Quantification of vulnerability to xylem embolism bench dehydration



Brendan Choat1, Danielle Creek1, Maria A. Lo Gullo2, Andrea Nardini3, Elisabetta Oddo4, Fabio Raimondo2, Tadeja Savi5, José M. Torres-Ruiz6, Patrizia Trifilo2, Alberto Vilagrosa7

Author Affiliations

1Hawkesbury Institute for the Environment, Western Sydney University
2Dipartimento di Scienze Chimiche, Biologiche, Farmaceutiche e Ambientali, Universit di Messina
3Dept. Life Sciences, University of Trieste
4 Department of Biological, Chemical and Pharmaceutical Sciences and Technologies, University of Palermo
5Department of Integrative Biology and Biodiversity Research, University of Natural Resources and Life Sciences, Vienna (BOKU)
6PIAF INRA-University Clermont-Auvergne
7Joint Research Unit, University of Alicante, CEAM


This technique was first proposed by Sperry and Tyree (1988) and describes vulnerability of a species to drought-induced embolism. Typically, branches are detached from the plant and allowed to dehydrate naturally on a laboratory bench top. The increase in xylem tension that occurs during the dehydration process results in the formation and spread of gas bubbles (embolism) in the xylem. Emboli block xylem conduits (vessels and tracheids) and reduce the hydraulic conductivity of the sample. The vulnerability of a species is described by a vulnerability curve, which is constructed by plotting the percent loss of hydraulic conductivity (PLC) against xylem water potential (ℼx) during the progressive dehydration of samples.

This technique is time consuming (samples may take 5-7 days to dehydrate) and destructive (typically require 5-30 large branches to be harvested). However, it is currently considered to be the most reliable reference technique for establishing vulnerability to embolism because (1) the embolism is induced naturally by transpiration (Sperry et al., 2012), and (2) large samples or whole plants can be dehydrated. It is also relatively easy to deploy under field conditions when a portable flow meter is used. The dehydration technique is preferred for species with long vessels (eg. ring porous tree, lianas), since more rapid techniques that rely on centrifugation or air injection may result in over estimation of vulnerability in these species (Choat et al. 2010; Cochard et al. 2010; Torres-Ruiz et al. 2014).


– Pressure chamber for determination of xylem water potential
– Safety glasses for pressure chamber measurments
– Flow meter or analytical balance (resolution to at least 4 decimal places)
– Aluminium foil and cling wrap or zip seal bags
– Large plastic bags (garbage bags)
– Plastic tubing
– Razor blades
– Plastic trough for cutting stems under water
– Syringes and hypodermic needles
– Syringe filters
– Captive air tank (optional)
– Stopcocks, 3 way manifolds and various fittings to connecting tubing
– Two or three prong extension clamp, and clamp holder, to hold graduated cylinder or syringe to make a pressure head
– Perfusing solution (usually 2-10 mM KCl, 1 mM CaCl, filtered to 0.2 μm)
– Small plastic beaker and pipette if using balance

Units, terms, definitions

PLC percentage loss of conductivity (%)
Ψx xylem water potential (MPa)
P50 xylem pressure at which PLC = 50%
k initial initial hydraulic conductance of the sample (kg s-1 MPa-1)
k max final hydraulic conductance of the sample (kg s-1 MPa-1)
Ks sapwood specific hydraulic conductivity (kg s-1 m-1 MPa-1)


Harvesting branches

Usually 5-30 large branches from separate trees are required to produce a vulnerability curve.

1. Selecting branches: Before selecting the length of the branches to be harvested, it is important to determine the maximum vessel length (MLV, see maximum vessel length protocol). The harvested branch length should be longer than the maximum vessel length (~1.5 MVL) to avoid air entry into the segment that will eventually be measured. In many cases, branches will need to be 1-2 m in length. The branch should ideally be healthy and sunlit. Where possible, branches should be cut from the plant underwater using a water filled funnel or plastic bag. However, this procedure is particularly difficult with tall trees.

2. Harvest branches before dawn if possible when ℼx is least negative (during or after rain event) to minimise problems associated with air entry into the cut surface. If this is not possible or plants have been grown under water deficit conditions, branches can be harvested in the evening and rehydrated overnight before measurements begin. See note 1.

FIGURE 1: Harvested branch with multiple branched sections. Segment for conductivity measurement selected and leaves wrapped in cling film and aluminium foil.

3. Once a branch is cut, place branch immediately into a bucket of water and recut the end. Once recut, remove from water and wrap the cut end with parafilm to limit evaporation. Seal the branch in a dark plastic bag humidified with damp (but not dripping) paper towel to prevent transpiration.

