macroporositycomodo官网

Size-dependentmortalityinaNeotropicalsavannatree:theroleofheight-relatedadjustmentsinhydraulicarchitectureandcarbonallocationpce_20121456.
.
1466YONG-JIANGZHANG1,2,3,FREDERICKC.
MEINZER4,GUANG-YOUHAO1,2,3,FABIANG.
SCHOLZ5,SANDRAJ.
BUCCI5,FREDERICOS.
C.
TAKAHASHI6,RANDOLVILLALOBOS-VEGA2,JUANP.
GIRALDO7,KUN-FANGCAO1,WILLIAMA.
HOFFMANN8&GUILLERMOGOLDSTEIN2,91KeyLaboratoryofTropicalForestEcology,XishuangbannaTropicalBotanicalGarden,ChineseAcademyofSciences,Mengla,Yunnan666303,China,2DepartmentofBiology,UniversityofMiami,POBox249118,CoralGables,FL33124,USA,3GraduateSchooloftheChineseAcademyofSciences,Beijing100039,China,4USDAForestService,ForestrySciencesLaboratory,3200SWJeffersonWay,Corvallis,OR97331,USA,5ConsejoNacionaldeInvestigacionesCienticasyTecnicas(CONICET)andLaboratoriodeEcologiaFuncional,DepartamentodeBiologia,UniversidadNacionaldelaPatagoniaSanJuanBosco,ComodoroRivadavia,Argentina,6DepartamentodeEcologia,UniversidadedeBrasilia,CaixaPostal04457,Brasilia,DF70904970,Brazil,7DepartmentofOrganismicandEvolutionaryBiology,HarvardUniversity,Cambridge,MA02138,USA,8DepartmentofPlantBiology,NorthCarolinaStateUniversity,Raleigh,NC27695-7612,USAand9ConsejoNacionaldeInvestigacionesCienticasyTecnicas(CONICET)andLaboratoriodeEcologíaFuncional,DepartamentodeEcologia,GeneticayEvolucion,FacultaddeCienciasExactasyNaturales,UniversidaddeBuenosAires,CiudadUniversitaria,Nuez,BuenosAires,ArgentinaABSTRACTSize-relatedchangesinhydraulicarchitecture,carbonallocationandgasexchangeofSclerolobiumpaniculatum(Leguminosae),adominanttreespeciesinNeotropicalsavannasofcentralBrazil(Cerrado),wereinvestigatedtoassesstheirpotentialroleinthediebackoftallindivi-duals.
Treesgreaterthan~6-m-tallexhibitedmorebranchdamage,largernumbersofdeadindividuals,higherwooddensity,greaterleafmassperarea,lowerleafareatosapwoodarearatio(LA/SA),lowerstomatalconductanceandlowernetCO2assimilationthansmalltrees.
Stem-specichydraulicconductivitydecreased,whileleaf-specichydraulicconductivityremainednearlyconstant,withincreasingtreesizebecauseoflowerLA/SAinlargertrees.
Leavesweresubstantiallymorevulnerabletoembolismthanstems.
Largetreeshadlowermaximumleafhydraulicconductance(Kleaf)thansmalltreesandalltreesizesexhib-itedlowerKleafatmiddaythanatdawn.
Thesesize-relatedadjustmentsinhydraulicarchitectureandcarbonallocationapparentlyincurredalargephysiologicalcost:largetreesreceivedalowerreturnincarbongainfromtheirinvest-mentinstemandleafbiomasscomparedwithsmalltrees.
Additionally,largetreesmayexperiencemoreseverewaterdecitsindryyearsduetolowercapacityforbufferingtheeffectsofhydraulicpath-lengthandsoilwaterdecits.
Key-words:carbonbalance;hydraulicconductivity;popula-tiondynamics;treedieback;xylemcavitation.
INTRODUCTIONTreemortalityandtheconsequentreleaseofcarbonandnutrientsareimportantprocessesthatinuencethestruc-ture,composition,dynamicsandfunctioningofwoodyeco-systems(Franklin,Shugart&Harmon1987).
Treedieback,thesynchronizedmortalityofanentirepopulationoracohortofthatpopulation,isaphenomenonthathasbeendocumentedinavarietyoftreespeciesandecosystemtypes(e.
g.
Watt1987;Woodman1987;Gerrish,Mueller-Dombois&Bridges1988;Crombie&Tippett1990;MacGregor&O'Connor2002;Riceetal.
2004).
Factorscontributingtotreediebackincludeairpollution(Woodman1987),her-bivorybyinsects(Haugen&Underdown1990),fungalinfection(Crombie&Tippett1990),climatechange(Watt1987),nutrientlimitation(Gerrishetal.
1988)anddrought(MacGregor&O'Connor2002;Riceetal.
2004).
Fre-quently,deathoftreesisattributedtomorethanonetrig-geringagent,andistheresultofacombinationofbioticandabioticfactors(Manion&Lachance1992).
However,diebackcouldalsobeanaturalrecurringphenomenonresultingfromcohortsenescence(Mueller-Dombois1985)orrelatedtoreproduction(Foster1977)ratherthanfromdiseasesorenvironmentalstresses.
DiebackinAustralianandAfricansavannashasbeenattributedtotheeffectsofdrought(MacGregor&O'Connor2002;Riceetal.
2004),buttoourknowledge,notreediebackhasbeenreportedandnostudieshavebeendonetolinktreephysiologicalprocessesandpopulationdynamicsforNeotropicalsavannas.
Sincesavannaecosystemsareusuallycharacterizedbystrongseasonalityinprecipitation,plant–waterrelationsareconsideredtobekeydeterminantsoftheirstructureCorrespondence:F.
C.
Meinzer.
Fax:+15417587760;e-mail:rick.
meinzer@oregonstate.
eduPlant,CellandEnvironment(2009)32,1456–1466doi:10.
