Acta Hortic. 1222. ISHS 2018. DOI 10.17660/ActaHortic.2018.1222.2 Proc. of the X International Workshop on Sap Flow Eds.: L.S. Santiago and H.J. Schenk

5

Changes in xylem conducting capacity and water storage across species: how can variable air content of xylem cells affect sap flow? A.J.McElrone1,2,a,J.M.Earles3,T.M.Knipfer2,C.P.Albuquerque2,C.R.Brodersen3andI.F.Cuneo41USDA-ARS,CropsPathologyandGeneticsResearchUnit,Davis,California,USA;2DepartmentofViticultureandEnology, University of California, Davis, California, USA; 3School of Forestry & Environmental Studies, YaleUniversity,NewHaven,Connecticut,USA;4EscueladeAgronomıa,PontificiaUniversidadCatolicadeValparaıso,Quillota,Chile.Abstract

Sap flowsensorsandother techniquesarecommonlyusedacrossspeciesandplant organs to quantify water use and storage, detect stress, and evaluate thecontributionofvarioustissuestoplant/organwaterbalance.Sapflowmethodsoftenrelyuponmodellingorassumptionsabouthowheatdeliveredbythesapflowsensorsis partitioned into convection and conduction into active sapwood xylem andsurrounding tissues.Dynamic changes in tissuewater contentover space and timecan impact the interpretation of plant and organ water use and how variouscompartments contribute to an integrated response to plant stress. Here,we firstsummarizeresultsfromavarietyofstudiesthatusedacombinationofsynchrotron-basedX-raymicroCTandMRI imagingtodemonstratehowwatercontentofvariousorgansandxylemcelltypescanchangetemporallyandhowthespatialdistributionofair-filled tissuemay impactpatternsof sap flowwithin thexylemnetwork.Resultsfromvisualizationtechniqueswerecomparedtothat fromtraditionalhydraulicandsap flowmethods to illustratepotentialdiscrepanciesparticularlywhen comparingdata from excised stems versus intact plants. Using a spatially explicitmodel,wedemonstrate how changes in the water content of various cell types can impactresulting interpretationof sensoroutput. Implications for the interpretationof sapflowandothersensordatabasedontheseresultsisdiscussed.

Keywords: thermaldissipationprobes,thermalconductivity,tissuemoisturecontent,heatpulse sensors, capacitance, water storage, xylem embolism, finite elementmodellingINTRODUCTIONNon-invasivevisualizationtechniques,suchasmagneticresonanceimaging(MRI)andx-raymicro-computedtomography(microCT),arebeingusedmorefrequentlytostudyplantwatertransportinvivo(e.g.,Holbrooketal.,2001;Kaufmannetal.,2009;Brodersenetal.,2010;Knipferetal.,2015,2016;Choatet al.,2015).These techniqueshaveprovidednewinsights into thestructureand functionofplanthydraulicnetworks that could impactourunderstanding and interpretation of datasets captured with sap flowmethods and otherplanthydraulicandwaterrelationstechniques.Leeetal.(2013)recentlymodelledsapflowinthree-dimensionalxylemnetworksderivedfrommicroCT,andfoundthatflowin27%ofthevesselswasorientedinthedirectionoppositethebulkflowinthexylemnetworkundernormal transpirational conditions. This example highlights that efforts should be mademoving forward to integrate findings across scales to improve sap flowmethods and ourunderstandingofdatageneratedfromthesetechniques.Non-invasive techniques have also raised doubts about some commonly heldparadigms in plant hydraulics. For example,while xylem embolismplays a critical role in aE-mail: ajmcelrone@ucdavis.edu