Preparing sections and measuring water potential

1. Identify an unbranched, healthy looking section on the branch for the hydraulic conductivity measurement. It is preferable to work with current-year material since older growth xylem my contain vessels that were embolised some time ago and are not refillable (see note 2). Mark the segment with an arrow indicating the normal direction of flow. The diameter of the segment is typically between 3-10mm in diameter. The length of the segment depends on the aims of each experiment and can be either (a) long segment (5-200 cm), longer than maximum vessel length, or (b) a short segment (1-5 cm) in which most vessels are cut open at both ends. A long segment will capture both lumen and pit component of hydraulic resistance in the branch. Short segments do not take into account pit hydraulic resistance but can still be used because only the ratio between initial and final hydraulic conductivity of the sample is required to calculate PLC. Short segments are also easier to flush since emboli are simply expelled from open vessels by positive pressure rather than needing to be dissolved back into solution. The influence of sample length on PLC curves has not been tested extensively but there is evidence that long and short segments produce similar results (Choat et al. 2010).

2. For determining ℼx, cover four leaves located near the segment chosen for the hydraulic measurements with aluminium foil covered plastic bags or alternatively plastic cling wrap and aluminium foil. In species with long vessels, it may be important to select leaves below the segment chosen for hydraulic measurement (i.e. upstream) to avoid possible air entry in the sample when leaves are excised. Place branch back into humidified garbage bag for at least 1 hour so that the water potential equilibrates across the sample.

3. After 1 hour, cut two bagged leaves off the sample and measure their water potential in the pressure chamber (link to pressure chamber protocol). The branch should be resealed in the plastic bag while conducting water potential measurements. This average value of bagged leaf water potential is equal to Ψx. If the two Ψx values are within 0.25 MPa, the sample can be prepared for hydraulic measurements. Otherwise, the branch should be left in the plastic bag to equilibrate for at least another 30 min before Ψx of the other two bagged leaves is measured.

FIGURE 2: Leaves covered for measurement of Ψx. A layer of cling wrap prevents transpiration from the leaf and a layer of aluminium foil prevents light from striking the leaf.

4. Cut the stem segment from the branch while it is submerged in water. Make the initial cut some distance upstream from intended final cut. This will allow xylem tension in the branch segment to relax, avoiding artefacts associated with the cutting under tension (Wheeler et al. 2013). Even when cuts are made under water, cutting under tension may result in the generation of emboli that were not present in the xylem prior to the cut and should therefore be avoided (Torres-Ruiz et al. 2015). The distance of the initial cut to the final cut depends on the species but should ideally be equal to or more than the maximum vessel length of the samples. In general it is best to make several cuts along the length of the branch before the final cut is made. In some cases, it may be necessary to allow the branch to rehydrate in a water filled container for 30 min before the initial cut is made (Wheeler et al. 2013). See note 3.

5. Carefully shave each cut surfaces of the stem segment with a sharp razor blade. This will ensure that xylem conduits are open during hydraulic measurements. In case of uneven bark surface, 1 cm of bark can be removed from each end of the sample.

6. Fit water filled soft tubing connectors onto each end of the sample, parafilm can be wrapped around either end to prevent leaks. Flush connectors with filtered purfusing solution using a syringe and hypodermic needle. If leaks are detected, connections can be tightened with cable ties or hose clamps. Caution must be used in order to avoid over tightening hose clamps and crushing the sample.

FIGURE 3: Prepared segment ready for connection to flow meter. Note direction of flow marked on stem and watertight seal obtained.

FIGURE 4: flow meter set up. Arrows indicate direction of flow.

FIGURE 5: Set up using an analytical balance. Arrows indicate direction of flow.

Measuring conductivity

1. Connect sample to flow meter or balance ensuring that stopcocks next to the sample are open to prevent pressure building up and pushing emboli out of the sample before the initial measurement is made. For details of hydraulic flow measurements see Constructing and operating a flow meter and Measurement of hydraulic conductivity protocols.

2. Ensure hydraulic head is set to desired pressure with filtered perfusing solution (see composition above). See note 4.

3. Start logging flow through the sample and wait for the flow rate to stabilise. This usually takes about 5 minutes. Once you have at least 1 minute of stable flow, stop logging and remove the sample from the flow meter. The initial hydraulic conductance of the sample (kinitial) is then calculated as:

kinitial = F / P

where F is the flow rate on to the balance (kg s-1) and P is the pressure drop across the sample (MPa) which is measured by the difference in height of water menisci in the supply reservoir and the balance. See note 4.

