Xylem sap flow as a major pathway for oxygen supply to the sapwood of birch (Betula pubescens Ehr.)



    1. Institut für Ökologische Pflanzenphysiologie und Geobotanik, Universität Düsseldorf, Universitätsstraße 1, D-40225 Düsseldorf, Germany
    Search for more papers by this author

D. Gansert. Fax: +49 211 81 13335; e-mail, gansert@uni-duesseldorf.de


The role of xylem sap flow as an aqueous pathway for oxygen supply to the wood parenchyma of Betula pubescens saplings was investigated. Using micro-optode sensors the oxygen status of the sapwood was quantified in relation to mass flow of xylem sap. Sap flow was gradually reduced by an increasing oxygen depletion in the root space. The effect of sap flow on radial O2 transport between stem and atmosphere was assessed by a stoichiometrical approach between respiratory CO2 production and O2 consumption. Restriction of sap flow set in 36.5 h after the onset of O2 depletion, and was complete after 71 h. Interruption of sap flow drastically increased the O2 deficit in the sapwood to 70%. Sap flow contributed about 60% to the total oxygen supply to the sapwood. Diurnal O2 flow rates varied between 3 and 6.3 nmol O2 m−2 leaf area (LA)  s−1 during night- and daytime, respectively. Maximum O2 flow rates of 20 nmol O2 m−2 LA s−1 were reached at highest sap flow rates of 5.7 mmol H2O m−2 LA s−1. Sap flow not only affected the oxygen status of the sapwood but also had an effect on radial O2 transport between stem and atmosphere.


It has long been recognized that the transpiration stream plays a vital role in aeration of living tissues surrounded by the cambium of woody plants (Bailey 1913; Haberlandt 1914). The importance of xylem sap flow for oxygen supply to and removal of carbon dioxide from parenchymatous tissues of the sapwood was clearly recognized by Bailey (1913) when he stated that the xylem of arborescent plants has an important function for transport of gases dissolved in water from the roots to the cambium and leaves. Axial oxygen transport in flood-tolerant plant species from aerial parts to the roots has been a subject of broad interest (Hook, Brown & Wetmore 1972) with special reference to physical processes involved in gaseous diffusion of oxygen through intercellullar gas-spaces such as thermo-osmosis or Graham's law of diffusion (Armstrong 1979; Grosse, Frye & Lattermann 1992; Grosse 1997). However, long-distance gaseous oxygen transport primarily takes places in the intercellular gas-space continuum of the outer cortex surrounded by the phellogen (Hook et al. 1972), which has been identified as a thermo-osmotically active partition in several tree species (Eschrich 1995; Grosse 1997).

Radial gaseous diffusion of atmospheric oxygen across the cambial sheath towards the wood parenchyma can only be effective if continuity through cambium pores is maintained. In trees of Betula pendula Roth radial influx of oxygen into the sapwood has been suggested as a gaseous pathway mainly effective at night, by which xylem sap will be enriched with dissolved oxygen (Gansert, Burgdorf & Lösch 2001). With respect to species-specific wood anatomy, oxygen supply to the various types of wood parenchyma (Braun 1970; Grosser 1977; Schweingruber 1978; Eschrich 1995) principally depends on both the abundance of intercellular spaces and their effectiveness for gaseous diffusion (i.e. the gaseous path), and xylem sap flow – the aqueous path – that primarily supports those parenchyma strands that shield water transport from embolism and regulate secretion of osmotica and hormones into the xylem sap. Oxygen concentrations measured in sapwood of different anatomical structure and metabolic activity vary over a broad range from pronounced hypoxia to near saturation (Eklund 1990, 1993, 2000; Gansert et al. 2001; Del Hierro et al. 2002; Mancuso & Marras 2003).

New types of miniaturized, high-sensitive oxygen sensors allow improved quantitative analyses of the oxygen status and oxygen flow rates in different compartments of the woody cormus on a high-resolution time scale. Previous efforts on the measurement of oxygen concentrations in the sapwood by use of these sensors focused on the preclusion from interference with atmospheric oxygen, and maintenance of xylem sap flow by avoidance of embolism. Thus, underwater access to the sapwood was used as a methodical approach for in situ measurements of diurnal and seasonal variations of endogenous oxygen concentrations in woody plants (Gansert et al. 2001). This approach may contribute to identifying circumstances and metabolic dependencies that are responsible for the different oxygen status of plant shoots under varying ontogenetic, environmental, and seasonal conditions. Using the new type of micro-optode oxygen sensors, the purpose of this study was: (1) to investigate the role of xylem sap flow as an aqueous pathway  for  oxygen  supply  to  the  sapwood  parenchyma of young mountain birch (Betula pubescens Ehr.); (2) to quantify its effect on the oxygen status of the sapwood; and (3) to evaluate the potential effect of xylem sap flow on oxygen transport between stem and atmosphere during stem respiration.


The experimental set-up

In early June 2002, two potted plants of 5-year-old field-grown B. pubescens saplings about 1.7 m tall were brought into a climate chamber and preconditioned to the experimental light and temperature regime for 10 d. The soil substrate was natural loess (volume = 8.3 × 10−3 m3) in which the roots grew undisturbed for 1 year. In order to simulate a near-daylight spectrum, HRI-T 400 W/D lamps (Radium; Osram, München, Germany) were used as the light source during a 13 h day period (0700 to 2000 h). Prevention of upper crown parts from excessive warming was achieved by insertion of a glass pane (4 mm thick) 8 cm beneath the light unit. Horizontal air ventilation between the lamps and the glass pane efficiently reduced the effect of excessive heat production on leaf gas exchange. Thus, air temperature measured at 27 cm beneath the lamps fluctuated between 21 and 23 °C during night-time and between 23 and 27 °C during daytime. Because air ventilation started about 30 min prior to the onset of light a drop in air temperature of 2–3 °C always marked the end of the night.

