The temperature dependence of shoot hydraulic resistance: implications for stomatal behaviour and hydraulic limitation


Correspondence: Steven Matzner. Fax: + 1 605 274 4718; e-mail:


Recent soil pressurization experiments have shown that stomatal closure in response to high leaf–air humidity gradients can be explained by direct feedback from leaf water potential. The more complex temperature-by-humidity interactive effects on stomatal conductance have not yet been explained fully. Measurements of the change in shoot conductance with temperature were made on Phaseolus vulgaris (common bean) to test whether temperature-induced changes in the liquid-phase transport capacity could explain these temperature- by-humidity effects. In addition, shoot hydraulic resistances were partitioned within the stem and leaves to determine whether or not leaves exhibit a greater resistance. Changes in hydraulic conductance were calculated based on an Ohm’s law analogy. Whole-plant gas exchange was used to determine steady- state transpiration rates. A combination of in situ psychrometer measurements, Scholander pressure chamber measurements and psychrometric measurements of leaf punches was used to determine water potential differences within the shoot. Hydraulic conductance for each portion of the pathway was estimated as the total flow divided by the water potential difference. Temperature-induced changes in stomatal conductance were correlated linearly with temperature-induced changes in hydraulic conductance. The magnitude of the temperature-induced changes in whole-plant hydraulic conductance was sufficient to account for the interactive effects of temperature and humidity on stomatal conductance.




Although chemical signalling between roots and shoots has been shown to influence stomatal opening (Dodd et al. 1996; Jia et al. 1996; Schurr & Schulze 1996; Tardieu 1996; Tardieu et al. 1996), for rapid diurnal responses to changing humidity gradients, the role of leaf water potential as a negative feedback mechanism regulating stomatal aperture has received considerable support recently (Saliendra et al. 1995; Fuchs & Livingston 1996; Comstock & Mencuccini 1998). These studies used a root pressurization technique to manipulate shoot water status independently and have shown that leaf water status could explain stomatal closure due to soil drying, leaf-to-air vapour pressure differences (but see Gollan et al. 1986; Schurr et al. 1992) and root chilling (Mencuccini & Comstock, unpublished). While leaf water potential could account for the stomatal responses to leaf-to-air vapour pressure differences (Δω) in these studies, the more complex temperature-by-humidity responses of stomata have not yet been explained fully (Lösch & Tenhunen 1981; Ball et al. 1987; Aphalo & Jarvis 1991; Mott & Parkhurst 1991). If temperature is raised without raising absolute humidity (i.e. increasing Δω), there is stomatal closure, but if Δω is kept constant and temperature is raised, there is actually an opening response. The mechanistic nature of this response and exactly what is sensed to mediate it remains an open controversy. Ball et al. (1987) argued that this interaction could be simplified empirically by expressing humidity as relative humidity and that plants may actually sense relative humidity; however, they provided only correlative evidence that this might be the case. In contrast, Aphalo & Jarvis (1991) did not find that relative humidity correlated with stomatal response, but found a separate response to both humidity and temperature. Mott & Parkhurst (1991) attempted to ascertain whether stomata respond to Δω or to the rate of water loss from the leaf. They concluded that plants (at constant temperature in their experiments) were sensing transpiration rather than Δω and that a simple hydraulic model could explain stomatal responses to Δω but could not explain the temperature and/or temperature-by-humidity interactive effects. We argue that when water relations are viewed at the whole-plant level and the extremely important effects of temperature on whole-plant water transport capacity are taken into account, the complex temperature-by-humidity responses of stomata are consistent with a simple hydraulic model

A simple hydraulic model based on an Ohm’s law analogy can be useful in describing whole-plant water transport capacity. The water potential difference between the soil and leaves (ΔΨ) is related to transpirational flux (E) and hydraulic conductance (K) according to the Ohm’s law analogy:


Despite several caveats (osmotic and capacitance effects, for example), Eqn 1 is a useful theoretical orientation for the measurement of hydraulic conductance in both roots and stems (Passioura 1988; Saliendra et al. 1995; Mencuccini & Comstock 1999). In this study, we used the Ohm’s law analogy to assess the effect of temperature on K. We measured E at steady state and the water potential difference between the soil (kept saturated and assumed to be close to 0) and the leaves and used Eqn 1 to calculate K at different temperatures. As seen in Fig. 1,E depends on gs and humidity, but temperature affects K directly, which in turn affects ΔΨ. Changes in K due to temperature (at a constant Δω) will cause changes in ΔΨ. Stomatal regulation is balanced between opening signals favourable to photosynthesis and the limitation imposed by feedback from a depressed ΨL. ΨL is postulated to have only a closing influence during stress, but never an opening influence at high ΨL.

Figure 1.

A concept model of stomatal regulation as a balance between opening signals from conditions favourable to photosynthesis (+ arrows on stomatal conductance) and the limitation imposed by feedback from depressed leaf water potential (0 or – arrow). The parameters in the shaded region represent Eqn. 1. The postulated behaviour of this model is that stomata open in response to positive signals and are little influenced by water potential until a critical threshold of Ψleaf is reached. Parameters include Δω, ci, K, E, A, gs, Ψleaf and Ψsoil.

If K is decreased (by lowering temperature at a constant Δω), then ΨL will decrease and exert a closing influence on gs. If K is increased (by raising temperature at a constant Δω), then ΨL may become more positive, relieving any closing influence being exerted on gs.

One mechanism for the effect of temperature on hydraulic conductance is simply the temperature-dependent changes in the physical properties of water. The long-distance flow of water in plants depends primarily on the flow in xylem vessels: the flow through a single vessel can be described by the Hagen–Poiseulle law (Tyree & Ewers 1991; for flow through multiple vessels, see Roderick & Berry 2001):


where Lp is the hydraulic conductance of one capillary (vessel), r is its radius, and ρ and η are the density and viscosity of water, respectively. While this equation is usually used to emphasize the fourth power dependency of flow on vessel radius, the temperature dependency of ρ/η is actually quite strong: the Q10 for the effect of ρ/η on flow is approximately 1·25 between 20 and 30 °C and 1·24 between 20 and 40 °C (Lide 1994) and could exert a significant influence on hydraulic conductance.

In addition to the effect of temperature on the physical properties of water, water flow through extra-xylary portions of the pathway that include movement through the symplast may involve temperature-dependent changes in membrane permeability (Boyer 1985; Passioura 1988; Canny 1990). These membrane permeabililty changes may have an effect beyond that of the changes in the physical properties of water. The interactive effects of temperature and humidity on gs and E may be the result of changes in both membrane permeability and water viscosity (Boyer 1985; Passioura 1988; Canny 1990).

By accounting for the temperature-dependent effects on the physical properties of water and in membrane permeability changes in the symplastic portions of the hydraulic pathways, our overall objective was to determine if the complex temperature-by-humidity effects were still consistent with a simple Ohm’s law analogy. Firstly, we compared the temperature-induced changes in KSoil–lamina, gs and E to determine if the magnitude of the temperature-dependent change in KSoil–lamina was large enough to account for changes in gs and E. Secondly, we compared the relative resistance of RLeaf vein–lamina with RStem–lamina to determine if there was a higher resistance in the RLeaf vein–lamina pathway. This may indicate a greater resistance to water flow through extra-xylary portions of the pathway that include movement through the symplast.

Materials and methods

Experimental material

Two common bean cultivars of the ‘pinto bean’ type (Phaseolus vulgaris L., Fabaceae) were grown in a greenhouse at the Boyce Thompson Institute (Ithaca, NY, USA). The two cultivars, G4523 and Othello, represent the Andean South American and Middle American centres of origin, respectively. G4523 is of the Nueva Grenada land race and is generally grown under rain-fed conditions, while Othello is of the Durango race and is generally grown under irrigated conditions (Mencuccini & Comstock 1999). The two cultivars differ markedly in leaf size and growth form.