1111/j.
1365-3040.
2009.
02012.
x2009BlackwellPublishingLtd1456andcomposition(Huntley&Walker1982).
Drought,ingeneral,notonlyinuencesthewaterrelationsandhydrau-licpropertiesofplantsbutalsonegativelyaffectsphotosyn-thesis(Chavesetal.
2002),aswellasresistancetoherbivory(Lowman&Heatwole1992)andinfectionbyfungi(Gibbsetal.
1990).
CentralBraziliansavannas(Cerrado)aresub-jectedtoapronounceddryseasonthatmaylastforaslongas5months.
Waterdecitsplayacrucialroleinthegrowthandphysiologyofsavannatreesbecauseofthehighatmo-sphericevaporativedemandandthelimitedamountofwaterthattreescanobtaindailyfromtheuppersoillayersduringthedryseason(Meinzeretal.
1999).
Althoughmanysavannawoodyspecieshavedeeprootsystemsthataccesswateravailableatdepthduringthedryseason(Rawitscher1948;Oliveiraetal.
2005),theparadigmthattheyallhavesimilaraccesstodeepsoilwaterreservesdoesnotappeartobeuniversallyapplicable(Jacksonetal.
1999;Buccietal.
2005;Goldsteinetal.
2008;Scholzetal.
2008).
Waterdecitsareampliedintalltreesowingtoincreasedtensioninthexylemrequiredtodrawwaterfromthesoiltothecanopy.
Increasedtensioninxylemconduitsintalltreesmayinduceembolismandhydraulicdysfunc-tion,leadingtoincreasedwaterdecits,decreasedstomatalconductance(gs)andphotosynthesis,lowergrowthrates(Ryan&Yoder1997;Kochetal.
2004;Woodruff,Bond&Meinzer2004;Ryan,Phillips&Bond2006;Domecetal.
2008)and,eventually,treemortality.
Inordertomitigatetheeffectsofincreasedwaterdecits,treesmaydecreasegs,down-regulatephotosynthesis,intensifyrellingofembo-lizedxylemand/ormodifytheirhydraulicarchitecture,inparticular,theleafareatosapwoodarearatio(LA/SA)(McDowelletal.
2002).
Recently,ithasbeenshownthatleavesareahydraulicbottleneckinthewatertransportpathwayoftheplant,representingabout30%ofwholeplantresistanceforarangeoflifeforms(Sacketal.
2003)andthatmaximumleafhydraulicconductance(Kleaf)decreaseswithincreasingheight(Woodruff,Meinzer&Lachenbruch2008).
Therelationshipbetweenadjustmentinhydraulicarchitecturewhengrowingtallerandwhole-treecarbonbalance,aswellastheirrolesintreemortality,however,ispoorlyunderstood.
SclerolobiumpaniculatumVog.
(Leguminosae)isadomi-nanttallevergreensavannaspeciesexhibitingconspicuousbranchdiebackandtreemortalityamonglargerindividuals.
Itisafast-growingpioneerspecies(Pires&Marcati2005)andisamongthefewBrazilianCerradotreespecieswithrelativelyshallowrootsystems(Jacksonetal.
1999;Scholz2006).
Theobjectiveofthisstudywastoidentifypotentialcausalrelationshipsbetweensize-relatedchangesintreehydraulicarchitecture,carbonallocation,growth,gasexchange,waterdecitsandmortalityinS.
paniculatumtreesgrowinginaCerradositewhererehasbeenexcludedduringthelast35years.
Weinvestigatedstemandleafhydraulicproperties,leafwaterstatus,growthrates,gasexchangeandotherfunctionaltraitssuchaswooddensityandleafmassperarea(LMA)inS.
paniculatumtreesofdifferentheights.
MATERIALSANDMETHODSStudysiteandplantmaterialThisresearchwascarriedoutinasavannasiteattheInsti-tutoBrasileirodeGeograaeEstatística(IBGE)reserve,aeldexperimentalstationlocated35kmsouthofBrasilia(15°56′S,47°53′W,elevation1100m).
Averageannualpre-cipitationinthereserveis1500mmwithapronounceddryseasonfromMaytoSeptember.
Averagerelativehumidityduringthedryseasonis55%andminimumrelativehumid-itycandroptovaluesaslowas10%.
Meanmonthlytem-peraturesrangefrom19to23°C.
Thesoilsareverynutrientpoor,deepandwell-drainedoxisols.
Soilbulkdensityisabout0.
99gcm-3,macroporosityisabout18%andtexturefraction(silt/clay)isabout0.
22intheupper100cmofatypicalsoilprole(Buccietal.
2008).
TheIBGEreservecontainsallmajorphysiognomictypesofsavannasfromveryopentoclosedsavannas.
Sclerolobiumpaniculatumisadominantspeciesinsavannaswithahightreedensityandexhibitshighchronicmortalityoflargetrees.
A'cerradodenso'sitewithrelativelyuniformtopogra-phyandsoilcharacteristicsandintermediatetreedensitywaschosentominimizeshadingeffectsandincreasethelikelihoodthatsmallandlargetreesexperiencedsimilarlightregimes.
Firehadbeenexcludedfromthesiteformorethan35years.
AlloftheS.
paniculatumtreesinsidea200-m-long,40-m-widetransectweresurveyedandlatercatego-rizedintofourheightclassesformeasurementsoffunctionaltraits(Table1).
Boththeheightandstemdiam-eterat1.
3m(DBH)oftheindividualsusedforphysiologi-calstudiesweredetermined.
ThepercentageofdeadbranchespertreeofeachS.
paniculatumindividualwasdeterminedbycountingthetotalnumberofdeadbranchesperplantinsmallertrees,orvisuallyestimatingthepercent-agesofdeadbranchesinthecrownsoftallertreesinwhichTable1.
