ARTICLE IN PRESS

Journal of Plant Physiology 163 (2006) 996—1007

0176-1617/$ - sdoi:10.1016/j.

AbbreviationE, total energyacoustic emiss�CorrespondE-mail addr

www.elsevier.de/jplph

Acoustic emission analysis and experiments withphysical model systems reveal a peculiar nature ofthe xylem tension

Ralf Laschimkea,�, Maria Burgera, Hartmut Vallenb

aZollern GmbH, Hitzkofer Str. 1, D-72488 Sigmaringen, GermanybVallen Systeme GmbH, Schaftlarner Weg 26, D-82075 Icking, Germany

Received 27 February 2006; accepted 10 May 2006

KEYWORDSAcoustic emissions;Cohesion theory;Gas bubbles;Peristaltic transport;Xylem transport

ee front matter & 2006jplph.2006.05.004

s: A, surface area; AE,; F, gravitational force;ions; e modulus of elasting author. Schlosshaldeess: Laschimke@t-onlin

SummaryAdvanced acoustic emission analysis, special microscopic examinations and experi-ments with physical model systems give reasons for the assumption that the tensionin the water conducting system of vascular plants is caused by countless minute gasbubbles strongly adhering to the hydrophobic lignin domains of the xylem vesselwalls. We ascertained these bubbles for several species of temperate deciduoustrees and conifers. It is our hypothesis that the coherent bubble system of the xylemconduits operates as a force-transmitting medium that is capable of transportingwater in traveling peristaltic waves. By virtue of the high elasticity of the gasbubbles, the hydro-pneumatic bubble system is capable of cyclic storing andreleasing of energy. We consider the abrupt regrouping of the wall adherent bubblesystem to be the origin of acoustic emissions from plants. For Ulmus glabra, werecorded violent acoustic activity during both transpiration and re-hydration. Thefrequency spectrum and the waveforms of the detected acoustic emissionscontradict traditional assumptions according to which acoustic emissions are causedby cavitation disruption of the stressed water column. We consider negative pressurein terms of the cohesion theory to be mimicked by the tension of the wall adherentbubble system.& 2006 Elsevier GmbH. All rights reserved.

Elsevier GmbH. All rights reserved.

acoustic emissions; AAE, audible acoustic emissions; CLSM, confocal laser scanning microscopy;n, moles; p, pressure; s, translation; S, entropy; V, volume; T, temperature; UAE, ultrasonicicity; m, chemical potential; s, surface tension; t, three-dimensional distortion.1, D-72479 Strassberg, Germany. Tel.: +49 7434 8288; fax: +49 7434 315 908.

e.de (R. Laschimke).

ARTICLE IN PRESS

Nature of xylem tension 997

Introduction

The predominantly passive character of long-distance water transport has frequently beenquestioned. However, it has been repeatedlydemonstrated, even in context of acoustic emis-sions (Borghetti et al., 1991), that the applicationof poisons as metabolic inhibitors to the watertransport system has little or no effect on watermovement, reinforcing the concept that non-livingcells are primarily involved in water transport andno essential input of metabolic energy takes place.Despite this, various suggestions of active pumpinghave been proposed time and again, but nomechanical or biochemical devices that couldachieve such pumping have been found. Also, thefrequently cited root pressure would not becapable of transporting water to the tops of talltrees. In sum, the idea of the cohesion theory thatwater ascends in the apoplast part of the xylem,largely under the direct influence of the evapora-tion from the leaves, is most likely the correctexplanation, but the transmission of pulling forcesthrough the water column is still puzzling.

Milburn and Johnson (1966) reported that tran-spiring plants generate acoustic emissions in theaudio frequency range. Since that time, Milburn’sfindings have been considered proof of the validityof the cohesion theory, according to which theascent of sap is caused by a gradient of negativepressure. The cohesion theory has been repeatedlyquestioned from a physical point of view becausethe claimed metastability of water under negativepressure has not been demonstrated for airsaturated water. However, the soil water taken upby the roots is air saturated. Despite this incon-sistency, challengers of the cohesion theory areusually rejected with reference to the acousticemissions discovered by Milburn. Seen in the lightof the cohesion theory, the acoustic emissions arecaused by cavitation-disruptions of the watercolumn under negative pressure. This argumentwas weakened when Tyree and Dixon (1983)introduced the ultrasonic frequency range intothe investigations. It has now become clear thatthe AAE is only the tip of the iceberg. Theoverabundance of the UAE turned out to be aproblem. Presupposing that each acoustic event isattributable to a cavitation event, it appearsimpossible to place the countless cavitation defectswithin the limited number of water conductingconduits. The problem is compounded since thecavitation-defects are supposed to be instantlyembolized by air, and therefore the affected vesseltubes malfunction. Several assumptions of inces-sant embolism repair have been made recently by