Plant propagation

Plants were germinated between December 1998 and August 1999 under shade cloth in 11·4 L pots and were then moved under lights. The soil mixture and greenhouse conditions were similar to those described in Mencuccini & Comstock (1999). The soil mixture consisted of fritted clay (Turface, Profile Products LLC, Buffalo Grove, IL, USA), silica sand, pasteurized soil, vermiculite and peat (6 : 2 : 2 : 2 : 1, by volume) amended with dolomitic lime, gypsum, superphosphate and Micromax (The Scotts Co., Marysville, OH, USA). Plants were watered twice daily and were fertilized periodically throughout the experiment. The environmental conditions within the greenhouses were controlled and monitored continuously. Plants received supplemental lighting using a combination of Na vapour and metal halide lamps. The photoperiod from combined artificial and natural lighting was 14 h, with a total irradiance (400–700 nm) of 44 mol m2 day−1. Day/night time conditions were approximately 28/20 °C, 40/80% relative humidity and 375/390 µmol mol−1 CO2. A set of several rotating fans stirred the air continuously during growth and was strong enough to cause leaf fluttering.

Experimental protocol

Experiments were conducted between January and December 1998. Measurements were made on plants between 27 and 47 d old. Day one of the experiments for each plant consisted of loading the plant into the root pressure chamber and the gas exchange cuvette and installing leaf thermocouples and in situ leaf psychrometers (see Gas exchange and root pressurization chamber and Psychrometric measurements sections). Leaf thermocouples were taped to the underside of leaves using ‘breathable’ medical tape. In addition, an initial estimate of leaf area was made (see Leaf area measurements section). Generally, during days two and three, measurements were taken at either 20 or 35 °C; on the fourth day, measurements were taken at 28 °C. Within each day, a range of transpiration rates was generated by varying Δω, light levels and root pressurization treatments. Maximum light levels at the canopy level in the cuvette were approximately 1·5 mmol m−2 s−1. Reduced light treatments were accomplished using mesh screens over the top of the cuvette, which reduced maximum light levels by about 66%.

Leaf area measurements

On day one, an initial estimate of leaf area was made based on the emperical relationship between leaf length × width and measured leaf area for each cultivar. At the end of the experiment, leaves were harvested and final leaf area was measured using a leaf area meter (LI-3200; Li-Cor Inc., Lincoln, NE, USA). Leaf area on days between the initial leaf area estimate and the final leaf area measurement was estimated based on a linear regression of the change in leaf area over time for each plant.

Gas exchange and root pressurization chamber

Gas exchange measurements were carried out with a single-pass system and a whole-plant cuvette, as described in Comstock & Mencuccini (1998) except with the addition of radiators and a frontal access panel. The radiators were installed to increase the speed of leaf temperature control and a frontal access panel, equipped with Teflon cuffs, was installed to minimize exchange with the outside environment and to facilitate plant sampling. The root system of the plants was sealed in a root pressure chamber (described in Comstock & Mencuccini 1998).

Adaxial and abaxial leaf boundary layer conductance was determined at three heights inside the cuvette, using leaf replicas made out of filter paper sealed on one side with sticky Teflon film. Adaxial and abaxial boundary layer conductance averaged 0·8 and 1·2 mol m−2 s−1, respectively. The stomatal ratio was assumed to be 0·3 for both cultivars, based on previous work (Comstock & Ehleringer 1993). Leaf temperature was calculated as the average of seven type-E thermocouples inserted into leaves in different parts of the canopy. Gas exchange calculations followed von Caemmerer & Farquhar (1981) and Comstock & Ehleringer (1993). Gas exchange parameters were logged automatically every 2 min during the course of the experiments; in addition, manually logged points were taken under steady-state conditions.