Numberofindividualsusedforphysiologicalmeasurements,height,DBHandrelativeabundanceoftreesineachsizeclassTreeheightclass(m)NumberofindividualsHeight(m)DBH(cm)Relativeabundance(%)8118.
950.
1916.
260.
8842TheheightandDBHvaluesaremeansSEfromallthesampledtreesusedtomeasurephysiologicalvariables.
TherelativeabundancecorrespondstoallSclerolobiumpaniculatumtreesinthestudysite,eveniftheyweredead.
Size-dependentmortalityinaNeotropicalsavannatree14572009BlackwellPublishingLtd,Plant,CellandEnvironment,32,1456–1466branchesweretoonumeroustoaccuratelycountfromtheground.
LeafwaterpotentialApressurechamber(PMS,Corvallis,OR,USA)wasusedtomeasureleafwaterpotential(YL)duringthedryseasonof2006.
Sixto10newestfullydevelopedmatureleavesfromsun-exposedterminalbranchesofdifferentindivi-dualswereselectedtomeasuredawnandmiddayYL.
SamplesfordawnYLwerecollectedbetween0630and0730h,whileleavesformiddaymeasurementswerecol-lectedbetween1230and1400h.
PreviousstudiesattheIBGEreservehaveshownthatdailymaximum(leastnega-tive)valuesofYLaretypicallyattainedbetween0600and0630hwithadeclineof0.
2MPaby0730h(Buccietal.
2003,2004a).
Thetoptwoleaetsofthelargecompoundleaveswereexcised,immediatelysealedinplasticbagsandkeptinacoolerwithsmallamountoficeuntilbalancingpressuresweredeterminedinthelaboratorywithin1hofsamplecollection.
Kleaf,pressure-volumerelationshipsandleafcapacitanceKleafwasestimatedusingthepartialrehydrationmethod(Brodribb&Holbrook2003).
SamplesfordawnKleafwerecollectedbetween0630and0730h,whileleavesformiddaymeasurementswerecollectedbetween1230and1400h.
Largebrancheswerecut,baggedandkeptinthedarkwithslightlywetpapertowelsforabout30minforwaterpoten-tialequilibrationofleavesandleaets.
TwoleaveswerechosentomeasuretheinitialYLinthetwotopleaetsofthecompoundleaf.
Then,thetoptwoleaetsoftwoadja-centleaveswerecutfromtherachisunderwater,andallowedtoabsorbwaterfor3to15sdependingontheinitialwaterpotential,afterwhich,theirrachisendsweredriedcarefully.
ThenalvalueofYLwasimmediatelymea-suredusingthepressurechamber.
Kleafwasthencalculatedfromtheequation:KCtleafof=*()lnΨΨ(1)whereCistheleafcapacitance,YoistheYLpriortorehy-drationandYfistheYLafterrehydrationfortseconds.
ValuesofleafcapacitanceusedtocalculateKleafwerederivedfrompressure–volumerelationships(Tyree&Hammel1972)usingthemethoddescribedbyBrodribb&Holbrook(2003).
Briey,theYLcorrespondingtoturgorlosswasestimatedastheinectionpoint(thetransitionfromtheinitialcurvilinear,steeperportionofthecurvetothemorelinearlesssteepportion)ofthegraphofYLversusrelativewatercontent(RWC).
TheslopeofthecurvepriortoandfollowingturgorlossprovidedCintermsofRWC(CRWC)forpre-turgorlossandpost-turgorloss,respectively.
LeafareatodrymassratioswereusedtonormalizeConaleafareabasis.
Pressure–volumecurvesweredeterminedforsixfullydevelopedexposedleavesfromdifferentindividuals.
Leavesforpressure–volumeanalyseswereobtainedfrombranchescutintheeldintheearlymorning,re-cutimmediatelyunderwaterandcoveredwithblackplasticbagswiththecutendinwaterforabout2huntilmeasurementsbegan.
Datawerettedbyapressure–volumeprogramdevelopedbySchulte&Hinckley(1985).
LeafvulnerabilitycurveswereplottedasKleafagainstinitialYLbeforerehydration.
ArangeofYLwasattainedthroughslowbenchdryingofleafybranchescollectedfromtheeldatdawn.
Branchesweredehydratedfordifferenttimeperiods,afterwhich,theywereenclosedinblackplasticbagswithslightlywetpapertowels.
After0.
5to1hequilibrationperiod,YL,beforeandafterrehydration,weremeasured.
StemhydraulicconductivityHydraulicconductivity(kh)wasmeasuredonsun-exposedterminalbranchesexcisedatdawn(0630to0730h)andmidday(1300to1400h)fromfourtoveindividualsofeachheightclassexceptthesmallest(6mtallexhibitedslightlylowervulnerabilitytoembolismthanthoseofsmalltrees.
TheY50was-0.
8MPainleavesoftrees8mtall(Fig.
3).
They-interceptofthefunctionttedtoYLversusKleafrelationships(KleafataYLof0MPa)was22mmolm-2s-1MPa-1fortrees>8mtall,whileitrangedfrom30to40mmolm-2s-1MPa-1fortreesintheotherheightclasses.
Leafturgorlosspointsrangedbetween-1.
5and-1.
7MPa,andcoincidedapproximatelywiththeYLatwhichKleafapproachedzero(Fig.
3).
Stem-specichydraulicconductivity(ks)didnotchangesignicantlybetweendawnandmidday(Fig.
4).
Ontheotherhand,middayKleafwassubstantiallylowercomparedwithdawnvaluesintreesfromallsizeclasses(Fig.
4).
Therewasamarginallysignicant(P=0.
06)declineindawnKleafwithincreasingtreeheight.
