Differences in hydraulic architecture account for near-isohydric and anisohydric behaviour of two field-grown Vitis vinifera L. cultivars during drought

Authors

  • H. R. SCHULTZ

    1. INRA/ENSA, UFR Viticulture, 2 Place Pierre Viala, 34060 Montpellier CEDEX, France and
    2. Institut für Weinbau und Rebenzüchtung, Fachgebiet Weinbau, Forschungsanstalt, D-65366 Geisenheim, Germany
    Search for more papers by this author

Hans R. Schultz, (permanent address) Institut für Weinbau und Rebenzüchtung, Fachgebiet Weinbau, Forschungsanstalt, D-65366 Geisenheim, Germany. E-mail: h.schultz@fa-gm.de

ABSTRACT

A comparative study on stomatal control under water deficit was conducted on grapevines of the cultivars Grenache, of Mediterranean origin, and Syrah of mesic origin, grown near Montpellier, France and Geisenheim, Germany. Syrah maintained similar maximum stomatal conductance (gmax) and maximum leaf photosynthesis (Amax) values than Grenache at lower predawn leaf water potentials, Ψleaf, throughout the season. The Ψleaf of Syrah decreased strongly during the day and was lower in stressed than in watered plants, showing anisohydric stomatal behaviour. In contrast, Grenache showed isohydric stomatal behaviour in which Ψleaf did not drop significantly below the minimum Ψleaf of watered plants. When g was plotted versus leaf specific hydraulic conductance, Kl, incorporating leaf transpiration rate and whole-plant water potential gradients, previous differences between varieties disappeared both on a seasonal and diurnal scale. This suggested that isohydric and anisohydric behaviour could be regulated by hydraulic conductance. Pressure-flow measurements on excised organs from plants not previously stressed revealed that Grenache had a two- to three-fold larger hydraulic conductance per unit path length (Kh) and a four- to six-fold larger leaf area specific conductivity (LSC) in leaf petioles than Syrah. Differences between internodes were only apparent for LSC and were much smaller. Cavitation detected as ultrasound acoustic emissions on air-dried shoots showed higher rates for Grenache than Syrah during the early phases of the dry-down. It is hypothesized that the differences in water-conducting capacity of stems and especially petioles may be at the origin of the near-isohydric and anisohydric behaviour of g.

Abbreviations
A

net photosynthetic rate

Amax

maximum net photosynthetic rate

AEs

ultrasonic acoustic emissions

g

stomatal conductance to water vapour

gmax

maximum stomatal conductance to water vapour

E

transpiration rate

Kh

hydraulic conductance per unit length, Kl, leaf specific conductance (whole plants)

LSC

leaf area specific conductivity (excised plant organs)

ΨPD

predawn leaf water potential

ΨM

midday leaf water potential

Ψleaf

leaf water potential

VPD

leaf-to-air vapour pressure deficit

LA

leaf area

LAI

leaf area index, namely the leaf area per total soil surface area

projected LAI

leaf area per soil surface area shadowed by the canopy

LPI

leaf plastochron index.

INTRODUCTION

Stomatal closure is the dominant factor limiting gas exchange during water deficit (e.g. reviews by Lösch & Schulze 1994; Davies & Gowing 1999). Stomata regulate transpiration so that sufficient carbon is gained while leaf water potential (Ψleaf) is prevented from becoming too negative and the break-down of the plants hydraulic system is avoided (Tyree & Sperry 1988; Jones & Sutherland 1991; Schultz & Matthews 1997). A decrease in stomatal conductance can correlate with a declining Ψleaf during soil drying, but can also occur before any measurable change in Ψleaf is recorded (Gollan, Turner & Schulze 1985; Trejo & Davies 1991). Differences in stomatal sensitivity during drought among cultivars or between species may serve to limit transpiration to compensate for differences in the vulnerability of xylem to cavitation (Jones & Sutherland 1991). Cultivated grapevine (Vitis vinifera L.) is a very heterogeneous species with an estimated 10–20 000 cultivars (Ambrosi et al. 1994) grown from the cool temperate 50° North latitude, through the dry mediterranean-type climates to the tropics. The diversity of this species with respect to its tolerance to drought seems large, yet it has been generally classified as ‘drought avoiding’ (Smart & Coombe 1983) or as ‘pessimistic’ following the ecological classification of Jones (1980) into ‘pessimists’ and ‘optimists’. The principal difference in strategies would be that ‘pessimists’ would modify their growth and physiology to conserve current resources and to control their demand for future resources whereas the ‘optimists’ use all the resources available to them in expectation of more arriving. This ecological classification is analogous to the physiological classification into isohydric and anisohydric plants (Stocker 1956; Tardieu & Simonneau 1998) and fundamentally linked to stomatal behaviour.