Psychrometric measurements

In situ leaf psychrometers (leaf hygrometers; Plant Water Status Instruments Inc., Guelph, Ont., Canada) were installed on three to four exposed leaves. A piece of sticky Teflon (about 2 cm2) with a circular hole for the psychrometer was applied to the target leaves. Cellite dissolved in distilled water was rubbed gently on the leaf surface exposed by the circular hole to increase the speed of water vapour equilibration between internal leaf spaces and the psychrometer chamber. The in situ psychrometers were positioned on the exposed leaf surface and a vapour-tight seal was made around the psychrometric chamber using dental compound (Reprosill Dental Supply Intl Inc., Milford, DE, USA). The psychrometers were then covered with foam insulation and two to three layers of aluminium foil to minimize thermal gradients.

Wescor C-52 psychrometers (Wescor Inc., Logan, UT, USA) were calibrated and leaf punch ΨL values were measured at 25 °C using a temperature-controlled box. Both the in situ psychrometers and the C52 psychrometers were controlled with a CR7 datalogger (Campbell Scientific Inc., Logan, UT, USA). Since leaves were sampled at 21, 28 and 35C, but psychrometric measurements were taken at 25C, we tested whether these temperature changes affected ΨL. Multiple leaf punches were taken from single leaves and ΨL was measured at either 21 or 35 °C (using psychrometers calibrated at those temperatures). Pressure chamber measurements (Soil Moisture Inc., Corvalis, OR, USA) were also taken on the same leaves that had been sampled with the leaf punch. There were no significant differences between leaf punch water potentials measured at either 21 or 35 °C. In addition, the psychometric measurements were not significantly different from the pressure chamber values.

Measurements of hydraulic conductance

Hydraulic conductance is defined as the flow rate of liquid water divided by the pressure difference across a defined flow path (Saliendra et al. 1995). We expressed hydraulic conductance as the leaf-specific conductance (K), which is the hydraulic conductance divided by the leaf area supplied by the path. Measurements of conductance were performed by both direct and indirect methods.

Direct measurements of total root conductance (KSoil–stem) were carried out using a pressure flux technique (Saliendra et al. 1995). After gas exchange measurements were completed, the shoot was detopped, leaving 2–4 cm of stem, which was attached to tubing connected to a reservoir on an analytical balance. The root pressure chamber was pressurized to at least three different pressures ranging between 0 and 0·75 MPa. Flow rates at each pressure were measured after a stabilization period. Flow rates per canopy leaf area were plotted as a function of applied pressure and KSoil–stem was estimated as the slope of the linear regression. Measurements were automated by interfacing the balance with a computer. For plants in which we did not directly measure KSoil–stem, this parameter was estimated based on relationships between KSoil–stem and leaf area, stem biomass and plant age.

Direct measurements of KStem–petiole were performed using a cylindrical vacuum chamber as described in Saliendra et al. (1995). Leaves were cut from the shoot at the petiole and the remaining shoot was cut at the base of the stem, attached to tubing connected to a balance and placed inside the cylindrical chamber. The measurement of flow were carried out at vacuums between 0 and 0·02 MPa. The slope of the regression of flow versus pressure gave the KStem–petiole value. The measurements of KStem–petiole were taken in a growth chamber at the Boyce Thompson Institute (Ithaca, NY, USA) at three temperatures between 15 and 35 °C. Preliminary experiments indicated that KStem–petiole did not change over a period of 12–18 h.

Indirect estimates of the total leaf-specific conductance of the soil to the lamina (KSoil–lamina) and of the soil to the petiole (KSoil–leaf vein) were calculated based on the following equations, assuming Ψsoil was equal to zero:


Estimates of KLeaf vein–lamina were obtained as the difference between resistances in series. Values of conductance are the inverse of the resistance:


Estimates of KStem–petiole were obtained as the difference between resistances in series:



Differences between the ‘composite’Q10 values (see Results and discussion section for explaination) for KSoil–lamina, E and gs and for Q10 values for KStem–lamina, KStem–leaf vein and KLeaf vein–lamina were compared using an analysis of variance model with cultivar as one factor. Q10 values were not transformed because the distribution was not significantly different from a normal distribution. Differences in measures of shoot resistance (%RStem–lamina, %RStem–leaf vein and %RLeaf vein–lamina) and measurements of KSoil–lamina, ΨL, A, E, gs and ci/ca were tested using repeated-measures analysis of variance models. Measures on individual plants on successive days at different temperatures (20, 28 and 35 °C) were the repeated factor. Differences between cultivars were tested as the between-subject effect.