ThemiddayKleafvalueswereslightlyhigherthanthevaluesderivedfromleafvulnerabil-itycurvesandmiddayYL,probablyasaresultoftheequili-brationbetweenleafandstemwaterpotentials,whilethebranchesremainedbaggedbeforemeasurements.
Sincenosignicantdifferencewasfoundbetweendawnandmiddaykh,ksandkl(datanotshownforkhandkl),themeansofmiddayanddawnvalueswereusedinFig.
5.
Therewasasignicant(P=0.
003)height-relateddeclineinkh(Fig.
5).
ksalsodecreasedsignicantly(P6mtallfallingfurtherbelowtheYLatturgorlossthanintrees8mtall.
Inplantsfromaridenvironments,rehydra-tionofsamplesforpressure–volumeanalysismaysome-timescauseartefactsbyshiftingtheturgorlosspointtolessnegativevalues(Meinzeretal.
1986).
However,CerradotreesexperienceYLcloseto0MPaatnightandFigure3.
Leafandstemhydraulicvulnerabilitycurvesfortreesindifferentheightclasses.
Asigmoidfunctionwasttedtothedata(P8Kleaf(mmolm–2s–1MPa–1)0102030*******Figure4.
Dawn(blackbars)andmidday(greybars)stem-specichydraulicconductivity(ks),andleafhydraulicconductance(Kleaf),oftreesfromdifferentheightclasses.
ksoftreesshorterthan3mwasnotmeasured(seemethods).
BarsaremeansSEforn=4to6.
SignicantdifferencesbetweendawnandmiddayKleafwerefound(**P8LA/SA(m2cm–2)0.
00.
10.
20.
30.
40.
50.
60.
7Treeheight(m)3to66to8>8Kl(x10–4kgm–1s–1MPa–1)024681012141462Y.
-J.
Zhangetal.
2009BlackwellPublishingLtd,Plant,CellandEnvironment,32,1456–1466atabout-1MPa,whereasstemswere50%embolizedbelow-3.
2MPa(Fig.
3).
StemxylemwasoperatingfarfromthepointofcatastrophicdysfunctionsensuTyree&Sperry(1988),andthereforeterminalstemshadawidersafetymarginintermsofwaterdecitsthanleaves.
LeafwaterpotentialsatwhichKleafreachedminimumvalueswereclosetotheirturgorlosspoints.
ThemortalityofS.
panicu-latumtreesisnotlikelytoresultfromcatastrophicstemxylemdysfunctionduringperiodsofdroughtbecauseminimumstemwaterpotentialsintheeldwerenotonlywellabovethosecorrespondingtoP50,butalsowellabovethewaterpotentialthresholdatwhichcavitationstartstoincreaseaccordingtothestemxylemvulnerabilitycurves.
Ontheotherhand,YLintheeldandleafvulnerabilitycurvessuggestthatembolisminleavesoccurredregularly.
LowermaximumKleafintallerS.
paniculatumtreesmayreectheight-relatedtrendsinxylemstructureassociatedwithreducedleafexpansion(Woodruffetal.
2008).
LeavesofS.
paniculatumshowedaconsistentdailypatternofdepressionandrecoveryofKleaf(Fig.
4).
DailychangesinKleafcouldbepartiallyexplainedbyembolismformationduringthemorningwhenYLdecreasedasevapo-rativedemandandtranspirationincreased.
Intheafternoonoratnight,Kleafincreased,consistentwithdailyembolismrepair.
ResultsofdyeexperimentsbyBuccietal.
(2003)supportthehypothesisthatdielvariationofpetiolekhofsavannatreeswasassociatedwithembolismformationandrepair.
Dyeexperimentsaredifculttoperformwiththeleaflamina,butresultsobtainedwithothertechniques,LMA(gcm–2)0.
0160.
0170.
0180.
0190.
0200.
021Leafsize(cm2)200250300350Treeheight(m)8Wooddensity(gcm–3)0.
560.
580.
600.
620.
640.
660.
68Figure6.
Leafmassperarea(LMA),leafsizeandwooddensityoftreesindifferentheightclasses.
BarsaremeansSEofsixindividuals.
gs(mmolg–1s–1)0123456A(mmolg–1s–1)0.
000.
020.
040.
060.
080.
100.
120.
14Treeheight(m)0246810A/gs(mmolmol–1)020406080100r2=0.
40r2=0.
30r2=0.
31Figure7.
Stomatalconductance(gs),netassimilationrate(A)andintrinsicwateruseefciency(A/gs)inrelationtoheightofS.
paniculatumtrees.
Thelinesarelinearregressionsttedtothedata.
Size-dependentmortalityinaNeotropicalsavannatree14632009BlackwellPublishingLtd,Plant,CellandEnvironment,32,1456–1466suchascryoscanningelectronmicroscopy(Canny2001;Woodruffetal.
2007)andacousticemission(e.
g.
LoGulloetal.
2003;Johnsonetal.
2009),indicatethatcavitationcommonlyoccursinleavesofmanyspecies.
Cochardetal.
(2004),ontheotherhand,suggestedthatdiurnalvariationsinKleafofconiferscouldbepartiallyexplainedbychangesinconduitdimensionsundercyclesoftensionincreaseandtensionrelease,reversiblyconstrictingthewaterowthroughtheleaf.
RegardlessofthemechanismgoverningdiurnalchangesinKleaf,thereversiblelossofhydraulicconductanceintheleaflaminamaybeanadaptivemeansofamplifyingtheevaporativedemandsignaltothestomatainordertoexpe-diteastomatalresponse(Brodribb&Holbrook2004).
Althoughwedidnotnddailychangesinksofstemsinourstudy,diurnaldepressionandrecoveryofstemkshasbeenobservedinotherspecies(Zwieniecki&Holbrook1998;Melcheretal.
2001).