Barley and sunflower for instance belong to the anisohydric class of plants (Tardieu, Lafarge & Simonneau 1996; Tardieu & Simonneau 1998) that have a Ψleaf which markedly decreases with increasing evaporative demand during the day and is lower in droughted than in watered plants. In contrast, other species such as cowpea (Bates & Hall 1981), maize, poplar (Tardieu & Simonneau 1998) or sugarcane (Saliendra & Meinzer 1989) maintain a near constant Ψleaf during the day at a value which does not depend on soil water status. The classification into isohydric and anisohydric plants so far only applies to different species. However, some Vitis vinifera L. cultivars of contrasting genetic origin show very different responses of Ψleaf during water stress, which suggests that a similar classification within the same species may exist (Düring & Scienza 1980; Chaves et al. 1987; Winkel & Rambal 1993; Schultz 1996).

The apparent differences in stomatal control of isohydric and anisohydric plants are thought to be due to differences in the perception of abscisic acid [ABA] (Tardieu & Simonneau 1998), the chemical signal coming from the roots and the most likely candidate for root-to-shoot signalling in stomatal control (e.g. Davies & Zhang 1991). Such differences in stomatal behaviour may be related to the presence or absence of a sensitivity with respect to high evaporative demand and high temperature, frequent environmental cofactors of developing water deficit, or Ψleaf itself, which can modify the response of stomata to [ABA] in isohydric but not in anisohydric plants (Davies & Zhang 1991; Tardieu & Simonneau 1998).

[ABA] has been demonstrated to act on stomatal conductance (g) of grapevines (e.g. Loveys 1991; Correia et al. 1995; Stoll, Loveys & Dry 2000; Lovisolo, Hartung & Schubert 2002), yet a definite control has only be shown for mid-morning maximum g (gmax) (Correia et al. 1995). However, g as well as leaf water status undergo large diurnal fluctuations without substantial changes in soil water content and the controlling mechanism may not be directly related to [ABA] (Assmann, Snyder & Lee 2000). For instance the rapid decrease in Ψleaf in grapevines observed in the field after sunrise, even in well-watered plants (Naor & Wample 1994; Schultz 1996), suggests the development of substantial water potential gradients in the soil–plant system, which, despite the increased transpiration rates, indicate large hydraulic resistances in the water-conducting pathway (Schultz & Matthews 1988a). Differences in the diurnal behaviour of Ψleaf of different cultivars under drought could thus be related to the water-conducting capacity and stomatal behaviour may respond to a hydraulic signal (Fuchs & Livingston 1996; Hubbard et al. 2001; Comstock 2002). Jones & Sutherland (1991) have proposed that stomata act primarily to avoid damaging water deficits causing cavitation in the xylem (Tyree & Sperry 1989). Although at first glance this would only fit to isohydric behaviour (Jones 1998), depending on the hydraulic capacity and the proportion of conducting tissue which could be sacrificed for embolisms to maximize stomatal aperture and, hence, short-term productivity (Jones & Sutherland 1991), it would be applicable also to plants with anisohydric behaviour (Hubbard et al. 2001; Comstock 2002). Woody plants such as Quercus (Cochard, Bréda & Granier 1996), Juglans regia (Cochard et al. 2002) or Piper auritum (Schultz & Matthews 1997) can achieve control of cavitation by stomatal closure, and this has also been suggested for grapevines in pot studies (Lovisolo & Schubert 1998). There have also been observations of reduced g in response to reduced hydraulic conductance but at a relatively constant Ψleaf, suggesting a feedback link between g and some form of hydraulic signal (Meinzer & Grantz 1990; Sperry, Alder & Eastlack 1993; Saliendra, Sperry & Comstock 1995; Meinzer et al. 1995; Fuchs & Livingston 1996; Meinzer et al. 1999; Salleo et al. 2000; Nardini, Tyree & Salleo 2001). The mechanisms underlying such a response are not clear, but may be related to localized water potential differences as a result of the hydraulic properties of the water pathway (Nonami & Boyer 1993), or through pressure–volume changes in sensing cells or localized cavitations (Canny 1997; Salleo et al. 2000; Nardini et al. 2001).