Zwieniecki and Holbrook (2000). However, a de-monstrable repair mechanism has not been foundto date.

Simultaneous monitoring of AAE and UAE byRitman and Milburn (1990) revealed a significanttime pattern of the acoustic emissions with leadingUAE for intense transpiration and rising AAEfor declining transpiration. Unfortunately, theseauthors eliminated the acoustic emissions duringthe nightly re-hydration phase. Apparently, theAE-signals of the re-hydration phase could not beexplained within the context of cohesion theory.Milburn (1993) suggested that the mysterious UAEbe ignored, and considered the AAE to be the mostreliable indicator of cavitation. In a previousinvestigation, we argued that the acoustic emis-sions from plants are not caused by cavitationunder negative pressure, but by sudden surfacerearrangements of groups of wall adherent micro-bubbles under positive pressure (Laschimke, 1989).The existence of gas bubbles in the xylem conduitshas been known for years. It is generally known thatgas bubbles, which fill the whole lumen of thevessels, form so-called Jamin chains. The Jaminchains do not contribute to the driving forces ofwater transport; rather, they are barriers. Thediscovered wall adherent microbubbles, however,are much smaller in diameter than the vesselstubes. Such bubbles are largely stable and do notimmediately result in embolism. When we pre-sented the first micrographs in 1989, the micro-bubbles seemed to be an absurdity, seen in the lightof the governing cohesion theory. On the otherhand, Zimmermann and co-workers regarded thewall adherent microbubbles as an attractiveconcept of water transport by Marangoni-forces(Zimmermann et al., 1993). Marangoni flow re-quires a permanent concentration gradient in thetransported water. This gradient has been taken forgranted by Zimmermann, but no evidence couldbe provided up to now. Because Marangoni flowscarcely generates acoustic emissions, Zimmer-mann took into consideration several other acousticorigins, such as aspirating pits, occurrence offlow turbulence, fusion of gas bubbles adhering tothe walls, cracks in the cell walls under strain,drying out of the bark, shrinking of the stemdiameter and xylem bending (Zimmermann et al.,2004). Aiming to shed more light on the mysteriousacoustic emissions from plants, we applied ad-vanced methods of event detection and AE-analy-sis. By means of a special arrangement of theacoustic sensors, we were able to eliminatesome of the named dubious origins of AE signals.In order to substantiate the results of our acousticemission testing, we furthered our investigations

ARTICLE IN PRESS

R. Laschimke et al.998

with special microscopic examinations and experi-ments with physical model systems.

Materials and methods

Acoustic emissions

We selected a 4-year-old potted Ulmus glabra forexamination. The broad leaves of Ulmus glabrafacilitate the linkage of the special sensors we usedin our investigation. Usually, acoustic sensors arelinked to partly de-barked living stems or branches.We placed the sensors directly on the upper surfaceof living leaves as shown in Fig. 1b. This unusualarrangement allows monitoring of acoustic eventsright next to the place of transpiration. Thetranspiration of the investigated leaf was sufferedpractically no interference because the majority ofthe transpiring stomata are located on the un-touched underside of the leaf. To record transpira-tion data simultaneously with acoustic data, wefitted the potted young elm tree to a computerizedelectronic scale as shown in Fig. 1a. To avoiduncontrolled evaporation, we covered the flower-pot with a polyethylene foil. The young elm waswell watered before data acquisition. The tem-perature in the laboratory was constant at 22 1Cand a relative humidity of 60%. Thus, the transpira-tion depended primarily on the intensity of thesunlight and the rhythm of day and night. Ourexperiments were carried out from the middle ofJuly to the middle of August 1998.

