranmetamorphosis.AmNat131:91–106

AliM,NiciezaA,WoottonRJ(2003)Compensatorygrowthinfishes:

aresponsetogrowthdepression.FishFish4:147–190

AlvarezD,NiciezaA(2002) Effectsoftemperatureandfoodquality on anuran larval growth and metamorphosis. Funct Ecol

16:640–648

AndersonTW(1958)Anintroduction tomultivariatestatistical analysis.Wiley,NewYork

BalinskyJB(1981)Adaptationofnitrogenmetabolismtohyperos-

moticenvironmentinamphibia.JExpZool215:335–350

BeachyCK,SurgesTH,ReyesM(1999)Effectsofdevelopmental andgrowthhistoryonmetamorphosis inthegraytreefrog,Hyla versicolor(Amphibia,Anura).JExpZool283:522–530

BenardMF(2004)Predator-inducedphenotypicplasticityinorgan- isms with complexlife histories.Annu Rev Ecol Evol Syst

35:651–673

BervenKA(1990)Factorsaffectingpopulationfluctuationsinlarval andadultstagesofthewoodfrog(Rana sylvatica).Ecology

71:1599–1608

BrunelliE,PerrottaE,TripepiS(2004)Ultrastructure anddevelop- mentofthegillsinRanadalmatina(Amphibia, Anura). Zoomorphology123:203–211

CapellanE,NiciezaAG(2007) Nonequivalenceofgrowth arrest inducedbypredation riskorfoodlimitation:contextdependent compensatory growth in anuran tadpoles. J Anim Ecol

76:1026–1035

ChelgrenND,Rosenberg DK,HeppellSS,GitelmanA(2006) Carryover aquaticeffectsonsurvival ofmetamorphicfrogs duringpondemigration.EcolAppl16:250–261

Chinathamby K,ReinaRD,BaileyPCE,LeesBK(2006)Effectsof salinityonthesurvival,growthanddevelopment oftadpolesof thebrowntreefrog,Litoriaewingii.AusJZool54:97–105

ChristyMT,DickmanCR(2002)Effectsofsalinityontadpolesof greenandgoldenbellfrog(Litoriaaurea). AmphibiaReptilia

23:1–11

DeBlockM,StoksR(2005)Fitnesseffectsfromeggto reproduction:

bridgingthelifehistorytransition.Ecology86:185–197

DenverRJ(1997)Proximatemechanisms ofphenotypicplasticityin amphibianmetamorphosis.AmZool37:172–184

DenverRJ(2009)Endocrinologyofcomplexlifecycles:amphibians.

In:PfaffD,ArnoldA,EtgenA,Fahrbach S,MossR,RubinR (eds)Hormones, brainandbehavior,2ndedn.Elsevier,San Diego,pp707–744

DenverRJ,MirhadiN,PhillipsM(1998)Adaptive plasticityin amphibianmetamorphosis:responseofScaphiopushammondii tadpolestohabitatdesiccation.Ecology79:1859–1872

DenverRJ,GlennemeierKA,BoorseGC(2002)Endocrinologyof complexlifecycles: amphibians.In:PfaffD,ArnoldA,EtgenA, FahrbachS,MossR,RubinR(eds)Hormones, brainand behavior.Elsevier,SanDiego,pp469–513

DietzTH,Alvarado RH(1974)NaandCltransport acrossgill chamberepitheliumofRanacatesbeianatadpoles.Am JPhysiol

226:764–770

DuellmanWE,TruebL(1994)Biologyof amphibians.McGraw-Hill, NewYork

Gomez-Mestre I,TejedoM(2003)Localadaptationofananuran amphibians to osmotically stressful environments. Evolution

57:1889–1899

Gomez-Mestre I,TejedoM,RamayoR,EstepaJ(2004)Develop- mentalalterationsandosmoregulatory physiologyoflarval anuranunderosmoticstress.PhysiolBiolZool77:267–274

Gomez-MestreI,SaccoccioVL,Iijima T, CollinsEM,RosenthalGG, Warkentin K(2010)Theshapeofthingstocome:linking developmentalplasticitytopostmetamorphicmorphology in anurans.JEvolBiol23:1364–1373

GordonMS,Schmidt-Nielsen K,KellyHM(1961)Osmoticregula- tion in the crab-eating frog (Rana cancrivora). J Exp Biol

38:659–678

GosnerKL(1960)Asimplifiedtableforstaginganuranembryosand larvaewithnotesonidentification.Herpetologica16:183–190

Haramura T(2008)Experimentallytestofspawningsiteselectionby Buergeriajaponica(Anura:Rhacophoridae) inresponseto salinitylevel.Copeia2008:64–67