Each pot was packed in a polyethylene bag leaving the top open, and immersed in the water bath of a cryostat (Type MC5S0; Colora Messtechnik GmbH, Lorch, Germany). In each pot three copper–constantan thermocouples were inserted at different soil depths (7, 13 and 17 cm from the surface) 8 cm radially off the stem base. A mean soil temperature of 17 ± 0.3 °C at a depth of 13 cm was maintained by permanent temperature control of the water baths. Figure 1 presents the experimental set-up for both trees schematically. Air temperature and relative humidity were simultaneously measured at 27 and 95 cm beneath the lamps by use of two Vaisala sensors (HMP 35 A; Vaisala, Helsinki, Finland). Photosynthetic photon flux density (PPFD) was measured at half crown length, 55 cm beneath the lamps by use of three sensors (SKP 215; Skye Instruments Ltd, Llandrindod Wells, Powys, UK) which were evenly positioned over a distance of 80 cm. Global radiation (Pyranometer SKS 1110; Skye Instruments Ltd) was also measured at the same height as the PPFD sensors. On both trees two Plexiglass cuvettes (own construction) were positioned on the stem at 96 and 101 cm above the stem base, respectively. At those heights the stems showed nearly the same diameter of about 8 mm. The upper cuvette was designed for measurement of O2 concentration in the cuvette atmosphere and in the sapwood, whereas the lower one was prepared for measurement of CO2 efflux rates. Both cuvettes are described in detail in the following section. To preclude interference in CO2 and O2 gas exchange due to bark photosynthesis, all internodes and the cuvettes were shielded from light by strips of black silk cloth wrapped around the woody plant parts. The data of all microclimatic variables were synchronously recorded at 10 s intervals and averaged over 5 min intervals by data loggers (Squirrel 1250 series; Grant Instruments Ltd, Barrington Cambridge, UK).

Figure 1.

Scheme of the experimental set-up applied to reversibly induce gradual reduction of xylem sap flow in 5-year-old saplings of B. pubescens by creation of oxygen deficiency in the root space (see text). Centimetres given in parentheses indicate the distance of each measuring spot beneath the light unit in the climate chamber. PPFD indicates sensors used to measure photosynthetic photon flux density. The cuvettes used to measure O2 and CO2 are shown in detail in Fig. 2.

Cuvettes for measuring stem gas exchange

Cuvettes for the measurement of O2 and CO2 gas exchange between ambient air and the stem were made out of Plexiglass as shown schematically (Fig. 2). Each cuvette consisted of two semi-cylinders (wall thickness = 5 mm, thickness of top and bottom = 6 mm) which enclosed a stem internode of up to 8 mm in diameter and 20 mm in length. A rubber seal fixed in a groove (2 mm wide, 1.5 mm deep) along the plane side of one semi-cylinder ensured a leak-proof seal when both semi-cylinders were pressed onto each other by two steel strings. Silicon rubber (Loctite no. 5910;  Loctite  GmbH,  München,  Germany)  was  used to seal the remaining gap between the stem and the cuvette. Each cuvette was equipped with four Plexiglass tubes (dout = 10 mm, l = 62 mm) with different inner diameters to be connected to different gas analysing instruments. The cuvette being used for O2 measurements was sealed from ambient air by insertion of a silicon septum (3 mm thick) into each Plexiglass tube. The one being used for CO2 measurements was run in a permanent open airflow modus (see below).

Figure 2.

Detailed drawing of the O2 (upper) and CO2 (lower) Plexiglass cuvette used for measurements of stem gas exchange of B. pubescens saplings (own construction). The O2 cuvette was hermetically sealed from the atmosphere. Measurements of O2 concentrations in the sapwood (indicated by the cannula to the right, housing the glass fibre with the O2 sensor at its tip), and in the cuvette atmosphere were carried out simultaneously. Atmospheric pressure inside the cuvette was maintained by using soda lime as CO2 absorbent (hatched area) to compensate for an increase of CO2 partial pressure due to CO2 efflux from the stem. CO2 efflux was continuously measured under open flow conditions by use of a porometer.


The concentrations of molecular oxygen were simultaneously measured both, in the cuvette atmosphere (Vair = 5.28 cm3), and in the aqueous phase of the sapwood at a depth of 2.9 mm from the stem surface. Measurements of molecular oxygen were performed by fibre-optical micro-optode sensors (see below). Two opposite tubes were used as mechanical guides for the plastic syringes and the attached cannulae that penetrated through the silicon septa. The third tube was filled with soda lime as the CO2 absorbent (Fig. 2). Thereby, during the experiments total pressure in the O2 cuvette remained near constant at atmospheric pressure because an increase of CO2 partial pressure due to continuous CO2 efflux from the stem, was compensated by an equivalent chemical binding of CO2. Maintenance of atmospheric pressure in the O2 cuvette was confirmed by use of a calibration manometer (DP 205; Ametek Precision Instruments GmbH, Meerbusch, Germany) attached to one of the tubes. The air temperature in the cuvette was measured  by a Pt-100 temperature sensor (2 × 2.3 × 0.5 mm; S105PD4A; Telemeter Electronic, Donauwörth, Germany), inserted into the cuvette through the fourth tube, and connected to the MICROX® oxygen analysing system (PreSens GmbH, Regensburg, Germany) Measurements of oxygen concentration and temperature were carried out at 5 min intervals.