We used the mass-loaded piezoelectric sensor SE-45 from Dunegan Engineering (USA), which exhibitsoptimum response in the range between 20 and120 kHz. Experience has shown that, in the fre-quently used MHz range, the density of the acousticemissions is extremely high. For the enormousnumber of the MHz signals, the storage of thewaveforms of the individual signals is practicallyimpossible, unlike the range of 20–120 kHz, wherethe signals are within reasonable limits. The

Figure 1. Simultaneous detection of acoustic emissions and trUlmus glabra with measuring sensor (left) and controlling se

threshold of signal detection was 26 dB, whichcorresponds to 20 mVat the sensor output. By meansof a controlling sensor that was not in contact withthe leaf surface, we discriminated and eliminatedbackground noise from the relevant acoustic data.To obtain maximum information, we employed theVallen system AMSY4 equipped with transientrecorders, which store the complete waveforms ofa maximum of 5000 signals.

Microscopic examination

Encouraged by the results of our acoustic emis-sion testing, we decided to attempt to confirm ourprevious findings of microbubbles adhering at thelignin domains of the xylem vessel walls. Lignin canbe detected by its yellow-green auto-fluorescencewhen excited by blue light. In a normal epi-fluorescence microscope, the intensity of the ligninfluorescence is low and in general, the source ofthe fluorescence cannot be localized sharply due tothe stray light and the low depth of focus. Thisdisadvantage is overcome by confocal laser scan-ning microscopy (Laschimke, 1991). The finerpoints of the surface structure of the vessel wallswere made visible by means of SEM. The non-uniform wetting behavior of the xylem vessel wallswere made visible by means of water immersionmicroscopy; a living branch of our young Ulmusglabra was bent down into a container of tap waterwith the temperature of the branch. Then a samplewith a length of about 3 cm was cut off, and atangential cut was made. Then the sample thatremained immersed in water was transferred underthe microscope equipped with high-resolutionlenses for water immersion (numerical aperture1.25). This procedure lasted only a few minutes.

Physical model systems

The special properties of wall adherent bubblesare evident through experiments with modelsystems. Although our experimental devices do

anspiration loss: (a) experimental set-up and (b) leaves ofnsor (right).

ARTICLE IN PRESS

Figure 2. Experiments with a glass capillary perforated by a laser beam: (a) 40 symmetrically arranged lateral holes;(b) 13 lateral holes in one row, sheathed with hydrophobic polyolefine. Formation of gas bubbles at the hydrophobicbottoms of the lateral blind holes after 2 h; and (c) the gas bubbles according to (b) after 5 h. Embolism due toformation of Jamin chains.

Figure 3. Capillary with 40 hydrophobic lateral blindholes and gas accumulations after 4 h. By virtue of thebubbles a longer continuous water column is held againstgravity, compared to a normal capillary (right).

Nature of xylem tension 999

not accurately simulate the structural features ofxylem vessels, these devices demonstrate impress-ively the physical behavior of wall adherent gasbubbles. Fig. 2a shows a glass capillary perforatedby means of a laser beam. We sheathed theperforated section with transparent hydrophobicpolyolefin and immersed the prepared capillarytube into air-saturated water. Shortly after immer-sion, air bubbles develop at the hydrophobicbottoms of the lateral blind holes. We then with-drew the capillary tube half out of the watercontainer and compared the water column withthat of a normal capillary tube. Fig. 3 shows that,by virtue of the lateral bubbles, the water columnis held on a significantly higher level comparedto a normal capillary tube. After some hours, thecapillary tube became embolized by expandingbubbles and the formation of Jamin chains (Fig. 2band c). It became obvious that lasting stability ofthe water column requires a much larger number ofmuch smaller hydrophobic blind holes, similar tothe xylem vessels walls, where the geometricalparameters have been optimized by nature. Con-sequently, we made efforts to miniaturize theperforations of the tube wall. This turned out tobe unrealistic, but we were successful in preparingimproved experimental devices with plane sur-faces. Fig. 4 shows an array of densely arrangedhydrophobic blind holes, each with a diameter of0.05mm. Related to the vessel pits of Ulmusglabra, even the miniaturized blind holes are stillin the ratio of 50:1. The disadvantage of ourminiaturized devices is that the peculiar shape ofthe wall adherent bubbles is not clearly visible

ARTICLE IN PRESS

Figure 4. Array of numerous hydrophobic blind holes at a hydrophilic plane surface: (a) flat-shaped bubbles withundulated rims indicate mechanical tension between the solid base and the bubbles; and (b) two bubbles (marked area)at the moment of coalescence. The blurred contours indicate violent vibration.