Results and discussion

In situ psychrometer measurements and measurements of leaf punches using Wescor C52 sample chambers were compared with bagged and unbagged leaf pressure chamber measurements, respectively (Fig. 2). All values fell closely along a 1 : 1 line and there was a significant correlation between pressure chamber measurements and psychrometer measurements (r = 0·957, P < 0·001). In situ or bagged measurements represent a non-transpiring leaf water potential and were used to calculate the water potential difference between the soil and the leaf vein. Leaf punch and unbagged measurements represent a transpiring leaf water potential and were used to calculate the water potential difference between the soil and the leaf lamina. In situ or bagged leaf water potential values were significantly different from leaf punch water potentials (P < 0·001). In situ or bagged leaf water potential values ranged between −0·1 and −0·6 MPa, while leaf punch and unbagged values were mostly between −0·45 and −0·9 MPa. The range of water potentials was achieved by varying humidity, temperature and light levels.

Figure 2.

The relationship between in situ or leaf punch psychrometer water potential measurements and bagged and unbagged Scholander pressure chamber measurements for leaves of P. vulgaris. bsl00000 = in situ psychrometer/bagged leaf comparison; s = leaf punch/unbagged leaf comparisons (one pressure chamber value compared with an average of 2–3 leaf punches from the same leaf).

Measurements of E and ΔΨ were used to calculate conductances based on a simple Ohm’s Law analogy (see Eqn 2). Hydraulic conductance was calculated as the slope of the relationship between the transpiration rate and the water potential difference (Ψsoil − Ψlamina or Ψsoil − Ψleaf vein). Comparing the differences in the slopes of the relationship assessed the effect of temperature on flow. One assumption in the calculation of conductance in this study is that the volume of the flow path was the same at the different temperatures (i.e. xylem cavitation did not occur). Plants were kept well-watered during the experiment, but it is possible that at high Δω, stem water potentials may have decreased enough to cause xylem cavitation. The strongest evidence that stem cavitation was not occurring in this experiment can be seen in Table 1. The Q10 values for KStem–leaf vein and KStem–petiole were nearly identical. A cylindrical vacuum chamber was used to measure KStem–petiole and it is unlikely that xylem cavitation occurred for these measurements (see Materials and methods section). Since KStem–leaf vein represents a similar portion of the hydraulic pathway, it appears that xylem cavitation was not occurring.

Table 1. Q10 values (average ± SE) for the viscosity of water (predicted value based on the Hagen–Poiseulle law) and leaf-specific conductances (K) for stem to lamina (n = 10), stem to leaf vein (n = 10), stem to petiole (n = 9) [Measurements made using a cylindrical vacuum chamber (see Materials and methods)] and leaf vein to lamina (n = 10). Composite Q10 values (not true Q10 values because root temperature was not changing) for transpiration (E, n= 10), stomatal conductance (gs, n= 10) and leaf- specific conductance (soil to lamina, n= 10) are indicated by an asterisk. Most Q10 or composite Q10 values were calculated based on six G4523 and four Othello individuals. The predicted Q10 for water was calculated from the Hagen–Poiseulle law and represents the temperature dependency of the density and viscosity of water.
 Q10 value+ SE– SE
Viscosity of water1·25  
KStem–leaf vein1·540·090·09
KLeaf vein–lamina1·800·090·08