6

plantmortalityunderconditionsofseverewaterstress (e.g.,Choat,2013),directmicroCTvisualizationsacrossnumerousspeciesandorgansshowthatminimalcavitationoccurs inxylemconduitsundermildtomoderatewaterstress(i.e.,Scoffonietal.,2017a,b;Cuneoetal.,2016;Hochbergetal.,2016;Choatetal.,2010;Brodersenetal.,2013).These findingssuggestthatdailycyclesofembolismformationandrepairarenotcommonplace(CochardandDelzon,2013).Furthermore,Knipferetal.(2017)foundthatrefillingofbothcapillarywaterstorage(i.e., fibers)andvesselsoccursrarelyorneverunderin-vivoconditionsovertwotimescales,24hand3weeks,forLaurusnobilissaplings,butpatternsofovernightsapflowintreeshavebeenattributedtodiurnalcyclesofstoragetissuedischargeandrecharge(Goldstein et al., 1998). Knipfer et al. (2017) also found that fibers refill when excisedsamples are attached to a water source, which impact sap flow sensor calibrationsconductedonexcisedplantsegments.Visualization techniques andmodelling provide an opportunity to evaluatewhetherdynamicchangesinxylemtissuewatercontent(i.e.,fiberandvessels)occuronadailybasisacrossplantspecies,whichcanimprovesapflowmethodsanddatainterpretationforintactplants. This could also help to resolvewhether artefactual refilling and internal hydraulicredistribution that occurs in excised samples (e.g., Knipfer et al., 2016) impacts sap flowsensor calibrations and whether dye infiltrations, which are commonly used for sensorcalibration and translation of flowvelocities to valuesof volumetricwateruse, accuratelyrepresentsapflowforintactplants.Thismanuscriptrepresentsathoughtexercisebasedonwhatwehavelearnedaboutxylem functioning from advanced visualization techniques in recent years, and how itpotentially impacts the interpretation of our sap flow and other sensors data. We showvariedexamplesofxylemtissuewatercontentinsinglestemsandacrossspecies,andthenuse finiteelementmodelling todemonstrateeffectsonsap flowsensoroutput,whichwasshown to vary dramatically depending on the hydration status of the tissues surroundingsapflowsensorheatingelements.MATERIALANDMETHODSPlantmaterials used in the experimentswere grown in a greenhouse facility at theUniversity of California, Davis. Laurus nobilis and Acer rubrum saplings (~60 cm height)wereobtainedfromcommercialnurseries.Vitisvinifera‘CabernetSauvignon’wereobtainedfromtwosources:1)own-rootedcuttingspropagatedinourgreenhouse;and2)UberVines(~80cmheight)obtainedfromDuarteNursery,Hughson,CA.ThebenchgraftedUberVine(magnum) utilizes an extra-long rootstock (110R) cane, so it can be tied directly to thetrainingwirewhenplanted.Thecaneiscutintheairduringthegraftingprocessandthusembolizes any existing xylem vessels in these vines. Thus the majority of xylem in ourscannedcross-sectionswereair-filledat the timeofourmicroCTscanning (Figure1),butnewxylemhadstartedtogrowontheouteredgeofthestem.Plants were grown in either 4- or 6-inch diameter plastic pots filled with soil mix(equalpartsofpeatmoss,compostedbark,sandandperlite)andmaintainedfor~6weeksunder typical summer greenhouse conditions (approximated day/night temperature of25/8°C,photoperiodof15/9h,relativehumidityof35%).Plantswereirrigateddailywithwatersupplementedwithmacro-andmicro-nutrients(similartoKnipferetal.,2015).PlantmaterialwasscannedattheX-raymicro-tomographyfacility(Beamline8.3.2)atthe Lawrence Berkley National Laboratory. Saplings were transported by car to theAdvancedLight Source (ALS, LawrenceBerkeleyNationalLab,Berkeley, CA) less than4hprior to analysis. For visualization of stem tissue, the potted sapling was placed in analuminumcageandwasscannedoverasmallaxialdistanceonthestems.Forsomevinesthestemportionscannedontheintactplantswereexcised,driedat40°Candrescanned~24hlater(Figure1).Stemswerescannedina21keVsynchrotronX-raybeamusingacontinuoustomographysettingyielding1025two-dimensional longitudinal images(resolutionof3.22µmpixel-1) thatwerecapturedonaCMOScamera(PCO.edge;PCO,Kehlheim,Germany)at350 ms exposure time. Acquired raw images were reconstructed into transverse imagesusingacustomsoftwarepluginforFijiimage-processingsoftware(www.fiji.sc,ImageJ)that