Figure 5. Formation of gas bubbles at four hydrophobic blind holes on a hydrophilic base: (a) initial stage; formation ofspherical bubbles and (b) coalescence of spherical bubbles, formation of a multi-feet non-spherical bubble.

R. Laschimke et al.1000

after their coalescence. Therefore, we made amodel illustrated in Fig. 5, which allows observa-tion of the modifications of these bubbles.

Results

Acoustic emissions

Figure 6a shows the transpiration loss for atesting period of 77 h. The regular course of thedaily transpiration loss and the nightly refillingindicates normal hydration of the elm. To improvethe assessment, we calculated the derivative curvecalled ‘‘transpiration intensity versus time’’, whichreveals minimum transpiration intensity at nightof about 15% of the maximum intensity duringdaylight (Fig. 6b). The diagram shows that acousticemissions take place incessantly during bothtranspiration, and re-hydration. To differentiatethe acoustic activity, we selected three types ofacoustic events by the dominant AE-signal fre-quency, which is the frequency where the spectrumexhibits the maximum magnitude. Fig. 6c shows thecorrelation between acoustic activity and tran-spiration intensity in the three selected frequency

ranges. We selected these special frequencyranges because the analysis of the AE signals inthese frequency ranges turned out to be moreconclusive than in other frequency ranges. Theemissions in the lowest frequency range risewhen the transpiration is rising. When the tran-spiration declines, the emissions in the lowestfrequency range stop, and the emissions in thenext higher frequency range start. The emissions inthe highest frequency range weaken during max-imum transpiration. The time pattern of theacoustic activity of the leaves of the investigatedUlmus glabra contradict the idea that acousticemissions from plants occur in conjunction withintense transpiration. It is obvious that the acousticactivity also remains undiminished during thenight refilling phase. The finer points of thedifferent frequency pattern during transpirationand re-filling cannot be explained with the currentlevel of data acquisition. Further investigations arenecessary.

To approach the mechanism of acoustic emissiongeneration, we analyzed the waveforms of the2200 events, which are compiled in diagram Fig. 6.The review of the waveforms revealed greatvariability of amplitude, duration and frequency.The majority of the waveforms conflict with

ARTICLE IN PRESS

Figure 6. Correlation between transpiration and acoustic activity of Ulmus glabra: (a) transpiration loss andtranspiration intensity versus time; (b) correlation between transpiration intensity and cumulative acoustic events; and(c) correlation between transpiration intensity and dominant frequencies.

Nature of xylem tension 1001

cavitation-disruptions of the water column. Aftercavitation-disruption, the end of the broken watercolumn should be violently retracted along theaffected vessel tube, accompanied by a fast fadingacoustic signal. We found this type of waveformonly by a very limited number of signals (Fig. 7a).For the most part, the waveforms indicate lastingoscillation of the acoustic origins, as shown in Fig.7b and c. This finding is not explicable in thecontext of cavitation-disruption.

Microscopic examinations

Figure 8 shows a SEM image of apoplast xylemvessel walls of Ulmus glabra. We noted fields of pitswith wide-open apertures and fields of pits with

narrow apertures and pit chambers. We assumedifferent function in water transport for thedifferent types of vessel pits, as explained in the‘‘Discussion’’ section, below. For closer investiga-tion, we prepared the pit chambers as in Fig. 9a.The SEM image and the corresponding CLSM imagereveal ring-shaped lignin concentrations at thebottom of the pit chambers.

Figure 10 shows water immersion micrographs ofthe vessel walls of Ulmus glabra and Quercus roburat the end of May. Gas bubbles are clinging to theapertures of the vessel pits. Under special illumina-tion conditions, the bubbles show total reflection,which is evidence of a liquid/gas interface. Prior tofoliation of the leaves, the pit chambers are onlypartly filled with gas as visible in Fig. 11a and b.After foliation, bubbles balloon at the pit apertures,

ARTICLE IN PRESS

Figure 7. Typical variations of the waveforms of acoustic emissions from Ulmus glabra: (a) waveform monitored atirregular intervals both, during transpiration and refilling. This waveform indicates abrupt disruption of the watercolumn. (b) Predominant waveforms during both transpiration and refilling, indicating vibration of the acoustic origin.(c) Frequently detected waveform, indicating overlapping of two vibrating acoustic origins.