Figure 3 shows the average relationships between E and the soil-to-leaf vein and soil-to-lamina water potential differences for the G4523 cultivar at both 20 and 35 °C. Different light, humidity and soil pressurization treatments were used to attain the range of transpiration rates and water potential differences. The highest values for E and the water potential differences are from root pressurization treatments (where Ψsoil ranged between +0·32 and +0·51 MPa). Generally, pressures of +0·51 MPa were applied, but less pressure was used (particularly at 20 °C) if in situ psychrometer measurements at the highest Δω indicated a potential for ‘flooding’ the psychrometer. The relationships between E and the soil-to-petiole and soil-to-lamina water potential differences for Othello (not shown) were similar to those for G4523, except that transpiration rates and slopes tended to be greater. Although in theory the slopes should pass through zero, this did not always occur. For the calculation of individual conductances, the slopes were forced through zero since the intercepts from individual plots were not significantly different from zero and were not different between cultivars, temperatures and either in situ or leaf punch psychrometers.

Figure 3.

The relationships between transpiration and either the soil-to-leaf vein (bsl00000) or soil-to-lamina (s) water potential differences for the G4523 cultivar of P. vulgaris at either 20 (a) or 35 (b) °C. Each point represents the average (± SE) of at least six individuals. The range of transpiration rates was obtained by two low-light levels, two Δω levels and a root pressurization treatment.

Increasing shoot temperature caused a significant decrease in shoot pathway resistances (%RStem–lamina, %RStem–leaf vein and %RLeaf vein–lamina) (P < 0·0001, Fig. 4, open symbols). The overall effect can be observed by the decrease in %RStem–lamina from 78 to 64% with increasing temperature. In contrast, %RSoil–stem increased with temperature: this is because root temperature did not change in this experiment. Since shoot temperature was changing and %RStem–lamina decreased with temperature, %RSoil–stem, which is a percentage of the total pathway resistance, increased necessarily. The responses of both %RStem–leaf vein and %RLeaf vein–lamina were more variable. Much of this variation occurred at the 28 °C point. It should be noted that the measurements at 28 °C were based on fewer individuals and were carried out on the final day of the experiment for each plant measured. The lower replication and the cumulative effects of 3 d in the plant cuvette may have influenced these values and therefore we tended to put more weight on the 20 and 35 °C points. The overall effect was a decline from 39 and 39% at 20 °C to 34 and 31% at 35 °C, for %RStem–leaf vein and %RLeaf vein–lamina, respectively.

Figure 4.

The change in resistance to liquid-phase water flow as a function of shoot temperature. Resistance is divided into four paths: soil to stem (%RSoil–stem, bsl00000), stem to lamina (%RStem–lamina, ♦), stem to leaf vein (%RStem–leaf vein, ○) and leaf vein to lamina %RLeaf vein–lamina, bsl00084). Note that the actual values of RSoil–stem were unchanging, but %RSoil–stem was changing because of a change in total resistance. Closed symbols represent values adjusted to 22.7C by linear regression.

There was a significantly higher resistance for %RStem–lamina compared with %RStem–leaf vein and %RLeaf vein–lamina (P < 0·001) with no difference between %RStem–leaf vein and %RLeaf vein–lamina. In addition, a significant (temperature × pathway) interactive effect (P < 0·004) and also a significant temperature–cultivar–pathway interaction in the response of the three measures of shoot resistance (%RStem–lamina, %RStem–leaf vein and %RLeaf vein–lamina) to temperature (P < 0·03) was observed. The temperature–pathway interaction appears to be due to the higher values of %RStem–leaf vein at 28 °C since the response of %RLeaf vein–lamina was not significantly different from %RStem–lamina. The temperature–cultivar–pathway interaction appears to be because of a higher value of %RLeaf vein–lamina at 20 °C (42%, SE ± 1) for the Othello cultivar compared with G4523 (37%, SE ± 3) and a lower %RLeaf vein–lamina for Othello at 35 °C (27%, SE ± 2) compared with G4523 (33%, SE ± 4). In general, G4523 did tend to have higher values of shoot resistance (except for the 20 °C %RLeaf vein–lamina point) but because of the relatively large variation and low replication (particularly for Othello), there was not a significant cultivar × temperature difference.