7

usedOctopus software (ver. 8.3;National Institutes forNuclear Science, GhentUniversity,Ghent, Belgium) in the background (Knipfer et al., 2016). Longitudinal images weregenerated using the slice tool in the software AVIZO (ver. 6.2; Visualization SciencesGroup/FEI,Hillsboro,Oregon,USA).

Figure1. Comparison of intact, excised, and dried states for a grapevine stem capturedusingmicroCTatthesameaxiallocationonthesameplant.Theblueareaforthefirst image highlights the fluid filled vessels in the intact state. After beingscanned, dye infiltration was performed on an actively transpiring shoot atmiddaywhilethebaseofthestemwassubmergedinthedyesolution.Imageontherightshowsthestemfullyembolizedafterbeingdried.Dyeinfiltrationswereconductedongrapevinestoassesstheconductivesapwoodareaafter microCT scanning of the intact plants using a method similar to that described inPearsall et al. (2014). The vines were maintained outside the ALS facility placed in thesunlightpriortotheintactscanandduringthedyeinfiltration.Vinesstemsweresubmergedinacrystalvioletdyesolution,thenthesegmentwascutwithsharppruningshears.Thecutendwasthenquicklyrecutwithafreshrazorbladetolimitvesselclogging,andthebottomof the stem was held submerged under the dye until it appeared in the leaf material(between30-60min).Transpirationrates forthesevineswereconfirmed justpriortodyeinfiltrationbymeasuringweightlossfromtheplant,wherethesoilandthepotweresealedindoubleplasticbags toprevent evaporation.Thesedatawere used to calculate sap flowvelocityintheremainingvesselsfromtheintactplantscanvs.thedriedsample.RESULTSANDDISCUSSIONThedye infiltration experimentwas performedon transpiring grapevines thatwerecutwhilethestemwassubmergedindyesolution.Thiswaseffectiveinidentifyingtheactivesapwood area, this is a standardmethod for assessing active sapwood area for sap flowtechniques(e.g.,Pearsalletal.,2014).MicroCTscansofagrapevineshootbeforeandafterthe excision showed that similarnumbersof vesselswere fluid filledbefore and after theinfiltration(Figure1,leftandmiddleimages),andtheactivexylemareacomposedofwater-filledvesselswaslimitedtotheveryouteredge(i.e.,theyoungestxylem).Ascanofthedriedsample (Figure 1, right) revealed that all the fluid-filled vessels in the dyed area emptiedupon drying suggesting these were filled with water in the intact plant. The grapevines(UberVines) were actively transpiring at the time of the dye infiltration experiment. Onaverage,thesevinesweretranspiringatarateof4.34gH2Oh-1duringtheperiodjustpriortodyeinfiltration.Incomparisontoalternativedyeinfiltrationmethods,thismethodofdyeinfiltrationispreferredtodyeinfiltrationonanexcisedstemsegmentwithtwoopenends(i.e.,cutatthebaseofthestemandjustbelowtheleaves,resultinginasinglesegment).If