Figure 8. Xylem vessel walls of Ulmus glabra. The SEMimage shows fields of wide-open pits and pits with narrowapertures and pit chambers.

R. Laschimke et al.1002

as illustrated in Fig. 11c. When the expandingbubbles come into mutual contact they coalesce inflat bubble aggregates as shown in Fig. 11d. Werealize that the observed bubbles are not exactlythe same shape as in the intact vessel tubes,because the opening of the vessel tubes instantlycauses a change of the hydraulic tension. However,spontaneous formation of new bubbles is incon-ceivable. We observed similar gas accumulationsalso at the xylem vessel walls of many other species.Figure 12 shows gas bubbles at the scalariform

structured vessel walls of Vitis vinifera and in thetracheids of Pinus silvestris.

Physical model systems

Figure 5 shows an arrangement of four sizablehydrophobic blind holes at a hydrophilic base.When immersed in air-saturated water, sphericalbubbles balloon at the blind holes. If the individualbubbles come into mutual contact, they coalesceinto non-spherical bubbles. The deformed non-spherical bubbles are coupled to the solid byseveral stretched feet like a polyp. Obviously, thebuoyancy alone cannot detach such multi-feet-bubbles, but a higher energy threshold must besurpassed. We consider the deformed multi-feet-bubble as being in a potential well. A potential wellis a region surrounding a local minimum of allinvolved potential energy forms as explained in the‘‘Discussion’’ (see below). Also, the gas bubbles atthe vessel pits we consider to be fixed in potentialwells. As visible in Fig. 4b, we were successful increating a snapshot of bubbles at the moment ofcoalescence. The blurred contours of the coales-cing bubbles are an indication of vibration. Byvirtue of the large vibration amplitude of thecomparatively sizeable bubbles, the vibration aftercoalescence was detectable by photograph. Someof the emerged bubble configurations cover up to30 blind holes. The undulated rims of the bubble

ARTICLE IN PRESS

Figure 9. Lignin concentrations at the bottom of the pit chambers of Ulmus glabra: (a) SEM image and (b) CLSM image;fluorescence mode; extended focus. The white line shows the lignin concentration along the horizontal line.

Figure 10. Gas bubbles at the apertures of the vesselpits of water conducting xylem vessel tubes in end ofMay. Water immersion micrographs: (a) Ulmus glabra and(b) Quercus robur.

Nature of xylem tension 1003

configurations (Fig. 4a) indicate considerable dis-tortion or tension between the bubbles and thesolid.

Discussion

The existence of wall adherent gas bubbles at thewalls of the operating xylem conduits cannot bereconciled with negative pressure in terms of thecohesion theory. We put forward a concept ofpassive water transport based on transpiration-induced pulling forces but exclusive of negativepressure.

The German forest botanist Hartig (1878) re-ported that 30 temperate angiosperm and gymnos-perm tree species show significant high watercontent in January/February. Similar results werelater reported by Lundegardh (1954), Gibbs (1958),and Glavac et al. (1990). The famous promoter ofthe cohesion theory Zimmermann (1983) concededthat the cohesion theory fails to explain the high

water content of deciduous trees in winter. It hasbeen speculated that the static hold of the watercolumn results from the perfect wettability of thexylem vessel walls. However, it has been ignoredthat a resting Newtonian liquid like water does notcreate resultant forces in a direction tangential tothe wetted wall. The same is true for osmoticforces. Only the forces of menisci might be capableof holding a continuous water column above thenormal capillary height. This problem was debatedby botanists 100 years ago, but extra menisci couldnot be found in the water transporting system. Wewant to emphasize that the discovered walladherent microbubbles are nothing but hiddenmenisci.