The hypothesis that different cultivars of the same species have different stomatal sensitivities to drought and may thus act as isohydric or anisohydric plants depending on their genetic background was tested in the field on two grapevine varieties of different geographical origin. One, Grenache, originates from the Mediterranean basin and is largely planted in southern France and northern Spain, the other, Syrah, is of mesic origin from the Rhone valley. The importance of plant hydraulic conductance in the control of stomatal aperture during soil drying was assessed. The argument that apparent differences in stomatal sensitivity may be due to differences in plant developmental rate and thus leaf area formation and subsequent different rates of soil water depletion rather than true genetic differences in stomatal control was also addressed (Borel et al. 1997; Tardieu & Simonneau 1998).

MATERIALS AND METHODS

Plant material and treatments

Eight-year-old-grapevines (Vitis vinifera L. cvs. Grenache and Syrah on Rupestris x Berlandieri rootstocks) grown side by side in a commercial vineyard near Montpellier, France (43°45′ N, 3°40′ E), were investigated during the 1994 and 1995 growing seasons. The site is on calcareous soil, imbedded in indigenous ‘garrigue (macchia)-type’ vegetation and the planting density was 3860 vines ha−1. Seven vines of each variety were irrigated weekly at 30 L per vine after bloom (15. June) in 1994. In 1995, drip irrigation was installed and about 30 vines were irrigated at maximum capacity (up to about 50 L per vine/week), the rest of the vineyard received only natural precipitation. The March–August rainfall was 169 mm in 1994 and 119 mm in 1995, and there was less than 50 mm of precipitation between May and September each year. In some cases, individual vines were re-watered in mid-August to study their recovery.

In a second experiment, 3-year-old vines of the same varieties and grown on Fercal rootstocks in an experimental field in Geisenheim, Germany (50°N, 8° E), were used in 2001 to assess for genetic differences in hydraulic architecture in the absence of water deficit.

Leaf gas exchange measurements

Leaf gas exchange measurements were conducted on mature, well-exposed leaves with a modified ADC-LCA 3-type open-system (ADC, Hoddesdon, Herts, UK) enclosed in a shade chamber to avoid overheating and equipped with a Parkinson leaf cuvette. Gas exchange parameters were calculated from raw data using the equations of von Caemmerer & Farquhar (1981). In most cases, photosynthetic rate (A), transpiration rate (E) and stomatal conductance (g) were measured throughout the day between sunrise and sunset. From these data sets maximum photosynthetic rate (Amax) and maximum stomatal conductance (gmax) were determined.

Leaf area and leaf plastochron index

Leaf area in the field in France and on excised shoots in Germany was determined non-destructively from measurements of leaf lamina length using a quadratic model (Schultz 1992). All leaves on 10 shoots per vine, on three to five vines per treatment were measured and individual leaf area calculated to estimate the average leaf area per shoot. The average number of shoots per vine was then counted on all irrigated vines and 15 vines in the stress treatment and multiplied by the average shoot leaf area to estimate total vine leaf area. Leaf area lost through senescence was estimated by counting the nodes on the measurement shoots without leaves and subsequently subtracting the leaf area calculated for these positions from measurements on shoots which had retained all leaves. The leaf plastochron index (LPI) was calculated as described previously (Schultz & Matthews 1993) to quantify physiological age and used to make comparisons at the same developmental stage.

Water relations measurements

Leaf water potential (Ψleaf) measurements were conducted with a pressure chamber (Soilmoisture Corp., Santa Barbara, CA, USA) according to Turner (1988) either predawn (0500 h; ΨPD), midday (1200–1600 h; ΨM), or immediately following leaf gas exchange measurements during the course of a day in the field. In the latter cases, Ψleaf was determined on the same leaves used for gas exchange measurements. Leaf-specific hydraulic conductance, Kl, was estimated from the slope of the relationship between leaf transpiration rate, E (mmol m−2 s−1) and Ψleaf, and was calculated as E/(Ψleaf − ΨPD) (Sperry & Pockman 1993).

In some experiments, hydraulic conductance was measured on excised shoot internodes and leaf petioles. Shoots were cut under water predawn in the field at the end of the growing season (October) in Geisenheim, Germany, transported to the laboratory and re-hydrated for 2 h before measurements. For analyses of hydraulic conductivity, Kh (m4 MPa−1 s−1), internodes and petioles were excised under water about 0.3 cm from the nodes and the junction into the leaf lamina (petioles) (Schultz & Matthews 1993). The samples were then sealed into a pressure chamber filled with degassed and deionized water, and pressurized at either 0.02, 0.05, 0.10, 0.20 and 0.3 MPa depending on the physiological age (LPI) of the sample (Schultz & Matthews 1993). Flow (q, m3 s−1) was measured by collecting the exudate with a preweighed vial containing cotton wool placed over the protruding sample to minimize evaporation. Steady flow rates (± 5%) were attained within 10–30 s, depending on sample size, and could be maintained for at least 10 min. Kh was calculated relating q to the driving force, namely the water-potential gradient, ΔΨx (MPa m−1), across the segment, Kh = q/(ΔΨx). Leaf area specific conductivity, LSC (m−2 MPa−1 s−1), was calculated as LSC = Kh/A, where A (m2) is the supplied leaf area by the segment (Zimmermann 1978; Salleo, Rosso & Lo Gullo 1982).