4. by the pressure drop across the segment. It is recommendable that the sample be kept under water or wrapped in wet paper to avoid evaporation during the measurements. Note that passive uptake of water by the sample can be an important source of error when the flow rate through the sample is low due to, e.g. high PLC levels or low-conductivity material (Torres-Ruiz et al. 2012). See note 5.

Flushing sample to measure kmax

1. Connect the sample to the captive air tank filled with filtered, degassed perfusing solution and set pressure to 100-150 kPa. Be careful not to over pressurize the tank! Flush the sample for 20-30 minutes at this pressure. If the sample is very short (most vessels open at each end) it can be flushed for just 10s in each direction using a syringe. See troubleshooting for how best to flush conifers.

2. Remove the sample from the pressure tank and reconnect it to the flow meter or balance. Start logging flow again and wait until it becomes constant. This can sometimes take longer after flushing and there is often an initial decline phase. The kmax of the sample can then be calculated as:

kmax = F / P

where F is the flow rate on to the balance (kg s-1) and P is the pressure drop across the sample (MPa).

3. In order to determine the optimum time to flush the sample, step 1 and 2 can be repeated until a constant value of kmax is reached. The sum of the times used to flush the sample will be equal to the time required to flush samples.

4. This procedure should be repeated on branches that have been dried on the bench over a period of hours/days. After each sample has been dried for a period of time it should be placed back in a plastic bag to allow water potential to equilibrate throughout the sample. See note 6 and 7.

Calculating PLC
The percentage loss of conductance (PLC) can be determined by:

PLC = 100(kmax – kinitial)/ kmax

If PLC=0%, none of the conduits were embolised.
If PLC=100%, all the conduits were embolised.
The xylem water potential is plotted on the x-axis and the PLC on the Y-axis to give you a vulnerability to embolism curve (See fitting vulnerability curves).

FIGURE 6: Completed vulnerability to embolism curve. Each point represents one stem sample. The solid red vertical line is equal to P50. Dashed vertical lines show 95% confidence intervals around P50.

PLC is can generally be calculated using ratio of initial and final hydraulic conductance. However, in some cases it may be preferable to plot a vulnerability curve based on decline in raw hydraulic capacity rather than PLC. Because hydraulic conductance (k) can vary dramatically between samples based on their physical and anatomical characteristics and independent of embolism, it is necessary to normalize k to the length (L) and sapwood area (As) of the stem sample used for measurement, yielding sapwood specific hydraulic conductivity (Ks, kg s-1 m-1 MPa-1; see hydraulic conductance and conductivity). When this formulation is used, only the initial measurement of flow is required since flushing is only necessary to calculate PLC.

1. Rehydrating samples: place branches with cut surface submerged in a bucket of water and leave covered with a large plastic bag in a cool dark place. If branches are initially collected from a tree by cutting in air, a second cut at the branch end should be done under water at a minimum distance of 0.2 times the MVL of the plant material to hydraulically reconnect part of the vessels and allow the branch to rehydrate. Overnight is generally sufficient for a complete rehydration.

2. When selecting the stem segment for measurement of PLC, it is preferable to work with current-year material since older growth xylem could contain embolised conduits that cannot be refilled by overnight rehydration. Flushing of these conduits during the hydraulic measurements can have a major influence in the kmax and therefore greatly inflate PLC values for that sample. This is particularly problematic for ring porous species in which past years of growth become embolised due to winter freezing and are not refilled in the spring.

3. To avoid possible excision artifacts (Wheeler et al. 2013; Torres-Ruiz et al. 2015), it is recommended to induce a rapid relaxation of xylem tension before hydraulic measurements. In some cases, longer time periods (~30 min) of rehydration may be necessary for relaxation of xylem tension in the segment prior to cutting. However, because this procedure might induce artificial refilling of embolized conduits (Trifilo et al. 2014), mechanical (sample girdling) or chemical (orthovanadate utilization) treatments to inhibit any refilling mechanism are recommended in some species, eg. Laurus noblis.

4. Selecting a hydraulic head: the hydraulic head is the pressure head set by the height of the water above your sample. The pressure of the hydraulic head is equal to the difference in height of water between the head and the container of water on the balance or height of the sample if using a flow meter. The height in m can be converted to the pressure in kPa using the equation:
P = h g

P= Pressure of the fluid (Pa) divide by 1000 to convert to kPa.
h= difference in height of top and bottom menisci (m)
=the density of the fluid (988.21 kg m-3 at 20C)
g= the acceleration due to gravity (9.806 m s-2)

The hydraulic head pressure must be sufficient to drive a resolvable flow rate (set by the resolution of the balance/flowmeter) but also not too high so as to push bubbles from the sample during measurement. This pressure can be calculated based on the maximum vessel lumen diameter for a species using the capillary equation:

P = 4 / Dmax

P= Pressure of the fluid (Pa) divide by 1000 to convert to kPa.
= the surface tension of water (0.0728 N m-1 at 20C)
Dmax= the maximum vessel diameter measured in cross section (m).