On comparison of Fig. 11a and c, the questionarises: what is the origin of the gas that causes thegradual expansion of the bubbles from April to Mayas visible in these figures? Most likely, the gas is CO2

that develops in the spring by dissimilation ofpolysaccharides. The dissolved CO2 diffuses to-gether with the mobilized saccharides from thesymplast part of the xylem into the water columnof the apoplast conduits. When the transpirationstarts, expansion of the wall adherent bubbles nextto the site of transpiration takes place. However,excessive expansion of any individual bubbleis avoided due to its cohesive interaction withneighboring bubbles. Instead of the local expansionof single bubbles, the expansion of the wholebubble layer propagates gradually in basal direc-tion. In parallel, the solid structure of the vesselwalls and the adjacent parenchyma cells becomeselastically distorted by interfacial forces accordingto our model system shown in Fig. 4a. Conse-quently, the water conducting system becomescharged with both surface energy and mechanicalenergy. In this phase, water is not taken up by theroots, but the water in the conduits becomesdisplaced in an apical direction. When the peri-stalsis-like expansion of the bubble system arrivesat the roots, the hydro-pneumatic system is able to

ARTICLE IN PRESS

Figure 11. Gas accumulations at the xylem vessel walls in different stages of expansion. Water immersion micrographs:(a) Ulmus glabra, middle of April; (b) Quercus robur, middle of April; (c) Ulmus glabra, middle of May; and (d) Ulmusglabra, end of June.

Figure 12. Gas accumulations at xylem vessel walls. Water immersion micrographs: (a) and (b) Vitis vinifera, end ofMay; and (c) and (d) Pinus silvestis, end of May.

R. Laschimke et al.1004

take up water by release of stored energy. Fromnow on, peristalsis-like drag flow takes place inapical direction. The drag flow develops because aNewtonian liquid is forced from the surface of abody, when the body is moved through the liquid.The propagating variation of the surface of thebubble system is equivalent to a moved body. Inhydrodynamics, the component of the force in thedirection of the moved body is called the drag. Thecomponent perpendicular to this direction is calledthe lift. In a narrow capillary tube, such as a xylemvessel tube, the lift superimposes in the centerof the tube. Thus, the water column will be movedin direction of the drag as outlined in Fig. 13.This type of flow could be described as inverseHagen–Poiseuille flow. Recent investigations byZholkevich (2001) ascertained, for Phaseolus vul-garis, peristaltic water transport in a 1-min rhythm.Zholkevich postulated a metabolic correspondingregulatory and signal system that controls anystill unknown mechanism of peristaltic watertransport. We suggest regulation by stomata-con-trolled transpiration according to Fig. 13. Thurmeret al. (1999) ascertained that height-varying ten-sion gradients developed in the vessels by varia-tions in light intensity for the liana Tetrastigmavoinierianum. To our knowledge, there are not any

comprehensive data available from the literaturefor peristaltic water movement in trees. However,periodical changes in stem diameters of transpiringtrees are well known.

It has long been recognized that plants can showrapid systemic response to localized stimuli, asreported by Malone and Alarcon (1995). ForTriticum, an increase in leaf thickness up to 100%was measured, caused by scorch wounding of aneighboring leaf. Malone suggested that the trans-mitted signals are hydraulic in nature, but notcompatible with other known models of signaling.The rate of propagation of the hydraulic signalsmeasured by Malone is tens of cm s�1. From ourpoint of view, the rapid swelling of the leaves canbe explained only by rapid expansion of gas spaces,such as wall adherent bubbles. It is possible thatthe expansion of the gas spaces propagates frombubble to bubble by means of a domino effect.

A minimal temperature difference betweenmoved air and leaves is enough for transpiration-induced transformation of heat into mechanicalwork of water transport. Every long-term energytransformation of heat into work requires a workingcycle. Amazingly, in the cohesion theory, there hasbeen claimed continuous passive water transportbut any underlying working cycle is missing. We

ARTICLE IN PRESS

Figure 14. Local working cycle, according to Fig. 13. Thegradual displacement of S along the time axis isunconsidered in this diagram. M minimum of free energyor potential well. S starting point of cyclic transformationof thermal energy into mechanical work. E embolismafter excessive accumulation of gas.