The optical measurement of molecular oxygen was performed by an improved MICROX system with a micro-optode sensor (sensor type B; PreSens GmbH). The MICROX system consists of a pulsing laser light source (excitation wavelength 505 nm) to excite the O2 sensor, an optical fibre as signal transducer, a photodetector and the optical sensor. The micro-optode consists of a plastic syringe (d = 7 mm) which houses the optical fibre with the O2 sensor – a luminophor – on its tapered tip (d ≤ 30 µm). In order to move the sensor out of the cannula (dout = 0.5 mm, l = 37 mm) the fibre is fixed to the plunger. The MICROX system measures the luminescence lifetime of an immobilized O2 sensitive luminophore which is inversely proportional to the number of oxygen molecules that reversibly react with the luminophor. By this method no oxygen is consumed during the measurement, and the signal is independent of changes in sap flow velocity. Details of this optic–chemical O2 measurement have been described elsewhere (Gansert et al. 2001).

CO2 cuvette

Measurements of CO2 concentration were performed by use of LCA-4 porometers (ADC BioScientific Ltd, Hoddesdon, UK) each being attached to the CO2 cuvettes of the trees as shown in Fig. 2. A permanent airflow at a rate of 0.2 L min−1 was maintained throughout the experimental period. Air temperature in the CO2 cuvettes was measured by the same type of Pt-100 sensor mentioned above, which was connected to the temperature input side of the porometer. Due to continuous operation of two porometers (one for either tree), porometer-specific drifts in CO2 analysing  accuracy  were  minimized  by  re-calibration  at 3 d intervals. Because the O2 and CO2 cuvettes were positioned close to each other on internodes of the same diameter it was assumed that the rate of CO2 efflux in the O2 cuvette was nearly the same as measured in the CO2 cuvette. Measurements of CO2 efflux were synchronized with measurements of O2 concentration and also recorded at 5 min intervals.

Leaf photosynthesis and transpiration

Rates of photosynthesis, dark respiration, and transpiration of leaves were measured by use of a third LCA-4 porometer with a non-climatized leaf cuvette. Measurements were carried out in two modes. In the random mode 10 leaves per tree were selected and all parameters necessary for calculating the CO2 and H2O flux rates were recorded several times throughout the day. During all measurements the leaves were handled in the same order. In the automatic record mode one leaf of the upper crown of a tree, that showed a representative photosynthetic rate, was inserted into the leaf cuvette and single measurements were automatically recorded at 5 min intervals for 1 week.

Xylem sap flow

Xylem sap flow was measured by use of Dynagage® stem flow sensors (SGA-10; Dynamax, Inc., Houston, TX, USA) combined with a ‘Flow2’ sap flow monitor system (Dynamax Inc.). The gauges, one per tree, were positioned about 20 cm above the stem base at stem diameters of 12.2 and 12.6 mm for saplings ‘A’ and ‘B’, respectively. Tight contact between the flexible heater and the stem surface was ensured by use of four cable ties per gauge in addition to the velcro straps.

The manufacturer's calibration procedure recommends that the Ksh values, namely the effective thermal conductance of the sheath of materials surrounding the heater, should be measured under conditions of no sap flow. For this purpose, the gauges were mounted around a dry stick of birch wood. Wrapping the gauges in several layers of aluminium foil ensured sufficient thermal insulation from ambient temperature variations. The measured Ksh values at zero sap flow gave 0.8349 (resistance of gauge 1 = 132.5 Ω) and 0.812 W mV−1 (resistance of gauge 2 = 134.6 Ω). Following Smith & Allen (1996) the voltage across the heater was adjusted to 3.58 V for both gauges. Thereby the risk of overheating the stems at low sap flow rates was removed. All voltages necessary for calculating mass flow of xylem sap were logged synchronously with CO2 and O2 measurements at 5 min intervals (CR10; Campbell Scientific, Inc., Logan, UT, USA).

The leaf area specific sap flow rate (FLA) was calculated using the total leaf area of a tree as the reference unit. Hence, at the end of the experiment the area of each leaf was measured using a LI-3100 area meter (Li-Cor Inc., Lincoln, NE, USA) and summed. The total leaf area of saplings ‘A’ and ‘B’ was 0.365 and 0.451 m2, respectively.

For reasons of quantitative analysis of the role of xylem sap flow for endogenous oxygen transport the experimental set-up used in this study focused on the simultaneous measurement of several ecophysiological parameters of relevance for the interpretation of oxygen transport processes. Limited resources of equipment confined the investigations to only two tree individuals of the same phenological state.

Experimental restriction of xylem sap flow

Xylem sap flow was gradually reduced by an increasing oxygen depletion in the root space that lasted for several days. For that purpose, atmospheric oxygen supply to the soil and root system was interrupted by sealing the pots with the polyethylene bags and warming them up to 32 °C in the water baths of the cryostats. It was assumed that the temperature-dependent exponential increase in respiration of roots and soil organisms should cause a corresponding oxygen deficiency and even hypoxia in the soil atmosphere. This in turn should lead to an interruption of water uptake by the roots. Recovery from restriction of xylem sap flow was induced by subsequent re-aeration of the soil atmosphere.