8

theexcisedsegmentisshorterthanthelongestvessels,thiscanresultinnumerousvesselsopenedatbothendsofthesegmentthatwouldcreatepreferentialflowpathsanddominatedyeinfiltrationandconductivitymeasurements,whichmaynotrepresentthestatusoftheintactxylemnetwork.HigherresolutionmicroCTscansbetter illustrate the thinactivebandofwater filledvessels in the imaged grapevine plant (Figure 2) and correspond well with the dyeinfiltration in theoutermostportionof the xylemnearest the cambium. Interestingly, theolderxylemwasalwaysembolized(i.e.,didnotrefill)despitetheseplantsbeingmaintainedinawell-wateredstateinthegreenhousesduring6weekspriortotheexperiment,enoughtime for embolism refilling in grapevines (Knipfer et al., 2015). Despite the fact thatgrapevines can refill embolized conduits (Holbrook et al., 2001; Brodersen et al., 2010;Knipfer et al., 2015), these older vessels remained air-filled, and recovery of xylemconductivitywasassociatedwithgrowthofnewvesselsuponreconnectionofthevasculartissueinthegraft.Theseimagesalsoillustratethedifficultiesofplacinginsertionsapflowprobesintotheactivesapwoodforplantsliketheseandofaccuratelymeasuringtheactivesap wood area. Flow rates under such conditions would likely involve very few vessels,especiallywhenaccountingforwoundingeffectsassociatedwithneedleinsertion,andcouldsuffer fromheatbeingdissipateddifferentially inparts of the cross sectionwithdifferentvesselembolismstatus.

Figure2. Exampleofthethinactivecross-sectionalareaofanUberVinegraftedgrapevineexamined with microCT (left panel) and excised dyed stem (right panel). Theactive (water-filled) vessels in themicroCT image are in the region outlined inblue. The very shallow active cross-sectional area in these two-year-old plantsresults from all of the previous year’s xylem becoming inactivewhen the scionwas grafted onto theUberVine rootstock. This plantwas actively transpiring atthetimeofdyeing,andtheolderxylemremainedinanemptystate(i.e.,theplantis relying on a very thin band of active xylem closest to the cambium fortransportingwatertotranspiringfoliage).MicroCTscanscollectedacrossseveralspeciesrevealedthatfiberlumencontent(i.e.,watervs.air)candifferacrossspeciesandinthesameplant(Figures3and4).ThemicroCTimageofVitisviniferashowedembolizedvessels,whilethesurroundingxylemtissue(fibersandparenchyma)remainedwater-filled (Figure3, left);grapevine fibers, likeother lianas,

9

are living atmaturity and contain a viable cellmembrane (Knipfer et al., 2016), thus canremainwater-filledevenunderseverestress.Theonly timewehaveseen fibersempty ingrapevinesiswhenplantsareinfectedwithfungalpathogensthattargetthelivingfibersandparenchyma cells (Czemmel et al., 2015;McElrone, pers. observ.). It should benoted thattracheids (imperforate tracheary elements) are also common in many liana species,including Vitis vinifera, and are characterized by distinctly bordered pits with intact pitmembranesthataresimilartovessel-vesselpitmembranes.Ontheotherhand,fibersdonotcontribute toaxial transportanddonotshowdistinctlyborderedpits.Fiberscanbe filledwithairorcapillarywaterorremainlivingforalongtime,whichwouldalsogivethesetheimpressionofbeingwater-filledinmicroCTscans.Oftentracheidsandfibersoccurtogether,whichmakes thecelldiversity rathercomplex,andmacerationscombinedwith tangentialandradialsectionsareusefultoobserveimperforatetrachearyelementsindetail.