Thus, for a proper estimate of kinitial and, therefore, PLC, the maximum hydrostatic pressure difference between the two ends of the sample must be within the black area:

FIGURE 7: Illustrative maximum water head to determine the kinitial properly by avoiding the displacement of air bubbles trapped in an open vessels according to their maximum lumen diameter (from Xyl’em manual, Bronkhorst 2013).

5. Accounting for passive uptake of water by samples: The hydraulic conductance (k) is defined as the mass flow rate of water through a sample (F) divided by the pressure gradient driving the flow ( P). This assumes that, when P is zero, F should be zero too. However, there is a passive flow to the sample in the absence of any pressure (F0) that can have important consequences for the accuracy of the k measurements, especially when samples show low k values. In these cases, k should be determined as the slope of the F by P linear regression for a least two data points (i.e. considering F0 0). F0 can be then determined as the constant term in the linear regression analysis, i.e. the value at which the fitted line crosses the y-axis (y intercept). It is important to check the F0 regularly during the construction of a vulnerability curves especially when PLC starts to show relatively high values (i.e. >50%) to determine when it should be taken into account. Thus, if F0 >5% F, k should be determined as the slope of the F by P linear regression for two or more points (i.e. considering F0). If F0 <5% F, then, k can be accurately determined from the F resulting at a single P point (i.e. not considering F0). See Torres-Ruiz et al. (2012) for more info about the passive water uptake and how accounting it.

6. It is possible to use a fan to speed dehydration but very fast dehydration should be avoided because it can induce a high heterogeneity of water stress in the branch. How long branches need to be dried for will depend on the species and conditions. As more samples are completed it will become obvious what values of water potentials need to be targeted. The objective is to produce a curve with an even spread of data points between 0 and 100% PLC. This will often require some trial and error and several collection phases for branch samples.

7. Curves are inherently variable so don’t be disheartened if you have a few measurements that don’t fit the theoretical curve.

Notes and troubleshooting tips

Declining conductance
It is quite common for the conductance of a segment to decline rapidly after connecting the sample to the apparatus with conductance tending to stabilise after about 5 minutes. If however the conductance continues to decline, this could be due to the presence of tiny particles in the solution progressively blocking the conduits. It is critical to use very well filtered water, periodically change the syringe filters and clean the tubing frequently (at least once a week) with a dilute bleach solution to minimise problems. Bacterial growth in samples can also become a problem over time, although this is generally only relevant over a time scale of days.

There is some evidence that flow may also be reduced by gas coming out of solution during the conductivity measurement (Espino and Schenk 2011). The impact of this problem varies between species and studies and is an area of ongoing research. Generally, if flow has stabilised within the first five minutes of measurement then this effect can be ignored. If flow continues to decline then the measurements can be made with degassed water. A contact degasser (mini-module) is generally best suited for the purpose of degassing the perfusate.

Some species, particularly in the tropics, may produce copious resin or mucilage when cut. In some cases it is possible to reduce the impact of this wounding effect by various means. Removing the bark can help overcome resin plugging the conduits. The end of the sample may also be shaved by a few mm with the fresh razor blade periodically to remove mucilage on the cut surface. In other cases it is not possible to make measurements on these species because flow ceases quickly after the sample is connected to the flow meter. In these cases alternative methodology must be used to generate VCs.

The effects of declining flow may also materialise between the initial and final flow measurements. If the final flow rate is lower that the initial then a negative PLC will be calculated indicating that clogging by particles, gas or resin/mucilage is a serious issue. This can be remedied by identifying the source of the problem and taking the appropriate precautions described above.

Flushing conifers
Conifer segments when flushed according to the method described often results in the pit membranes being permanently pushed against the cells walls of tracheids (aspirated) resulting in very low to no flow of water through the segment. To overcome this, embolism is best removed by placing the stem segments in filtered perfusing solution under vacuum for 24 hours and then continuing with the protocol.

Figure 8: Transverse HRCT micrographs of a Pinus sylvestris current-year shoot at ℼx of -3.0 MPa (panel a) and after being vacuum infiltrated for 8-10 hours (panel b). Images show the effectiveness of the vacuum infiltration technique for removing the embolisms in conifer shoot samples. Cavitated or air-filled tracheids are observed as black and the functional ones in grey (Torres-Ruiz et al. unpublished data).