According to Henry's law, the concentration of dissolved oxygen in water saturated with air, inline image(mol L−1), was calculated as:


where inline image is the partial pressure of oxygen in air at atmospheric conditions (21 278 Pa), Cw the number of moles H2O per litre (55.6), and inline image (Pa × 109) the solubility coefficient of oxygen in water. Within the temperature range (T) from 0 to 30 °C the temperature dependence of inline image was calculated as:


derived from linear regression analysis of inline image values determined at different temperatures (Von Willert, Matyssek & Herppich 1995). The relative concentration of dissolved oxygen (% water saturation) at a given temperature could then be converted into the absolute concentration in the aqueous solution, [O2], given in µmol L−1. The oxygen flow rate (µmol O2 m−2 leaf area s−1) in the sapwood was then calculated from the known values of [O2] and the leaf area specific sap flow rate FLA(mmol H2O m−2 s−1).

The temperature dependence of CO2 efflux from the stems was quantified by use of the Arrhenius-equation:


where inline image = CO2 efflux rate, here given as µmol CO2 m−2 s−1 (the superscript ‘st’ stands for ‘stem’), T = temperature (K), and R = 8.314 (J K−1 mol−1). The coefficients a and b were calculated by non-linear regression analysis according to the database available (SigmaPlot 6.00; SPSS Inc., Chicago, Il, USA) . The coefficient b is the activation energy (kJ mol−1); namely the amount of heat energy necessary to activate respiration metabolism (Schopfer & Brennicke 1999).


Xylem sap flow and leaf gas exchange under control conditions

In mid-June, the experiment started after 10 d of acclimatization of the plants to the prevailing light regime (530 µmol photons m−2 s−1 on average in the upper crown), and room temperature conditions (20 °C at the end of the night to maximum values of between 26 and 28 °C at the end of the day), and lasted for 2 weeks. The mean rate of leaf net CO2 gas exchange (inline image) in the upper crown of both trees was 6–8 µmol CO2 m−2 s−1 with little variation throughout the day (Fig. 3a). Mean stomatal conductance (gs)  of  the  leaves  ranged  from  100  to  150 mmol H2O m−2 s−1 indicating open stomata. Thus, the leaves transpired at mean rates between 1.2 and 1.8 mmol H2O m−2 s−1. Dark respiration  rates  were  nearly  constant  at  1 µmol CO2 m−2 s−1 during night-time. These data of leaf gas exchange were measured at constant soil temperature conditions of 17 °C.

Figure 3.

(a) Diurnal pattern of net CO2 exchange rate (inline image) and stomatal conductance (gs) continuously measured on a single leaf of B. pubescens in a climate chamber  in  mid-June  2002.  Daily  means of inline image (squares) and standard deviation of 10 leaves randomly distributed in the upper crown are also indicated. inline image and gs show a transient increase and decrease after the onset of light at 0700 h, due to abrupt temperature changes inside the non-climatized cuvette. (b) Diurnal course of temperature of ambient air (T-room), in the leaf cuvette of the porometer (T-cuvette), of leaves inserted into the cuvette (T-leaf, cuvette), and at ambient conditions 32 cm beneath the light source (T-leaf, ambient) during an 11 h night (hatched area) and a 13 h day period. (c) Photosynthetic photon flux density was measured at half crown length, 55 cm beneath the light source (PPFD, ambient), and incident on the leaf enclosed in the porometer cuvette (PPFD, cuvette). Global radiation was also measured at half crown length (see Fig. 1).

When sap flow was unimpeded, an increase of gs during night-time was observed several times, as on 14 June (Fig. 3a) or 18 June (day 6 of the experiment, Fig. 4b). Stem sap flow under these control conditions was in a range between 1.1 and 1.6 mmol H2O m−2 LA s−1 in the light period, and 0.6–0.9 mmol H2O m−2 LA s−1 on average in darkness.

Figure 4.

(a) Diurnal temperature regime in the climate chamber (T-room), in the porometer cuvette (T-cuv), and at 13 cm soil depth (T-soil) of potted B. pubescens saplings (sapling ‘B’) during the experimental period from 15 to 22 June 2002 (days 3–10). (b) Continuous record of net CO2 exchange rate (inline image) and stomatal conductance (gs) of a single leaf before, during, and after recovery from gradual interruption of xylem sap flow. Daily means of inline image (squares) and standard deviation of 10 leaves randomly distributed in the upper crown are also indicated. (c) Leaf area specific sap flow rate (FLA), and transpiration rate of a single leaf (E) as well as mean values (n = 10) of upper crown leaves (dots) indicate restriction of the transpiration stream on day 6 of the experiment from 0930 h onward. (d) Daily variations of the relative O2 concentrations in the cuvette atmosphere (O2 cuvette), and dissolved in the sapwood at 2.9 mm depth from the stem surface (diameter = 7.6 mm). On day 4, from 2100 h onward, O2 depletion in the root space was induced (indicated by arrow).

Restriction of xylem sap flow

Oxygen depletion in the root space started on day 4 of the experiment from 2100 h onward, when the pots were sealed and warmed up to 32 °C (T-soil, Fig. 4a). In the following light period leaf area specific sap flow rate (FLA) increased by 34% compared with previous rates at 17 °C soil temperature (Fig. 4c). Thus, highest rates of inline image (10 µmol CO2 m−2 s−1) and gs (150 mmol H2O m−2 s−1) were also measured (Fig. 4b). On the second day of oxygen depletion (day 6), restriction of sap flow was observed by a decrease of FLA from 0930 h onward towards 0.3 mmol H2O m−2 s−1 at the end of the day (Fig. 4c). Due to reduced sap flow, mean leaf transpiration rates (E) dropped from 1.94 to 0.96 mmol H2O m−2 s−1, and inline image was reduced by half. On day 7, from 2000 h onward, no sap flow was recorded by the Dynagage sensors, which was 71 h after the onset of O2 depletion (Fig. 4c). Cessation of transpiration and net CO2 uptake by the leaves indicated that interruption of sap flow caused a complete closure of the stomata from the fourth day of oxygen depletion onward. The majority of the leaves could not tolerate such water shortage so that trees ‘A’ and ‘B’ lost 92 and 87% of their total leaf area, respectively, within the following days.