In order to compare shoot and root resistances, shoot resistances were adjusted to match root temperatures (Fig. 4, closed symbols). Since root temperature averaged 22·7 °C (root temperature was kept close to ambient room temperature), %R values for the shoot were adjusted to 22·7 °C using linear regression. Because of the greater variation and lower confidence in the 28 °C points for %RStem–leaf vein and %RLeaf vein–lamina, only the 20 and 35 °C points were used to estimate resistance values at 22·7 °C. At 22·7 °C, %RSoil–stem was approximately 25% and %RStem–lamina was approximately 75%. Values of %RStem–leaf vein and %RLeaf vein–lamina were nearly identical at 22·7 °C: both were approximately 38%.

The partitioning of the resistance into various components indicates that the majority of the resistance was in the shoot (approximately 75%). Using the above estimates for the partitioning between the roots, stem and leaves, the ratio of resistances would be 1 : 1·6 : 1·5, which differs from the calculated ratios of 1 : 0·5 : 0·75 given by Nobel (1999) for sunflower, bean and tomato. The main difference is a lower root resistance and a substantial axial xylem resistance. The assumption is often made that the major resistance to water flow in plants is in the non-vascular or radial pathways (Passioura 1988; Frensch & Steudle 1989). This study found that nearly 40% of the total plant resistance was in the axial pathway. A previous study by Mencuccini & Comstock (1999) also found a substantial axial resistance for this species. One possible explanation for this much higher resistance is the fact that the xylem elements end at stem nodes for P. vulgaris. Added resistances at these nodes and in the petiole may be responsible for the higher Q10 values for KStem–petiole found in this study.

To test whether the magnitude of the effect of temperature on conductance was large enough to account for the changes in gs and E, composite Q10 values were calculated for K, gs and E (Table 1). These are not true Q10 values because they include the root path, which did not change in temperature. Comparisons of E, gs and KSoil–lamina composite Q10 values did not reveal significant differences. This indicates that the changes in K were large enough to account for the temperature-induced changes in E and gs. In addition, the composite Q10 value for KSoil–lamina was significantly different from 1·25 (P < 0·01, d.f. = 9, one-tailed T-test); 1·25 is the Q10 for the Hagen–Poiseuille law (based on the changing viscosity and density of water from 20 to 30 °C; see Eqn 2). This indicates a significant change in conductance beyond the effect of changes in the density and viscosity of water.

Table 1 also includes Q10 values for various parts of the plant hydraulic path, including KStem–lamina, KStem–leaf vein, KStem–petiole and KLeaf vein–lamina. Shoot conductance Q10 values were all significantly different from 1·25 (P < 0·005, d.f. = 8 or 9, one-tailed T-test). There were also differences between KStem–lamina, KStem–leaf vein and KLeaf vein–lamina (P < 0·05), with KLeaf vein–lamina significantly different from KStem–leaf vein (P < 0·02). KLeaf vein–lamina had the highest Q10 value, while KStem–lamina– which is made up of both the shoot xylem and the leaf extra-xylary component – had an intermediate Q10 value. Values of Q10 for KStem–leaf vein and KStem–petiole are for similar parts of the pathway but were calculated based on different methods (see Materials and methods section) and gave remarkably similar results that were not significantly different (1·52 versus 1·54). The high Q10 value for KLeaf vein–lamina may be associated with extra-xylary symplastic water movement (Boyer 1985; Passioura 1988; Canny 1990). The relatively high Q10 value for KStem–petiole and KStem–leaf vein may be because these measures included flow through the petioles, which may have added significantly to the resistance. In addition, the xylem elements in P. vulgaris end at stem nodes. Added resistances at these nodes and in the petiole may be responsible for the higher Q10 values for KStem–petiole and KStem–leaf vein found in this study.