Figure3.TransversemicroCTimagesforthreespecies(grapevine–leftpanel;redmaple–middlepanel;andbaylaurel–rightpanel)underwaterstressdemonstratingthevariabilityintissuewatercontentthatcanexist.FibersinV.viniferaarelivingatmaturity and remain water-filled unless this tissue is infected with a fungalpathogen.ThefibersinAcerrubrumforthissampleemptiedundermilddroughtstress and long before embolism formation. Laurus nobilis exhibits patternswherethevesselsandfiberscanemptyatsimilarstressleaves.Thefibersinthelattertwospeciesaredeadatmaturity.Incontrasttograpevines,themicroCTimageofAcerrubrumexhibitedemptyfibersinthematurexylemwhiletherayparenchymawashydrated,andallxylemvesselsfilledwithwater(Figure3,middle),andLaurusnobilisshowedmostlywater-filledvesselsandfibers,butsomepatchesofair-filledfibers(Figure3,right).AscanofanadditionalL.nobilisplantrevealed this pattern could be quite variable evenwithin the sameplant (Figure 4). Fourpossiblecombinationswithxylemhydrationstatuswereseenintheseplants:1)water-filledvesselsandfiberslocatedclosethecambium(blueboxinFigure4,right,FV-FF);2)water-filledvesselsandempty(i.e.,air-filled)fibers(greenboxinFigure4,right,FV-EF);3)emptyvesselsandwater-filledfibers(greyboxinFigure4,right,EV-FF);and4)emptyvesselsandemptyfibers(blackboxinFigure4,right,EV-EF).Suchpatternswithinasingleplantcouldcause significant problemswith sap flowmethods given that heat dissipationwould varydramatically along the length of an insertionprobe inserted into this tissue. For example,heatconductivepropertiesofairandwatervaryconsiderably,andflowrateswherevesselsare functional, but fibers are either water or air-filled would introduce errors into theapparentversusactualflowratesateachpoint.To test this idea,we used a simplified finite elementmodelwhere filled andwaterconducting vessels were surrounded by either water- or air-filled fibers (Figure 5).Temperature trace outputs from this simulation were fed into the heat ratio method(Burgess et al., 2001) algorithms and revealed that apparent velocities were significantlyhigherwhenfiberswereair-filledcomparedtothewater-filledstate(Figure6).Underboth

10

low and high flow conditions (i.e., 5 and 50 cm h-1, respectively), the apparent flow rateswere dramatically higher when conducting vessels were surrounded by air-filled fibers(Figure6).

Figure4. Example of an intact plant stem (Laurus nobilis) exhibiting xylem tissue withvaried states of lumen filling. Varied states include water-filled vessels (FV),water-filled fibers (FF), empty vessels (EV), empty fibers (EF). Image wascapturedwithsynchrotron-basedx-raymicroCTatALSonanintactpottedplant.This figure illustrates that sensor placement, sensor output, and active cross-sectional areaof sucha stemwouldcreatedifficulties inunderstanding the sapflowsensoroutput.

Figure5. Illustration of the finite element model (left image) used to simulate heatdissipation away from a heater element and into the surrounding tissue viaconductanceandconvection.Resultingtemperaturetracesforthetwofiberstatesunderlowflowconditionsarepresented(rightpanels)inaheatratiomethodsapflow configuration. Vh is the resultant heat pulse velocity for this theoreticalsensor.