Embolismformationandrellingisprobablyoflimitedsignicanceforstemsofwoodyplantsatmosttimesbecauseofthehighenergeticcostsofrellingalargevolumeoftissuethatisdistantfromthesitesofcarbohydratesynthesisintheleaves.
Comparedwithstems,diurnalrellinginleavesmaybelessenergeticallycostlyandmayinvolvesimplermechanisms(Buccietal.
2003;Brodribb&Holbrook2004).
Sclerolobiumpaniculatumstemswerelessvulnerabletocavitationthanleavesononehand,andalsowereprotectedbyregulationoftranspira-tionthroughdiurnaldepressionandrecoveryofKleaf.
DuetothehighsafetyofS.
paniculatumstems,andtheeffectiveregulationofKleafandgs,thediebackoftallindividualscouldnotbeexplainedbycatastrophicxylemdysfunctionwhenexperiencingdroughtbutmayberelatedtochangesinwhole-treewaterandcarbonbalanceresultingfromsize-relatedstructuralchanges,asdiscussedbelow.
CarbonbalanceinrelationtohydraulicarchitectureandcarbonallocationTreesadjustedtheirbranchhydraulicarchitecturewithincreasingheight,resultinginlowerLA/SAintallertreesandarelativelyconstantkl,sotheamountofwaterthatcouldbedeliveredbythevascularsystemperunitleafareawassimilarindifferentsizetreesdespiteasharpreductioninbranchkswithincreasingtreesize.
Theserelationshipswerenotdirectlycharacterizedatthewhole-treelevel,butsize-dependentincreasesinbranchdiebackandsimilarvaluesofdawnandmiddayYLacrosssizeclasseswereconsistentwithmaintenanceofwhole-treeleaf-specicconductance.
SeasonaladjustmentsinleafareaofCerradotreeshavebeenshowntoreduceseasonalvariationinmiddayYLandwhole-plantleaf-specicconductance(Buccietal.
2005).
Otherstudieshaveshowncompensatoryadjustmentsintreeallometryandstandstructurethatcon-tributetohomeostasisofminimumYL(Whitehead,Jarvis&Waring1984;Williams&Cooper2005).
Nevertheless,adjustmentsthatmaintainanadequatewaterbalancecouldresultinanunsustainablesituationintermsofwhole-treecarbonbalanceinS.
paniculatum.
ThelargestS.
paniculatumtreesarelikelytoreceiveasubstantiallylowerreturnincarbongainfromtheirinvestmentinstemandleafbiomasscomparedwithsmallertreesasexplainedinthefollowingexercisebasedonourdataonsize-dependentchangesinallometry,carbonallocationandgasexchange.
Foragiveninvestmentinsapwoodarea,thelargesttrees(>8mtall)displayonly55%asmuchleafareacomparedwith3to6-m-talltrees(LA/SA=0.
33and0.
60m2cm-2intrees>8mtalland3to6mtall,respectively).
WhenthedifferencesabovearenormalizedbythedifferenceinbranchwooddensityandLMAbetweenthelargesttreesandsmallertrees(anestimateoftheamountofleafmassdevelopedperbiomassinvested),thenthelargesttreesdevelopedonly54%asmuchleafmassperbiomassinvestedinstemtissue.
Furthermore,accordingtothelinearrelationshipbetweenA(massbasednetCO2uptake)andH(treeheight)(A=0.
122-0.
0062H),Aintrees10mtallwouldbeabout64%ofthatintrees3to6mtall.
Thus,largetreesmayonlyget35%asmuchcarbonreturnperbiomassinvestedassmalltreesdo.
ItisnoteworthythatAofdetachedshootsdeclinedby~50%between1and10m(Fig.
7),suggestingthatthisheight-relatedchangewasasso-ciatedwithinherentleafstructuralandphysiologicalcon-straintsongasexchange(e.
g.
Niinemets2002;Woodruffetal.
2009)ratherthanextrinsiceffectsofhydraulicpath-lengthresistances.
Gasexchangeofattachedshootsislikelytodeclinemoresteeplywithincreasingheightbecauseofhydraulicconstraints(Schfer,Oren&Tenhunen2000;McDowell,Licata&Bond2005).
Sclerolobiumpaniculatumisafast-growingpioneerspecieswitharelativelyshortlifecycle.
Thesespeciesshouldhaveallocationpatternsthatfavoursurvivalandgrowthofyoungtrees,eveniftheperformanceofthesametreeiscompromisedwhenolder(Williams1957).
InS.
pan-iculatum,patternsofhydraulicarchitectureadoptedearlyindevelopmentappeartoconstrainwaterandcarbonrela-tionslateindevelopment.
Highadultmortalityhasalsobeenfoundinmonocarpicspecies(Stearns1992),alifehistorystrategyinwhichindividualsreproduceonlyonceandsubsequentlydie.
SeveralspeciesofTachigali,agenusnowconsideredsynonymouswithSclerolobium(Lewis2005),aremonocarpic(Foster1977;Poorteretal.
2005),anextremelyrarecharacteristicfortropicaltrees.
IftheobservedpatternsofhydraulicarchitectureandcarbonimbalancerepresentmoregeneralcharacteristicsoftheSclerolobium–Tachigaliclade,thissuggeststhathydraulicconstraintsmayhaveresultedinthelife-historytrade-offresponsibleforthemonocarpyinsomeTachigalispecies.
TheheightoftallS.
paniculatumtreesmayconferanadvantageincompetitionforlightandestablishingdomi-nance,particularlyindenseCerradophysiognomies,butmayalsocarrytheriskofgreaterwaterdecitsinexcep-tionallydryyears.
Compensatoryadjustmentsinleafandbranchhydraulicarchitecturewithincreasingtreeheightappearedtocarryalargephysiologicalcost:apoorreturnincarbongainforagiveninvestmentinstemandleafbiomasscomparedwithsmalltrees,whicheventuallycouldleadtodiebackofthewholetree.