Temperature can affect gs because of changes in both K (which affects ΔΨ) and Amax (which affects ci/ca) (Fig. 1). In order to assess these effects, values of gt and ΨLeaf were adjusted to Δω = 20 using linear regression; values of E, gs, ci/ca and Amax were calculated based on adjusted values of gt. A Δω of 20 mbars/mbar was chosen because it was close to the highest and lowest Δω values at 20 °C and 35 °C, respectively. Temperature-induced changes in K were correlated linearly with increases in gs for both Othello and G4523 (r = 0·85; Fig. 5, Table 2). There was a significant increase in KSoil–lamina (P < 0·007), E (0·014), gs (P < 0·009) and ci/ca (P < 0·001) with temperature, with no significant difference in the shape of the response of the two cultivars (Table 2, Fig. 5). Othello did have significantly higher values of KSoil–lamina (P < 0·010), E (P < 0·005), gs (P < 0·007) and ci/ca (P < 0·003) (Table 2, Fig. 5). Othello also tended to have higher Amax values compared with G4523, although this was only marginally significant (P < 0·06). Values of Amax were highest at 28 °C for Othello, but exhibited less response to temperature for G4523. Values of ΨL also tended to be more negative for Othello compared with G4523; however, this effect was not found to be significant. Temperature also had no significant effect on ΨL. With the exception of the Othello 35 °C point, there did appear to be a tendency for ΨL to be most negative at 20 °C. An increase in ΨL with temperature could occur if the increase in KSoil–lamina with temperature was greater than the increase in E. The higher Q10 values for KSoil–lamina (1·44) compared with the Q10 for E (1·32) tend to support this, although the difference was not found to be a significant (Table 1).

Figure 5.

The relationship between stomatal conductance and leaf-specific conductance for two cultivars of P. vulgaris, G4523 (bsl00000) and Othello (○). Values of stomatal conductance were adjusted to Δω = 20 mbars/mbar.

Table 2.  Values of transpiration (E), leaf-specific soil to lamina conductance (KSoil–lamina), leaf water potential (ΨL), stomatal conductance (gs), the ratio of leaf intercellular to ambient concentrations of CO2 (ci/ca) and maximum assimilation rates (Amax) for two cultivars of P. vulgaris at three temperatures . Values of E, gs, ci/ca, ΨL and Amax were adjusted to Δω = 20mbars/mbar for four Othello and six G4523 individuals
(mmol m−2 s−1)
(mmol m−2 s−1 MPa)
(mmol m−2 s−1)
(µmol m−2 s−1)

The complex interaction between the effects of temperature on K and Amax can be seen in Table 2. Increased temperature caused significant increases in K, whereas ΨL actually tended to decrease or only increase slightly. This relieved any closing effect on gs by ΨL and gs increased linearly with K (Fig. 5). Amax can influence gs through ci; however, the effects of K masked any effects on gs in this experiment. It should be noted that although gs and K are highly correlated, their relationship is not 1 : 1 – the increase in gs is less than the increase in K. For G4523, this makes sense because with gs increasing and Amax decreasing with temperature, the increasing ci/ca should tend to close stomata. Looking at Othello, the complex interaction between Amax and K can also be seen. Amax is highest at 28 °C, causing a lower ci/ca value despite the fact that gs is intermediate at 28 °C.


If the effects of temperature on whole-plant water transport are accounted for, a simple hydraulic model is capable of explaining the complex temperature × humidity responses of plants. The increase in shoot conductance with increasing temperature was correlated linearly with gs and the temperature-induced increases in shoot conductance were large enough to account for the increase in gs and E with increasing temperature at a constant Δω. Division of the hydraulic pathway into its component parts indicated large resistances within both the stem and the leaves. The large temperature-dependent changes in conductance that were found in the leaves may reflect temperature- dependent changes in permeability within the symplastic portions of the pathway.


We would like to thank the many people who contributed to this investigation. The work was supported by USDA grant #95-37100-1640.