After re-aeration of the roots on day 8 of the experiment from 2100 h onward, FLA rose to 2.3 mmol H2O m−2 s−1 within 20 min, in spite of the darkness. It gradually increased further to a final maximum rate of 5.7 mmol H2O m−2 s−1 (Fig. 4c). Thus, at its maximum, FLA showed a near four-fold increase after re-aeration. Continuous porometric measurements also revealed recovery from water stress.

The effect of sap flow restriction on changes in oxygen concentration in the sapwood

The concentrations of molecular oxygen in the O2 cuvette atmosphere were always clearly different between day and night. During daytime, the O2 concentration was reduced to about 94% of the atmospheric concentration (100% air saturation corresponds to 21% O2 concentration of the atmosphere). During night-time, the O2 concentration was in the range between 96 and 100% of air saturation. This diurnal pattern was measured with no ongoing bark photosynthesis and seemed not to be affected by the gradual interruption of sap flow (Fig. 4d).

Under conditions of unimpeded sap flow, the O2 concentration measured in the sapwood also showed a clear alternating diurnal pattern, but at a level considerably below concentrations of an aqueous solution of the same temperature saturated with air. During the day, [O2] was reduced to 75% of water saturation, whereas it increased to 82% on average at night (Fig. 4d). Thus, the sapwood of the B. pubescens saplings showed a permanent oxygen deficit. This O2 deficit alternated between 15% during the night, rising to 30% during the day.

The gradual reduction of sap flow caused a corresponding decrease of the O2 concentration in the sapwood. As a result of the first decrease of FLA on day 6 of the experiment, [O2] decreased from 75 to 67% at the end of the day. During the following night, recovery towards concentrations of previous nights took place when FLA also rose. However, the increase in [O2] was greater than might be expected from the sap flow increase. During the next day, [O2] dropped further to 58% of water saturation. When sap flow ceased completely on day 8, [O2] also reached its minimum of 30% of water saturation at the end of the light period (Fig. 4d). Hence, corrected for differences in temperature, sap flow rates between 1 and 1.5 mmol H2O m−2 LA s−1 corresponded to an oxygen concentration of about 119 µmol O2 L−1 which is 61% of the initial concentration of dissolved oxygen in the sapwood at unimpeded sap flow (Fig. 5). A similar result was obtained from the second tree, in which sap flow contributed 58% to the O2 concentration in the sapwood during daytime.

Figure 5.

Daily variation of dissolved oxygen concentration and leaf area specific oxygen flow rate in the sapwood of a 5-year-old B. pubescens sapling in mid-June 2002. Decreasing oxygen concentration towards a minimum of 77 µmol O2 L−1 is in correspondence with the gradual reduction of xylem sap flow. Note the nocturnal increase in oxygen concentration even when sap flow was interrupted.

After re-aeration of the roots and recommencement of sap flow to 2.3 mmol H2O m−2 s−1[O2] initially increased from 30 to 48% (77–134 µmol O2 L−1) within 3.5 h (Fig. 4d). Under these conditions the specific oxygenation rate was 2 nmol O2 L−1 s−1 per mmol H2O m−2 LA s−1 sap flow. However, the process of O2 re-loading of the sapwood was retarded so that it took 42 h longer to approximate the initial [O2] of 200 µmol L−1.

Leaf area-based oxygen flow rates in the sapwood

From a knowledge of both, the temperature-corrected absolute concentration of dissolved oxygen and the FLA, the oxygen flow rate in the sapwood can be calculated using the tree-specific total leaf area (LA) as the reference unit. During the light periods and under conditions of unimpeded sap flow, the oxygen flow rate in the sapwood reached 5–6.3 nmol O2 m−2 LA s−1 in saplings ‘B’ and ‘A’, respectively (Fig. 5). In line with the diurnal variation of FLA the oxygen flow rate decreased to about 3 nmol O2 m−2 LA s−1 at night. Recommencement of sap flow after re-aeration of the roots caused a rapid increase in oxygen flow to rates of about 6 nmol O2 m−2 LA s−1 within the first 20 min of re-aeration. Highest oxygen flow rates of 20 nmol O2 m−2 LA s−1 were reached  at  highest  sap  flow rates  of 5.7 mmol H2O m−2 LA s−1.

Stoichiometrical relations between CO2 efflux and O2 exchange across the stem surface

The rates of CO2 efflux from the stem (inline image) of either tree sapling were not affected by the gradual decrease of sap flow. For example on day 8 of the experiment, when no sap flow was measured, inline image was 2.4 µmol m−2 s−1 on average during daytime. Similar rates were also measured on days 4–6, when sap flow occurred (Fig. 6). A correlation of inline image versus cuvette temperature showed a characteristic exponential temperature-dependent relationship of CO2 release, as described by the Arrhenius-equation with correlation coefficients of 0.74 and 0.8 for saplings ‘A’ and ‘B’, respectively (Fig. 7). Both results indicate that inline image was primarily due to oxidative respiration of stem tissues and was little influenced by sap flow.

Figure 6.