Samples with low hydraulic conductivity
For samples with low values of stem conductivity (eg. seedlings or species growing in arid environments), it is important to evaluate the type of sample and the maximum and minimum values of conductivity for each species. If the native conductivity is very low then it may be below level of resolution of the device being used to measure conductivity (i.e. the flow meter sensor). In this case it is important to make adjustments so that flow rates are within the optimal range for each flow meter. This can be done by increasing the hydraulic head pressure, but caution must to taken that it is not increased over the limit that will cause air bubbles to be displaced from the stem (see note 4). The sample can also be shortened to reduce hydraulic resistance but this may result in more vessels being cut open.

Links to resources and suppliers

Other methods for measuring vulnerability to embolism described here
Methods for the construction of a flow meter can be found here.
See using a pressure chamber here.
Maximum vessel length protocol: Link TBA
Measuring xylem hydraulic conductivity: Link TBA
Sperry Lab Methods: http://biologylabs.utah.edu/sperry/methods.html#hydraulic_conductivity

Literature references

Bronkhorst (2013) XYL’EM instruction manual, version 2.1.

Click to access xyl’emplusmanual.pdf

Choat B, Drayton WM, Brodersen C, Matthews MA, Shackel KA, Wada H, McElrone AJ (2010) Measurement of vulnerability to water stress-induced cavitation in grapevine: a comparison of four techniques applied to a long-vesseled species. Plant Cell and Environment 33, 1502-1512.

Cochard, H., Badel, E., Herbette, S., Delzon, S., Choat, B., Jansen, S. (2013) Methods for measuring plant vulnerability to cavitation: A critical review. Journal of Experimental Botany, 64 (15), 4779-4791

Cochard H, Herbette S, Barigah T, Badel E, Ennajeh M, Vilagrosa A. 2010. Does sample length influence the shape of xylem embolism vulnerability curves A test with the Cavitron spinning technique. Plant, Cell and Environment 33: 1543-52.

Espino, S., and H. J. Schenk. (2011) Mind the bubbles: Achieving stable measurements of maximum hydraulic conductivity through woody plant samples. Journal of Experimental Botany 62:1119-1132

Sperry JS, Christman M, Torres-Ruiz JM, Taneda H, Smith DD. 2012. Vulnerability curves by centrifugation: is there an open vessel artefact, and are “r” shaped curves necessarily invalid Plant, Cell and Environment 35: 601-610.

Sperry JS, Tyree MT (1988) Mechanism of water stress-induced xylem embolism. Plant Physiology 88, 581-587.

Torres-Ruiz JM, Sperry JS, Fernandez JE (2012) Improving xylem hydraulic conductivity measurements by correcting the error caused by passive water uptake. Physiologia Plantarum 146, 129-135.

Torres-Ruiz JM, Cochard H, Mayr S, Beikircher B, Diaz-Espejo A, Rodriguez- Dominguez CM, Badel E, Fernández JE (2014) Vulnerability to cavitation in Olea europaea current-year shoots: further evidence of an open-vessel artefact associated with centrifuge and air-injection techniques Physiologia Plantarum 152, 465-474.

Torres-Ruiz JM, Jansen S, Choat B, McElrone AJ,Cochard H, Brodribb TJ, Badel E, Burlett R, Bouche PS, Brodersen CR, Li S, Morris H, Delzon S (2015) Direct X-Ray microtomography observation confirms the induction of embolism upon xylem cutting under tension. Plant Physiology 167, 40-43.

Trifilo P, Raimondo F, Lo Gullo MA, Barbera PM, Salleo S, Nardini A (2014) Relax and refill! Xylem rehydration prior to hydraulic measurements favours embolism repair in stems and generates artificially low PLC values. Plant, Cell and Environment 37, 2491-2499.

Wheeler J. K., Huggett B. A., Tofte A. N., Rockwell F. E., Holbrook N. M. (2013). Cutting xylem under tension or supersaturated with gas can generate PLC and the appearance of rapid recovery from embolism. Plant Cell Environ, 36(11), 1938-1949.

Health, safety & hazardous waste disposal considerations

1. Use caution when shaving stems with a razor blade or scalpel. It is easy to slice open a finger in this situation!

2. Follow all safety instructions for using the pressure chamber when making measurement of Ψx. This instrument uses very high gas pressure!

3. When using a captive air tank to flush samples, be careful never to over-pressurize the tank! It is preferable to use a regulator that can be set to below the maximum pressure threshold of the tank.

Leave a Reply