Figure 13. Transpiration-induced working cycle of xylemwater transport: (a) total working cycle dismantled tomany interlocked little working cycles and (b) phasesof peristalsis-like water transport. (A) Full hydration.(B) Beginning of transpiration; gradual expansion of thewall adherent bubble system; displacement of water tothe site of transpiration. (C) Intense transpiration,propagation of the expansion of the wall adherent bubblesystem; displacement of water to the site of transpira-tion; elastic distortion of the adjacent parenchyma;charge of mechanical energy. (D) Charge of energycomplete; transpiration stop; basal uptake of water;discharge of energy; beginning of drag flow (darkmarked) in apical direction. (E) Propagating dischargeof mechanical energy; re-hydration by propagating dragflow of water in apical direction. (F) Full hydration.

Nature of xylem tension 1005

suggest a working cycle according to Fig. 13a and b.We outlined this working cycle in a very simplifiedway, because, in our opinion, the formulation of athermodynamic state equation seems to be nearlyimpossible. The concerned energy forms, however,can be easily expressed by the following Gibbs-equation:

dE ¼ T dSþ F ds� pdV þ sdAþ �dtþ Smi dni

where dE is the total energy; T dS thermal energy;Fds gravitational energy; �p dV compression en-

ergy; s dA surface energy; e dt energy of the elasticdistortion; Smi dni chemical energy of the differentchemical components. The interaction of thenumerous energy forms corroborates the ‘Multi-Force’ Theory by Zimmermann et al. (2004).

The proposed working cycle starts after diffusionof sufficient gas into the permanently existing gasspaces of the pit chambers. Rising xylem tensioncorresponds with increasing size of the walladherent bubbles. The working cycle breaks downafter accumulation of a critical mass of gas, asoutlined in Fig. 14. Every working cycle can bedismantled to interlocked little working cycles. Indiagram Fig. 13a, the interlocked working cyclesare not in phase. These differences of phase areconsistent with our assumption that long distancewater transport resembles peristaltic movements.

The hydration status is usually assessed by meansof the water potential, which is the free energy ofthe water column associated with water per unitvolume. To be precise, the water potential is afunction of the solute potential, the pressurepotential and the potential due to gravity. Theseterms in Jm�3 are equivalent to pressure unitssuch as Pascals (Nm�2). In our concept, additionalenergy terms are implicated: the surface energy ofthe bubble system sdA and the energy of theelastic distortion of the solid structure e dt. Thesetwo terms are not associated with water per unitvolume. The surface energy is partly associatedwith water molecules, but not with water per unitvolume. The elastic energy is associated with themolecular structure of the solid. The term �pdV ismainly associated with the volume of gas contained

ARTICLE IN PRESS

R. Laschimke et al.1006

in the wall adherent bubbles. There is hardly anyfree energy associated with the liquid phase undereither positive or negative pressure. In considera-tion of this, we suggest redefining a ‘‘xylempotential’’, which includes the customary waterpotential, the gas potential of the bubble system,the surface potential of the bubble system and theelasticity potential of the solid structure.

The cohesive attraction between the individualbubbles enables the hydro-pneumatic system trans-porting surface energy in many small quantitiesfrom bubble to bubble. The gradually propagatingexpansion of the surface area of the whole bubblesystem represents, in fact, a flow of energy in thebasal direction. It is our statement that thistransport of non-metabolic energy from the leavesto the roots is the most important precondition oflong-distance water transport in plants (Laschimke,1990). In the last analysis, the water transportingsystem of plants can be viewed as an energyoscillator. To initiate the energy oscillation, thesystem must be deflected from the minimum offree energy or the potential well. This deflectionrequires an input of metabolic energy as outlined inFig. 14. After foliation and at the beginning oftranspiration, further input of metabolic energy isnot necessary, provided that the transpiration-induced input of thermal energy is enough tocompensate the friction loss of the water move-ment. It is clear that long-distance water transportrequires lateral passages. Such passages are pro-vided by the large number of pits with wide-openapertures (Fig. 8), which are free of gas accumula-tions and thus capable of transporting waterthrough the cell wall perforations.