HensleyFR (1993)Ontogeneticlossof phenotypicplasticityofageat metamorphosis.Ecology74:2405–2412

HuynhH,FeldtLS(1970)Conditionsunderwhichmeansquareratios

inrepeated measurementsdesignshaveexact F-distributions. JAmStatAssoc65:1582–1589

HuynhH, Feldt LS (1976)Estimation of the Boxcorrectionfor

degreesoffreedomfromsampledatainrandomizedblockand split-plotdesigns.JEducBehavStat1:69–82

KaplanEL,MeierP(1958)Nonparametricestimationfromincom- pleteobservations.JAmStatAssoc53:457–481

KarrakerNE,GibbsJP,VoneshJR(2008)Impactsofroaddeicing salt on the demographyof vernalpool-breedingamphibians. EcolAppl18:724–734

KishidaO,TrussellGC,MougiA,NishimuraK(2010)Evolutionary

ecologyofinduciblemorphologicalplasticityinpredator–prey

interaction:towardthepracticallinkswithpopulationecology. PopulEcol52:37–46

MetcalfeNB,MonaghanP(2001)Compensationfora badstart: grow now,paylater?TrendsEcolEvol16:254–260

MoreyS,ReznickD(2000)Acomparativeanalysisofplasticityin

larvaldevelopmentinthreespeciesofspadefoottoads.Ecology

81:1736–1749

NewmanRA(1992)Adaptiveplasticityinamphibianmetamorphosis.

Bioscience42:671–678

NiCC(2003)Off-shoreIslandsofTaiwan.WalkerCulture,Taipei

Pechenik JA(2006)Larvalexperienceandlatenteffects—metamor- phosisisnotanewbeginning.IntegrCompBiol46:323–333

PotvinC,LechowiczMJ,TardifS(1990)Thestatisticalanalysisof

ecophysiological response curves obtained from experiments involvingrepeatedmeasures.Ecology71:1389–1400

RoseCS(2005)Integrating ecologyanddevelopmentalbiologyto explainthetimingoffrogmetamorphosis.TrendsEcolEvol

20:129–135

RudolfVHW,Ro¨delMO(2007)Phenotypicplasticityandoptimal timingofmetamorphosis underuncertaintimeconstraints.Evol Ecol21:121–142

Semlitsch RD,ScottDE,Pechmann JHK(1988)Timeandsizeat metamorphosisrelatedtoadultfitnessinAmbystomatalpoideum. Ecology69:184–192

SilleroN,RibeiroR(2010)Reproduction ofPelophylaxpereziin brackishwaterinPorto(Portugal).HerpetolNotes3:337–340

SmithDC(1987)Adultrecruitmentinchorusfrogs:effectsofsize anddateatmetamorphosis.Ecology68:344–350

Squires ZE,BaileyPCE,ReinaRD,Wong BBM(2010) Compen- satorygrowthintadpoles aftertransient salinitystress.Mar FreshwRes61:219–222

Sumida M,KotakiM,IslamMM,Djong TH,IgawaT,KondoY, MatsuiM,AnslemDS,Khonsue W,Nishioka M(2007) Evolutionaryrelationships andreproductiveisolatingmecha- nisms in the Rice Frog (Fejervarya limnocharis) species complex fromSriLanka,Thailand,TaiwanandJapan,inferred frommtDNAgenesequences,allozymes,andcrossingexperi- ments.ZoolSci24:547–562

UchiyamaM,YoshizawaH(1992)Salinitytolerenceandstructureof externalandinternalgillsintadpoles ofthecrab-eatingfrog, Ranacancrivora.CellTissueRes267:35–44

UltschGR,BradfordDF,FredaJ(1999)Physiology,copingwiththe environment.In:McDiarmidRW,AltigR(eds)Tadpoles,the biologyofanuranlarvae.University ofChicagoPress,Chicago, pp189–214

VoneshJR,Warkentin KM(2006)Oppositeshiftsinsizeat metamorphosisinresponsetolarvalandmetamorphpredators. Ecology87:556–562

WernerEE(1986)Amphibianmetamorphosis:growthrate,predation riskandtheoptimalsizeattransformation.AmNat128:319–341

Wilbur HM, Collins JP (1973) Ecological aspects of amphibians

metamorphosis.Science182:1305–1314

WuCS,KamYC(2009)Effectsofsalinityonsurvival, growth, development,andmetamorphictraitsofFejervaryalimnocharis tadpoleslivinginbrackishwater.ZoolSci26:476–482

YangYR(1998)AfieldguidetothefrogsandtoadsofTaiwan.

ChinesePhotographyAssociation,Taipei