Daily variation of CO2 efflux from a stem (d = 7.8 mm) of a B. pubescens sapling with day- and night-time means (dots), and daily pattern of O2 exchange between stem and cuvette atmosphere (bold line, defined in the text). During daytime, a change from O2 efflux (days 4–6) to O2 influx into the stem (days 7–8) occurred when O2 depletion in the roots caused a gradual decrease of xylem sap flow. The O2 deficit in the cuvette atmosphere, given as difference between CO2 efflux rate and O2 exchange rate, increased from smaller to higher values than that equivalent to respiratory oxygen consumption. Equivalence (‘eq’) between CO2 production and O2 consumption was reached when there was no O2 deficit in the cuvette.

Figure 7.

Temperature dependence of CO2 efflux from the stems (inline image) of two 5-year-old B. pubescens saplings (sapling ‘A’: d = 7.6 mm; sapling ‘B’: d = 7.8 mm) grown in a climate chamber at a temperature regime between 21 and 27 °C from 10 June to 6 July 2002. Correlation coefficients (r) of 0.74 and 0.8 indicate that CO2 efflux was primarily due to respiration of stem parenchyma.

If sap flow contributes to the oxygen supply of the sapwood, one may question whether sap flow also affects radial oxygen transport between stem and atmosphere? Respiratory O2 consumption of wood parenchyma interferes with physico-chemically driven radial O2 transport. To examine this implication, rates of O2 exchange between stem and cuvette atmosphere, and of CO2 efflux can be related to each other on the basis of three assumptions: (1) the respiratory quotient (RQ) is close to unity. This is the case when glucose is the primary substrate respired; (2) no O2 deficit occurs in the cuvette atmosphere, that is, the O2 concentration in the cuvette reaches 100% air saturation; (3) no sap flow is measured at that time. The second and third assumptions were fulfilled during night-time on day 8 of the experiment (Fig. 4d). At that time, the stem respired at a near constant rate of 1.35 µmol CO2 m−2 s−1. It was therefore concluded that O2 consumed by respiration originated from the stem at a rate equivalent to the rate of respiratory CO2 production (Fig. 6). This equivalence was used as the reference level for respiration-dependent oxygen consumption which partially contributed to the O2 deficit usually measured in the cuvette atmosphere. The remaining portion of the O2 deficit could then be attributed to radial O2 influx from the cuvette atmosphere into the stem. With regard to the alternating decrease and increase of oxygen concentration in the cuvette atmosphere between day and night, respectively, the stoichiometrical approach provided an estimate of radial O2 flux in the stem during the light and dark periods (Fig. 6). It revealed that nocturnal O2 deficits in the cuvette atmosphere, namely the difference between CO2 efflux and O2 exchange rate (bold line), were considerably smaller than by day. This is due to both temperature-dependent reduction of respiratory oxygen consumption, and higher O2 efflux from the stem. At unimpeded sap flow (days 4–6) the mean O2 efflux rate was + 0.33 µmol O2 m−2 s−1, and thus, the O2 deficit in the cuvette atmosphere was smaller than expected from respiration during the day. The gradual reduction of sap flow caused a change from O2 efflux into an increasing O2 influx into the stem (day 7, −0.31 µmol O2 m−2 s−1; day 8, −0.54 µmol O2 m−2 s−1). Therefore, the O2 deficit in the cuvette atmosphere increased above that expected from respiration. When sap flow recommenced, reversion from oxygen sink to source towards a rate similar to that measured at normal sap flow took place. Thus, according to the stoichiometrical relations applied, xylem sap flow also affected radial oxygen exchange between stem and atmosphere in B. pubescens saplings. At rates from 1 to 1.5 mmol H2O m−2 LA s−1 source to sink relations were in a range of up to 0.9 µmol O2 m−2 stem surface s−1.


Reversible interruption of xylem sap flow by oxygen depletion in the roots

At conditions of unimpeded sap flow the leaf area-specific flow rates of B. pubescens saplings (1–1.5 mmol H2O m−2 s−1) were highly consistent with flow rates derived from mature B. pendula trees (1.2 mmol H2O m−2 s−1), naturally grown in a mixed birch-pine forest (Backes 1996). Similar rates of FLA of 1.8–3.6 mmol H2O m−2 s−1 at maximum on sunny days were also reported for mature B. pendula trees (Ladefoged 1963).

In this study xylem sap flow was reversibly reduced, and even stopped, by a combined treatment of warming and oxygen depletion in the root space. Three experimental findings are of interest: (1) warming up to 32 °C caused an initial increase in FLA by 34% compared with previous rates at 17 °C, followed by an increase of stomatal conductance and photosynthesis; (2) restriction of sap flow occurred first after 36.5 h from the onset of oxygen depletion; and (3) sap flow recommenced within 20 min after re-aeration of the roots.

Because the apoplastic path of radial water transport across the root cortex towards the xylem is interrupted or at least drastically reduced at the endo- and often at the exodermis, all water molecules must pass cell membranes. Thus, the modes of water movement across membranes are of major importance for water uptake by the roots (Steudle 1997, 2000a). Aquaporins contribute considerably to the hydraulic conductivity of plant cell membranes (Maurel 1997; Steudle 1997; Schäffner 1998; Tyerman et al. 1999). Here, water supply to the shoot is under metabolic control, and becomes affected by temperature and oxygen supply. Regulation of radial water flow by aquaporins is based on opening or closing of existing aquaporins or by variation of their density in membranes (Steudle 2000b). The rapid recommencement of sap flow in B. pubescens saplings within 20 min after re-aeration points to a reversible metabolic regulation of transmembrane water flow rather than an increase in aquaporine density.