The hydration state of plant material can beassessed by both direct and indirect measuringmethods. The indirect measuring pressure chambertest is based on the idea that negative pressureremains preserved in a detached sample, such as atwig. The preservation of negative pressure hasbeen concluded from the retraction of the menisciof the cut water column into the cut xylem conduits.There is no denying that this retraction results fromcontractive forces in the water conducting system.However, tension cannot be preserved in the watercolumn itself. From our point of view, the testingpressure in the pressure chamber primarily com-presses the expanded wall adherent bubble system.The direct measurement of the xylem tension by thepressure probe test is based on self-acting intake ofwater into the xylem conduits. There is no doubtthat this attraction of liquid is an indication ofcontractive forces in the water conducting system.However, in our view, the contractive forces are notcaused by a gradient of negative pressure in the

water column. In our opinion, the contractive forcesresult from the stressed hydro-pneumatic bubblesystem. That means we presume that negativepressure in terms of the cohesion theory is mimickedby the contractive forces of the wall adherentbubbles system.

The fragmentation of the water column cannotbe prevented even in our transport model. Airdissolved in the transported water diffuses inces-santly into the wall adherent bubbles. This way,embolism and dysfunction of the conduits isprogrammed as explained by Fig. 14. However,the dysfunction takes place after of a long period ofproblem-free water transport. To establish ourfindings, much more work is necessary. However,we hope that other investigators will criticallyevaluate our conclusions even in the present form.

References

Borghetti M, Edwards WRN, Grace J, Jarvis PG, RaschiA. The refilling of embolized xylem in Pinus sylvestris.Plant Cell Environ 1991;14:357–69.

Gibbs RD. Pattern in the seasonal water content of trees.In: Thimann KV, editor. The physiology of forest trees.New York: Ronald; 1958. p. 43–69.

Glavac V, Konies H, Ebben U. Auswirkung sommerlicherTrockenheit auf die Splintholz-Wassergehalte imStammkorper der Buche (Fagus sylvatica L.). HolzRoh Werkst 1990;48:437–41.

Hartig, Th., 1878. Anatomie und Physiologie der Holz-pflanzen, Berlin.

Laschimke R. Investigation of the wetting behaviour ofnatural lignin – a contribution to the cohesion theoryof water transport in plants. Thermochim Acta1989;151:35–56.

Laschimke R. Die Kohasionstheorie des Wasserferntrans-portes. Allg Forst Z 1990;36:993–7.

Laschimke R. Localisation of lignin in wood by fluores-cence confocal laser scanning microscopy. Mitt Bun-desforschungsanst Forst Holzwirtsch 1991;68:243–8.

Lundegardh H. Transport of water in wood. Ark Bot1954;3:89–119.

Malone M, Alarcon JJ. Only xylem-borne factors canaccount for systemic wound signalling in the tomatoplant. Planta 1995;196:740–6.

Milburn JA. Cavitation. A review: past, resent and future.In: Water transport in plants under climatic stress.Cambridge: Cambridge University Press; 1993. p. 14–26.

Milburn JA, Johnson RPC. The conduction of sap.Detection of vibrations produced by sap cavitation inRhicinus xylem. Planta 1966;69:43–52.

Ritman KT, Milburn JA. Monitoring of ultrasonic andaudible emissions from plants with or without vessels.J Exp Bot 1990;42:123–30.

Thurmer F, Zhu JJ, Gierlinger N, Schneider H, Benkert R,Geßener P, et al. Diurnal changes in xylem pressure

ARTICLE IN PRESS

Nature of xylem tension 1007

and mesophyll cell turgor pressure of the lianaTetrastigma voinierianum: the role of cell turgor inlong-distance water transport. Protoplasma 1999;206:152–62.

Tyree MT, Dixon MA. Cavitation events in Tuja occidenta-lis? Ultrasonic acoustic emissions from the sapwoodcan be measured. Plant Physiol 1983;72:1094–9.

Zholkevich VN. Water transport in plants and itsendogenous regulation. Moscow: Nauka; 2001 (inRussian).

Zimmermann MH. Xylem structure and the ascent of sap.New York: Springer; 1983.

Zimmermann U, Haase A, Langbein D, Meinzer F.Mechanism of long distance water transport in plants:a re-examination of some paradigms in the light ofnew evidence. Philos Trans R Soc London 1993;B341:19–31.

Zimmermann U, Schneider H, Wegner LA, Haase A. Waterascent in tall trees: does evolution of land plants relyon a high metastable state? New Phytol 2004;162:575–615.

Zwieniecki MA, Holbrook NM. Bordered pit structure andvessel wall surface properties. Implications for embo-lism repair. Plant Physiol 2000;123:1015–20.