11

Figure6. Apparent velocities versus the imposed actual velocities used to drive the FEMsimulationcomparedtothe1:1line(solidblackline).CONCLUSIONSItiscurrentlyunknownhowpatternsoftissuehydrationvarywithinplantsandacrossspecies, and it is impossible to currently test this for installations under field conditions.Knipferet al. (2017) recently showed thatxylem fibers rarely refilled in intactplants,butfibertissuerefilledrapidlywhenthetissuewasexcisedandattachedtoawatersource.Suchrehydrationmeasurements are commonlyused to inferdetails about tissuewater storagecapacity.Knipfer et al. (2016)also showed thatwater canbe internally redistributed (i.e.,movingoutofvesselsandintosurroundingxylemcells)inexcisedgrapevinestemsattachedtoawatersource.Giventhatcalibrationsofsapflowandothersensorsareoftenconductedonexcisedsamplesinthelab,itisimportanttotestwhetherxylemtissuehydrationstatusischanging under these artificial conditions and to determine how this impacts ourcalibrations.Futuresapflowtechniquedevelopmentswouldlikelybenefitfrompairingwithxylemnetworkpropertiesderivedfromnon-invasivetechniqueslikemicroCT.Samplesfromintact field plants may be able to be excised from flash frozen cross sections (thusmaintainingtheirnaturalstatus)toassesshydrationvariabilitybetweenxylemcelltypesindifferentpartsofthecross-section.ACKNOWLEDGEMENTSAJM would like to thank the symposium organizers and ISHS for the invitation,opportunity,andfinancialsupporttopresentakeynotetalk.WorkpresentedinthistalkwasfundedbytheUSDA-ARSSustainableViticultureCRIS(grantno.5306-21220-004-00).TheAdvancedLightSourceissupportedbytheDirector,OfficeofScience,OfficeofBasicEnergyScience, of the US Department of Energy under contract no. DE-AC02-05CH11231, TheCalifornia Grapevine Rootstock Improvement Commission, and The American VineyardFoundation.CPAthanksCAPES/BrazilandtheUCDavisDepartmentofViticulture&Enologyforfinancialsupport.LiteraturecitedBrodersen,C.R.,McElrone,A.J., Choat,B.,Matthews,M.A., andShackel,K.A. (2010).Thedynamicsof embolismrepairinxylem:invivovisualizationsusinghigh-resolutioncomputedtomography.PlantPhysiol.154(3),1088–1095https://doi.org/10.1104/pp.110.162396.PubMedBrodersen,C.R.,McElrone,A.J.,Choat,B.,Lee,E.F.,Shackel,K.A.,andMatthews,M.A.(2013).Invivovisualizationsof drought-induced embolism spread in Vitis vinifera. Plant Physiol. 161 (4), 1820–1829 https://doi.org/10.1104/pp.112.212712.PubMed