Oxygen deficiency not only affects radial water transport by a changed activity of aquaporins – lack of phosphorylation de-activates water channels (Johansson et al. 1996; Kjellbom et al. 1999) – but also reduces energy-dependent active ion pumping. This affects the passage of water by changing the osmotic gradient across plasma membranes. In primary roots of Zea mays that were grown hydroponically anoxia caused a decrease of the ATP/ADP ratio by 64% after 15 h of deoxygenation (Birner & Steudle 1993). These authors stated that under anaerobiosis the reduction in energy charge of the root tissues causes a successive switching off of ion pumps located at the xylem and at the cortical plasmalemma, respectively.

The experimental procedure of oxygen depletion in the root space applied in this study provides some evidence that radial movement of water towards the xylem in roots of B. pubescens is primarily under metabolic control. The increase in temperature from 17 to 32 °C caused an overall increase in root metabolism including those processes being involved in water transport as long as oxygen availability did not limit respiratory ATP supply. The gradual depletion of oxygen in the root space must have affected root respiration and thus, the energy charge of the cells most likely decreased as well. Due to a low energy charge both the de-activation of aquaporins, and loss of active ion pumping are probably the key-processes for the gradual interruption of xylem sap flow. The fact that about 90% of the leaves of either tree could not survive the water-stress conditions induced by root hypoxia may prove that interruption of sap flow was complete. It may indicate further that water transport bypassing the root endodermis was negligible. Finally, the rapid and full recommencement of sap flow after re-aeration of the roots within 20 min, and the resulting stomatal opening of remaining leaves also provides evidence for a metabolically induced interruption of water transport which was neutralized when oxygenation set in.

The role of xylem sap flow for the oxygen status of the sapwood

In June, the sapwood of B. pubescens saplings showed a high oxygen status during unimpeded sap flow. At standard temperature and pressure (T = 20 °C, P = 100 kPa) daily maxima of [O2] of up to 245 µmol O2 L−1 occurred. Such values are in line with those reported for different tree species when the growing season begins (Table 1). Even the daily minimum of [O2] was high (about 210 µmol O2 L−1), whereas in spring, in mature B. pendula the minimum value dropped to 70 µmol O2 L−1 (Gansert et al. 2001). Similar minima were also reported for Picea abies, Quercus robur or Acer platanoides during the growing season in July and August (Table 1). Although the diurnal span of concentration of dissolved oxygen in young B. pubescens was less distinct than in mature B. pendula, the daily course with an early morning maximum and minimum in the late afternoon was basically the same. However, due to constant temperature conditions in the climate chamber, the nocturnal rise in [O2], typically observed under natural conditions, was missing.

Table 1.  Concentrations of molecular oxygen (min – max) dissolved in the sapwood of different tree species cited in literature
Tree speciesd.b.h.
Tree height
Tree age
(µmol L−1)
  1. d.b.h., diameter at breast height. For compatibility, the data presented by the authors were corrected for an aqueous solution saturated with air at T = 20 °C and atmospheric pressure (285 µmol O2 L−1).

Picea abiesn.s.15–25n.s.Mar – Jan 27 (Aug)-285 (Jan)Eklund (1990)
P. abies15–1813–1530Mar – Nov 41 (Jul)−285 (Mar)Eklund (2000)
Quercus robur1510n.s.May – Sept 68 (Jul)−231 (May)Eklund (1993)
Acer platanoides1510n.s.May – Sept 81 (Jul)−217 (Sept)Eklund (1993)
Betula pendula10 8n.s.Mar – Apr 71–171Gansert et al. (2001)
B. pendula28 8n.s.Aug – Oct 65–142Gansert et al. (2001)
Laurus nobilis 7 (stem base)n.s.21Dec172–258Del Hierro et al. (2002)
Olea europaea  1.5–1.8 4Apr 28–100Mancuso & Marras (2003)

The gradual reduction of sap flow provided a quantitative assessment of its effect on oxygen concentrations and flow rates in the sapwood of B. pubescens saplings. Because the day/night temperature regime variation was nearly constant over the 3 d period from day 6–8 of the experiment, the decrease in daytime [O2] from 75% (188 µmol O2 L−1) at unimpeded sap flow to a minimum of 30% (77 µmol O2 L−1) when sap flow was interrupted could not be attributed to an increase in respiratory oxygen consumption. Rather, this provides evidence that sap flow accounts for some 60% of the oxygen concentration in the sapwood. Undoubtedly, sap flow rate, oxygen loading of xylem sap and oxygen withdrawal from it by oxidative respiration of living cells of the shoot, are factors which synergistically influence the extent of oxygen available from the aqueous transport path. The observed oxygen depletion down to 77 µmol O2 L−1 when sap flow ceased is comparable with the low value of 42 µmol O2 L−1 observed in B. pendula during flushing; namely when sap flow was negligible (Gansert et al. 2001). From wood anatomical studies of B. pendula and B. pubescens (Braun 1970; Grosser 1977; Schweingruber 1978) it can be assumed that, with the exception of multiseriate rays, the parenchymatous tissues of birch sapwood have little or no contact with the intercellular gas-space continuum. The drastic reduction of [O2] in B. pubescens caused by interruption of sap flow supports this assumption. Investigations on oxygen supply to the sapwood of Olea europaea saplings also indicated that about 80% of the oxygen concentration in the sapwood was delivered by xylem sap flow (Mancuso & Marras 2003). Interestingly, a similar pattern of gradual oxygen depletion in the sapwood of O. europaea was observed with the strongest hypoxia on the third day of root anaerobiosis.