Ourresultsprovideapotential1464Y.
-J.
Zhangetal.
2009BlackwellPublishingLtd,Plant,CellandEnvironment,32,1456–1466explanationforthemassmortalityinlargeS.
paniculatumtrees,aswellasinsightsintothephysiologicalcostsofsize-relatedchangesincarbonallocationpatterns.
ACKNOWLEDGMENTSWethanktheReservaEcologicadoInstitutoBrasileirodeGeograaeEstatistica(IBGE)forlogisticsupport.
WealsothankEricManzanéforhelpwitheldwork,aswellasDavidJanos,CatalinaAristizábal,TaniaWyss,SarahGaramzegiandXinWangforhelpfulcommentsonthemanuscript.
ThisstudywassupportedbyNationalScienceFoundation(USA)grants#0296174and#0322051.
ThisworkcomplieswithBrazilianlaw.
REFERENCESAbramoffM.
D.
,MagelhaesP.
J.
&RamS.
J.
(2004)Imageprocess-ingwithImagej.
BiophotonicsInternational11,36–42.
BrodribbT.
J.
&HolbrookN.
M.
(2003)Stomatalclosureduringleafdehydration,correlationwithotherleafphysiologicaltraits.
PlantPhysiology132,2166–2173.
BrodribbT.
J.
&HolbrookN.
M.
(2004)Stomatalprotectionagainsthydraulicfailure:acomparisonofcoexistingfernsandangiosperms.
NewPhytologist162,663–670.
BucciS.
J.
,ScholzF.
G.
,GoldsteinG.
,MeinzerF.
C.
,SternbergL.
&DaS.
L.
(2003)Dynamicchangesinhydraulicconductivityinpetiolesoftwosavannatreespecies:factorsandmechanismscontributingtotherellingofembolizedvessels.
Plant,Cell&Environment26,1633–1645.
BucciS.
J.
,ScholzF.
G.
,GoldsteinG.
,MeinzerF.
C.
,HinojosaJ.
A.
,HoffmannW.
A.
&FrancoA.
C.
(2004a)Processespreventingnocturnalequilibrationbetweenleafandsoilwaterpotentialintropicalsavannawoodyspecies.
TreePhysiology24,1119–1127.
BucciS.
J.
,GoldsteinG.
H.
,MeinzerF.
C.
,ScholzF.
G.
,FrancoA.
C.
&BustamanteM.
(2004b)Functionalconvergenceinhydraulicarchitectureandwaterrelationsoftropicalsavannatrees:fromleaftowholeplant.
TreePhysiology24,891–899.
BucciS.
J.
,GoldsteinG.
,MeinzerF.
C.
,FrancoA.
C.
,CampanelloP.
&ScholzF.
G.
(2005)Mechanismscontributingtoseasonalhomeostasisofminimumleafwaterpotentialandpredawndis-equilibriumbetweensoilandplantsinNeotropicalsavannatrees.
Trees19,296–304.
BucciS.
J.
,ScholzF.
G.
,GoldsteinG.
,HoffmannW.
A.
,MeinzerF.
C.
,FrancoA.
C.
,GiambellucaT.
&Miralles-WilhelmF.
(2008)Con-trolsonstandtranspirationandsoilwaterutilizationalongatreedensitygradientinaNeotropicalsavanna.
AgriculturalandForestMeteorology148,839–849.
CannyM.
(2001)Embolismandrellinginthemaizeleaflaminaandtheroleoftheprotoxylemlacuna.
AmericanJournalofBotany88,47–51.
ChavesM.
M.
,PereiraJ.
S.
,MarocoJ.
,RodriguesM.
L.
,RicardoC.
P.
,OsorioM.
L.
,CarvalhoI.
,FariaT.
&PinheiroC.
(2002)HowplantscopewithwaterstressintheeldPhotosynthesisandgrowth.
AnnalsofBotany89,907–916.
CochardH.
,FrouxF.
,MayrS.
&CoutardC.
(2004)Xylemwallcollapseinwater-stressedpineneedles.
PlantPhysiology134,401–408.
CrombieD.
S.
&TippettJ.
T.
(1990)Acomparisonofwaterrela-tions,visualsymptoms,andchangesinstemgirthforevaluatingimpactofPhytophthoracinnamomionEucalyptusmarginata.
CanadianJournalofForestResources20,233–240.
DomecJ.
-C.
,LachenbruchB.
,MeinzerF.
C.
,WoodruffD.
R.
,WarrenJ.
M.
&McCullohK.
A.
(2008)Maximumheightinaconiferisassociatedwithconictingrequirementsforxylemdesign.
ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica105,12069–12074.
FosterR.
B.
(1977)TachigaliaversicolorisasuicidalNeotropicaltree.
Nature268,624–626.
FranklinJ.
F.
,ShugartH.
H.
&HarmonM.
E.
(1987)Treedeathasanecologicalprocess.
Bioscience37,550–556.
GerrishG.
,Mueller-DomboisD.
&BridgesK.
W.
(1988)NutrientlimitationandMetrosiderosforestdiebackinHawaii.
Ecology69,723–727.
GibbsR.
,CraigI.
,MyersB.
J.
&YuanZ.
Q.
(1990)EffectofdroughtanddefoliationonthesusceptibilityofeucalyptstocankerscausedbyEndothiagyrosaandBotryoshaeriaribis.
AustralianJournalofBotany38,571–581.
GoldsteinG.
,MeinzerF.
C.
,BucciS.
J.
,ScholzF.
G.
,FrancoA.
C.
&HoffmannW.
A.
(2008)WatereconomyofNeotropicalsavanna:sixparadigmsrevisited.
TreePhysiology28,395–404.
HaugenD.
A.
&UnderdownM.
G.