12

Burgess, S.S.O., Adams, M.A., Turner, N.C., Beverly, C.R., Ong, C.K., Khan, A.A.H., and Bleby, T.M. (2001). Animprovedheatpulsemethodtomeasurelowandreverseratesofsapflowinwoodyplants.TreePhysiol.21(9),589–598https://doi.org/10.1093/treephys/21.9.589.PubMedChoat,B.(2013).Predictingthresholdsofdrought-inducedmortalityinwoodyplantspecies.TreePhysiol.33(7),669–671https://doi.org/10.1093/treephys/tpt046.PubMedChoat, B., Drayton, W.M., Brodersen, C., Matthews, M.A., Shackel, K.A., Wada, H., and McElrone, A.J. (2010).Measurementofvulnerabilitytowaterstress-inducedcavitationingrapevine:acomparisonoffourtechniquesappliedtoalong-vesseledspecies.PlantCellEnviron.33(9),1502–1512.PubMedChoat,B.,Brodersen,C.R.,andMcElrone,A.J.(2015).SynchrotronX-raymicrotomographyofxylemembolisminSequoia sempervirens saplings during cycles of drought and recovery. New Phytol. 205 (3), 1095–1105https://doi.org/10.1111/nph.13110.PubMedCochard,H.,andDelzon,S.(2013).Hydraulicfailureandrepairarenotroutineintrees.Ann.For.Sci.70(7),659–661https://doi.org/10.1007/s13595-013-0317-5.Cuneo, I.F., Knipfer, T., Brodersen, C.R., andMcElrone, A.J. (2016).Mechanical failure of fine root cortical cellsinitiates plant hydraulic decline during drought. Plant Physiol. 172 (3), 1669–1678 https://doi.org/10.1104/pp.16.00923.PubMedCzemmel, S., Galarneau, E.R., Travadon, R., McElrone, A.J., Cramer, G.R., and Baumgartner, K. (2015). GenesexpressedingrapevineleavesreveallatentwoodinfectionbythefungalpathogenNeofusicoccumparvum.PLoSONE10(3),e0121828https://doi.org/10.1371/journal.pone.0121828.PubMedGoldstein,G.,Andrade,J.L.,Meinzer,F.C.,Holbrook,N.M.,Cavelier,J.,Jackson,P.,andCelis,A.(1998).Stemwaterstorage and diurnal patterns ofwater use in tropical forest canopy trees. Plant Cell Environ.21 (4), 397–406https://doi.org/10.1046/j.1365-3040.1998.00273.x.Hochberg,U.,Albuquerque,C.,Rachmilevitch,S.,Cochard,H.,David-Schwartz,R.,Brodersen,C.R.,McElrone,A.,and Windt, C.W. (2016). Grapevine petioles are more sensitive to drought induced embolism than stems:evidencefrominvivoMRIandmicroCTobservationsofhydraulicvulnerabilitysegmentation.PlantCellEnviron.39,1886–1894https://doi.org/10.1111/pce.12688.PubMedHolbrook, N.M., Ahrens, E.T., Burns, M.J., and Zwieniecki, M.A. (2001). In vivo observation of cavitation andembolism repair using magnetic resonance imaging. Plant Physiol. 126 (1), 27–31https://doi.org/10.1104/pp.126.1.27.PubMedKaufmann, I., Schulze-Till, T., Schneider, H.U., Zimmermann, U., Jakob, P., andWegner, L.H. (2009). Functionalrepairofembolizedvesselsinmaizerootsaftertemporaldroughtstress,asdemonstratedbymagneticresonanceimaging.NewPhytol.184(1),245–256https://doi.org/10.1111/j.1469-8137.2009.02919.x.PubMedKnipfer,T.,Eustis,A.,Brodersen,C.,Walker,A.M.,andMcElrone,A.J.(2015).Grapevinespeciesfromvariednativehabitatsexhibitdifferencesinembolismformation/repairassociatedwithleafgasexchangeandrootpressure.PlantCellEnviron.38(8),1503–1513https://doi.org/10.1111/pce.12497.PubMedKnipfer,T.,Cuneo,I.F.,Brodersen,C.R.,andMcElrone,A.J.(2016).In-situvisualizationofthedynamicsinxylemembolism formation and removal in the absence of root pressure: a study on excised grapevine stems. PlantPhysiol.171(2),1024–1036.PubMedKnipfer,T.,Cuneo,I.F.,Earles,J.M.,Reyes,C.,Brodersen,C.R.,andMcElrone,A.J.(2017).Storagecompartmentsforcapillarywater refill in excised stems but rarely refill in intactwoody plants (Laurusnobilis): visualization oftissue-specific dynamics using microCT imaging. Plant Physiol. 175, 1649–1660 https://doi.org/10.1104/pp.17.01133.PubMedLee,E.F.,Matthews,M.A.,McElrone,A.J.,Phillips,R.J.,Shackel,K.A.,andBrodersen,C.R.(2013).AnalysisofHRCT-derived xylem network reveals reverse flow in some vessels. J. Theor. Biol. 333, 146–155https://doi.org/10.1016/j.jtbi.2013.05.021.PubMedPearsall, K.,Williams, L.E., Castorani, S.E., Bleby, T.M., andMcElrone, A.J. (2014). Evaluating the potential of anoveldualheatpulsesensortomeasurevolumetricwateruse ingrapevinesunderarangeof flowconditions.Funct.PlantBiol.41(8),874–883https://doi.org/10.1071/FP13156.Scoffoni,C.,Albuquerque,C.,Brodersen,C.R.,Townes,S.V., John,G.P.,Bartlett,M.K.,Buckley,T.N.,McElrone,A.J.,and Sack, L. (2017a). Outside-xylem vulnerability, not xylem embolism, controls leaf hydraulic decline duringdehydration.PlantPhysiol.173(2),1197–1210https://doi.org/10.1104/pp.16.01643.PubMedScoffoni,C.,Albuquerque,C.,Brodersen,C.R.,Townes,S.V.,John,G.P.,Cochard,H.,Buckley,T.N.,McElrone,A.J.,andSack,L.(2017b).Leafveinxylemconduitdiameterinfluencessusceptibilitytoembolismandhydraulicdecline.NewPhytol.213(3),1076–1092https://doi.org/10.1111/nph.14256.PubMed