However, radial gaseous diffusion of atmospheric oxygen through lenticels and the cambium into the sapwood cannot be left out of consideration. For example, a nocturnal increase of [O2] could be measured on days 7 and 8 of the experiment, irrespective of sap flow being reduced to half the rate of previous nights or even completely blocked (Figs 4d & 5). Under these conditions radial gaseous diffusion of oxygen into the sapwood accounted for 0.2–0.5 nmol O2 L−1 s−1. For comparison, participation of sap flow after re-aeration accounted for a specific oxygenation rate of 2 nmol O2 L−1 s−1 per mmol H2O m−2 LA s−1 transpired, which is one order of magnitude higher than the diffusion component of oxygen transport. Taking into account the anatomical features mentioned above, and experimental findings that both CO2 efflux from stems of different birch species such as B. pendula or B. ermanii (Levy et al. 1999; Gansert unpubl.), and O2 exchange across the stem surface (see below) are affected by sap flow, the cambium of birch can hardly be seen as an impervious sheath. Rather, it appears to allow the passage of gases preferentially through the intercellular gas-spaces of the rays (Larson 1994).

The oxygen relations in the sapwood of B. pubescens saplings presented here substantiate the importance of sap flow for oxygen supply to the sapwood as has been shown earlier in the field for mature B. pendula (Gansert et al. 2001). It also supports the concept of a dual transport system that supplies wood parenchyma of trees with oxygen via radial gas diffusion and axial flow of oxygen dissolved in the xylem sap as suggested earlier (Hook et al. 1972; Gansert et al. 2001). During daytime, the xylem sap flow appears to be the major path for transport of dissolved oxygen axially through the sapwood. During night-time, when sap flow approximates zero, the gaseous path for O2 transport, driven by diffusion gradients radially through intercellular gas-spaces prevails (Gansert et al. 2001). The water-saturated cell walls of the apoplast function as the primary absorbing matrix for gaseous oxygen supplied by radial gaseous diffusion. Thus, xylem sap can be enriched with oxygen using the overall surface of the above-ground woody cormus as the absorbing area. On the other hand, xylary diffusion namely the axial transport of dissolved oxygen in the xylem sap of tracheids and vessels only by diffusion, can hardly be considered as an efficient pathway for oxygen supply from the roots to the above-ground sapwood when there is no sap flow at night. According to Fick's second law, movement of oxygen in water over a distance of 1 m driven by diffusion alone takes several years, depending on the concentration gradient, so that this pathway can only be effective over short distances smaller than 100 µm (von Willert et al. 1995).

Respiration of wood parenchyma represents an intrinsic biogenous factor which complicates quantitative estimates of diurnal oxygen supply to the sapwood from measurements of the local [O2] inside the stem. Respiratory oxygen consumption of parenchymatous tissues of the wood strongly depends on their quantity and actual metabolic conditions with respect to growth, storage or secretion activities. During the day, respiratory oxygen consumption exponentially increases with a rise in temperature, and the reverse pattern is shown at night. Thus, O2 consumption approximates the daily minimum before dawn. The oxygen status of the sapwood is therefore a resultant of oxygen consumption (O2 sink) by respiration of wood parenchyma as a function of temperature and metabolic condition, and oxygen supply (O2 source) via the aqueous and gaseous pathways. The contribution of the aqueous path in the form of xylem sap flow depends above all on canopy transpiration which is a function of light incident on the leaves and the leaf-to-air water vapour pressure deficit (VPD). Gaseous diffusion is driven by concentration, temperature and pressure gradients. Therefore, the diurnal variation of the source–sink relation of oxygen in the sapwood is differentially affected by abiotic variables. The stoichiometrical approach on the relation between CO2 efflux and O2 consumption applied here, provided an estimate of oxygen exchange across the stem surface which seems to be physico-chemically coupled with xylem sap flow. Principally, in the absence of bark photosynthesis the stem was never a net source for oxygen. However, enhanced nocturnal O2 efflux contributed to smaller O2 deficits in the cuvette atmosphere than during daytime. This observation is in line with a nocturnal increase in [O2] of xylem sap measured under field conditions in spring and summer (Gansert et al. 2001). Conversely, oxygen deficits higher than those solely due to respiratory oxygen consumption mark the existence of an increasing endogenous oxygen sink as was measured in the sapwood when sap flow was interrupted. These results may indicate that xylem sap flow not only affects the oxygen status of the sapwood itself, but also has an effect on radial oxygen transport across the stem surface. Exposure of mature stems of B. pendula to a hypoxic atmosphere and prevention of bark photosynthesis caused an oxygen efflux from the stem which was coupled with xylem sap flow (Gansert & Burgdorf, unpublished). Hence, the cambium of birch species (B. pubescens, B. pendula, and B. ermanii) is not an impervious sheath, but allows radial fluxes of O2 and CO2 into and out of the stem through the intercellular gas-space continuum.

Although Bailey (1913) clearly recognized the importance of the transpiration stream as an aqueous pathway for gases 90 years ago, the quantitative analysis of the role of xylem sap flow for endogenous source-sink relations of O2 and CO2, and gas exchange between woody plant parts and atmosphere has just begun. The causal understanding of these relations will depend on the success in quantitative differentiation between biogenous processes such as respiration and bark photosynthesis, and physico-chemical processes such as diffusion, solubility and dissociation or flow in the gaseous and aqueous phase in woody parts of arborescent plants.


The author is indebted to Professor Dr R. Lösch for his encouragement, scientific and financial support. Many thanks are also due to the precision mechanical engineers Mr W. Seidel and Mr M. Meiercord for their co-operation in constructing the Plexiglass cuvettes. I am also grateful to A. Stöhr for his support in the SHB technique, to K. Kiefer for the schematic drawings, and to Dr J. R. Yates for his linguistic support.