(1990)Sirexnoctiliocontrolprograminresponsetothe1987GreenTriangleoutbreak.
AustralianForestry53,33–40.
HuntleyB.
J.
&WalkerB.
H.
(1982)EcologyofTropicalSavanna.
Springer,Heidelburg,Germany.
JacksonP.
C.
,MeinzerF.
C.
,BustamanteM.
,GoldsteinG.
,FrancoA.
,RundelP.
W.
,CaldasL.
S.
,IglerE.
&CausinF.
(1999)Parti-tioningofsoilwateramongtreespeciesinaBrazilianCerradoecosystem.
TreePhysiology36,237–268.
JamesS.
A.
,ClearwaterM.
J.
,MeinzerF.
C.
&GoldsteinG.
(2002)Heatdissipationsensorsofvariablelengthforthemeasurementofsapowintreeswithdeepsapwood.
TreePhysiology22,277–283.
JohnsonD.
M.
,MeinzerF.
C.
,WoodruffD.
R.
&McCullohK.
A.
(2009)Leafxylemembolism,detectedacousticallyandbycryo-SEM,correspondstodecreasesinleafhydraulicconductanceinfourevergreenspecies.
Plant,Cell&Environment32,828–836.
KochG.
W.
,SillettS.
C.
,JenningsG.
M.
&DavisS.
D.
(2004)Thelimitstotreeheight.
Nature428,851–854.
LewisG.
P.
(2005)LegumesoftheWorld.
RoyalBotanicGardens,Kew,Richmond,UK.
LoGulloM.
A.
,NardiniA.
,TriioP.
&SalleoS.
(2003)ChangesinleafhydraulicandstomatalconductancefollowingdroughtstressandirrigationinCeratoniasiliqua(Carobtree).
Physiolo-giaPlantarum117,186–194.
LowmanM.
D.
&HeatwoleH.
(1992)Spatialandtemporalvari-abilityindefoliationofAustralianeucalypts.
Ecology73,129–142.
McDowellN.
,BarnardH.
,BondB.
J.
,etal.
(2002)Therelationshipbetweentreeheightandleafarea:sapwoodarearatio.
Oecologia132,12–20.
McDowellN.
G.
,LicataJ.
&BondB.
J.
(2005)Environmentalsen-sitivityofgasexchangeindifferent-sizedtrees.
Oecologia145,9–20.
MacGregorS.
D.
&O'ConnorT.
G.
(2002)PatchdiebackofColo-phospermummopaneinadysfunctionalsemi-aridAfricansavanna.
AustralEcology27,385–395.
ManionP.
D.
&LachanceD.
(1992)ForestDeclineConcepts.
APSPress,St.
Paul,MN,USA.
MeinzerF.
C.
,RundelP.
W.
,ShariM.
R.
&NilsenE.
T.
(1986)TurgorandosmoticrelationsofthedesertshrubLarreatriden-tata.
Plant,Cell&Environment9,467–475.
MeinzerF.
C.
,GoldsteinG.
,FrancoA.
C.
,BustamanteM.
,IglerE.
,JacksonP.
,CaldasL.
&RundelP.
W.
(1999)AtmosphericandhydrauliclimitationsontranspirationinBrazilianCerradowoodyspecies.
FunctionalEcology13,273–282.
MeinzerF.
C.
,CampanelloP.
I.
,DomecJ.
C.
,GattiM.
G.
,GoldsteinG.
,Villalobos-VegaR.
&WoodruffD.
R.
(2008)ConstraintsonphysiologicalfunctionassociatedwithbrancharchitectureandSize-dependentmortalityinaNeotropicalsavannatree14652009BlackwellPublishingLtd,Plant,CellandEnvironment,32,1456–1466wooddensityintropicalforesttrees.
TreePhysiology28,1609–1617.
MelcherP.
J.
,GoldsteinG.
,MeinzerF.
C.
,YountD.
,JonesT.
,Hol-brookN.
M.
&HuangC.
X.
(2001)WaterrelationsofcoastalandestuarineRhizophoramangle:xylemtensionanddynamicsofembolismformationandrepair.
Oecologia126,182–192.
Mueller-DomboisD.
(1985)Ohi'adiebackandprotectionmanage-mentoftheHawaiianrainforest.
InHawaii'sTerrestrialEcosys-temsPreservationandManagement(edsC.
P.
Stone&J.
M.
Scott)pp.
403–421.
CooperativeNationalParkResourcesStudiesUnit,UniversityofHawaii,Honolulu,HI,USA.
Niinemets,.
(2002)StomatalconductancealonedoesnotexplainthedeclineinfoliarphotosyntheticrateswithincreasingtreeageandsizeinPiceaabiesandPinussylvestris.
TreePhysiology22,515–535.
OliveiraR.
S.
,BezerraL.
,DavidsonE.
A.
,PintoF.
,KlinkC.
A.
,NepstadD.
C.
&MoreiraA.
(2005)DeeprootfunctioninsoilwaterdynamicsincerradosavannasofcentralBrazil.
FunctionalEcology19,574–581.
PiresI.
P.
&MarcatiC.
R.
(2005)AnatomiaeusodamadeiradeduasvariedadesdeSclerolobiumpaniculatumVog.
dosuldoMaranhao,Brazil.
ActaBotanicaBrasilica19,669–678.
PoorterL.
,ZuidemaP.
A.
,Pena-ClarosM.
&BootR.
G.
A.
(2005)Amonocarpictreespeciesinapolycarpicworld:howcanTachigalivasqueziimaintainitselfsosuccessfullyinatropicalrainforestcommunityJournalofEcology93,268–278.
RawitscherF.
(1948)ThewatereconomyofthevegetationofthecamposcerradosinsouthernBrazil.
JournalofEcology36,237–267.
RiceK.
J.
,MatznerS.
L.
,ByerW.
&BrownJ.
R.
(2004)PatternsoftreediebackinQueensland,Australia:theimportanceofdroughtstressandtheroleofresistancetocavitaion.
Oecologia139,190–198.
RyanM.
J.
&YoderB.
J.
(1997)Hydrauliclimitstotreeheightandtreegrowth.
Bioscience47,235–242.
RyanM.
J.
,PhillipsN.
&BondB.
J.
(2006)Thehydrauliclimitationhypothesisrevisited.
Plant,Cell&Environment29,367–381.
SackL.
,CowanP.
D.
,JaikumarN.
&HolbrookN.
M.
(2003)The'hydrology'ofleaves:coordinationofstructureandfunctionintemperatewoodyspecies.
Plant,Cell&Environment26,1343–1356.
SalaA.
&HochG.
(2009)Height-relatedgrowthdeclinesinponderosapinearenotduetocarbonlimitation.
Plant,Cell&Environment32,22–30.
SchferK.
V.
R.
,OrenR.
&TenhunenJ.
D.
(2000)Theeffectoftreeheightoncrownlevelstomatalconductance.
Plant,Cell&Environment23,365–375.
ScholzF.
G.
(2006)Biosicadeltransportedeaguaenelsistemasuelo-planta:redistribucion,resistenciasycapacitanciashidrauli-cas.
PhDthesis,UniversityofBuenosAires,Argentina.
ScholzF.
G.
,BucciS.
J.
,GoldsteinG.
,MeinzerF.
C.
,FrancoA.
C.
&Miralles-WilhelmF.
(2007)Biophysicalpropertiesandfunc-tionalsignicanceofstemwaterstoragetissuesinNeotropicalsavannatrees.
Plant,Cell&Environment30,236–248.
ScholzF.
G.
,BucciS.
J.
,GoldsteinG.
,MoreiraM.
Z.
,MeinzerF.
C.
,DomecJ.
-C.
,VillalobosVegaR.
,FrancoA.
C.
&Miralles-WilhelmF.
(2008)Biophysicalandlifehistorydetermi-nantsofhydraulicliftinNeotropicalsavannatrees.
FunctionalEcology22,773–786.
doi:10.
1111/j.
1365-2435.
2008.
01452.
xSchulteP.
J.
&HinckleyT.
M.
(1985)Acomparisonofpressure–volumecurvedataanalysistechniques.
JournalofExperimentalBotany36,590–602.
StearnsS.
C.
(1992)TheEvolutionofLifeHistoryStrategies.
OxfordUniversityPress,Oxford,UK.
TyreeM.
T.
&HammelH.
T.
(1972)Themeasurementoftheturgorpressureandthewaterrelationsofplantsbythepressure-bombtechnique.
JournalofExperimentalBotany23,267–282.
TyreeM.
T.
&SperryJ.
S.
(1988)DowoodyplantsoperatenearthepointofcatastrophicxylemdysfunctioncausedbydynamicwaterstressPlantPhysiology88,574–580.
TyreeM.
T.
&SperryJ.
S.
(1989)Vulnerabilityofxylemtocavitationandembolism.
AnnualReviewofPlantPhysiologyandPlantMolecularBiology40,19–48.
WattK.
E.
F.
(1987)AnalternativeexplanationfortheincreasedforestmortalityinEuropeandNorthAmerica.
DanskSkov-foreningsTidsskrift72,210–224.
WhiteheadD.
,JarvisP.
G.
&WaringR.
H.
(1984)Stomatalconduc-tance,transpirationandresistancetowateruptakeinaPinussylvestrisspacingexperiment.
CanadianJournalofForestResearch14,692–700.
WilliamsC.
A.
&CooperD.
J.
(2005)Mechanismsofripariancottonwooddeclinealongregulatedrivers.
Ecosystems8,382–395.
WilliamsG.
C.
(1957)Pleiotropy,naturalselectionandtheevolu-tionofsenescence.
Evolution11,398–411.
WoodmanJ.
N.
(1987)Pollution-inducedinjuryinNorthAmericanforests:factsandsuspicions.
TreePhysiology3,1–15.
WoodruffD.
R.
,BondB.
J.
&MeinzerF.
C.
(2004)DoesturgorlimitgrowthintalltreesPlant,Cell&Environment27,229–236.
WoodruffD.
R.
,MccullohK.
A.
,WarrenJ.
M.
,MeinzerF.
C.
&LachenbruchB.
(2007)ImpactsoftreeheightonleafhydraulicarchitectureandstomatalcontrolinDouglas-r.
Plant,Cell&Environment30,559–569.
WoodruffD.
R.
,MeinzerF.
C.
&LachenbruchB.
(2008)Height-relatedtrendsinleafxylemanatomyandshoothydraulicchar-acteristicsinatallconifer:safetyversusefciencyinwatertransport.
NewPhytologist180,90–99.
WoodruffD.
R.
,MeinzerF.
C.
,LachenbruchB.
&JohnsonD.
M.
(2009)Coordinationofleafstructureandgasexchangealongaheightgradientinatallconifer.
TreePhysiology29,261–272.
ZimmermannM.
H.
&JejeA.
A.
(1981)Vessel-lengthdistributionofsomeAmericanwoodyplants.
CanadianJournalofBotany59,1882–1892.
ZwienieckiM.
A.
&HolbrookN.
M.
(1998)Diurnalvariationinxylemhydraulicconductivityinwhiteash(FraxinusamericanaL.
),redmaple(AcerrubrumL.
)andredspruce(PicearubensSarg.
).
Plant,Cell&Environment21,1173–1180.
Received4March2009;receivedinrevisedform20May2009;acceptedforpublication20May20091466Y.
-J.
Zhangetal.
2009BlackwellPublishingLtd,Plant,CellandEnvironment,32,1456–1466