Xylem cavitation

Vulnerability to cavitation was investigated in terms of continuous counts of ultrasound acoustic emissions (Tyree & Dixon 1983) parallel to hydraulic conductance measurements on separate excised shoots from the experimental field in Geisenheim, Germany. A two-channel AMSY-4 ultrasonic acoustic emissions (AEs) detector system (Vallen-Systeme, Icking, Germany) equipped with a multiplexer (33 MUX1 model; Vallen Systeme) for 12 acoustic emission sensors (VS 150-M; Vallen-Systeme) was used. Signals were amplified in the 150–400 kHz range (peak response at 150 kHz, AEP 4; Vallen Systeme), where several plants have been shown to emit peak signals (Tyree & Sperry 1989). The gain was set at a peak amplitude of 33–90 dB. Data were recorded automatically on a PC. Due to limited plant material and the advanced time during the season, only one replicated experiment was conducted on cut shoots of the two cultivars in the laboratory. In each case, three sensors per shoot per cultivar were attached to leaf petioles and three sensors to internodes (stem). The positions were chosen to represent the upper portion (LPIs 4–6), the middle portion (LPIs 11–13) and the basal portion (LPIs 18–22) of the shoots, which were then left to dry in air over a period of several hours. Concomitant to the recording of the AEs, leaf water potential was determined over the dry-down period using a pressure chamber.

RESULTS

Single leaf gas exchange

The course of Amax and gmax during the season was similar for Syrah and Grenache for any given day (Fig. 1a & b), yet the degree of water deficit, expressed as ΨPD, and ΨM, at which these values occurred evolved differently for the two varieties during dry down (Fig. 1c, d & Fig. 2). The Amax and gmax were more sensitive to water deficit in Grenache, where ΨPD decreased only to a low of −0.85 MPa at the end of August [day of year (DOY), 240], as compared to Syrah, where ΨPD at the same date had reached −1.4 MPa (Fig. 1c), yet with similar photosynthetic rates and stomatal conductances (Figs 1a–c & Fig. 2). An exponential decay function was fitted to data relating gmax to ΨPD[gmax = a exp(–b ΨPD)] using non-linear least square analysis (Sigma Plot 5.0; SPSS, Chicago, IL, USA). The function was first fitted to all data irrespective of variety and treatment and then separately to the data from each variety. This was repeated using only the data from the stress treatments. Genotypic differences in the fitting curves were then assessed for using a F-test (systat 10; SPSS) performed between single models grouping all data of both varieties and their subgroups. The fitted curves were significantly different at either the 1% (pooling stressed and control plants) or the 5% (only stressed plants) probability level.

Figure 1.

Changes in Amax (a); gmax (b); predawn (c), and midday Ψleaf (d), for irrigated (cont.) and water-stressed (stress) Grenache and Syrah vines during the 1994 season. Data are means ± SE for measurements on five to six fully exposed leaves per treatment.

Figure 2.

Relationship between gmax and ΨPD for irrigated (control) and water-stressed (stress) Grenache and Syrah vines during the 1994 season. Data are means ± SE for measurements on five to six fully exposed leaves per treatment. Curves are from non-linear regression analysis fitting an exponential decay function to the stress treatment data (Grenache: gmax = 212.58 exp(−2.2957 ΨPD), R2 = 0.94; Syrah: gmax = 190.69 exp(−1.4927 ΨPD), R2 = 0.91). Curves are significantly different at the 5% probability level.

The shown differences imply very different strategies in water use. Whereas ΨPD and ΨM were always lower for stressed as compared to control vines for Syrah, ΨM of stressed Grenache was not different from the control (Fig. 1d), despite large differences in ΨPD (Fig. 1d). This indicated a strong stomatal closure response to maintain diurnal leaf hydration at near control levels despite low soil water contents and closely resembled classic isohydric behaviour. Recovery of Amax and gmax after rainfall on 8 September (DOY 251) was similar and occurred equally fast (within 24 h) for both varieties, yet prestress values were not attained (Fig. 1a & b). The results also indicate that for the irrigated controls a water supply of 30 L vine−1 week−1 was insufficient at mid-season to ensure the maintenance of springtime values of A and g (Fig. 1).

The diurnal course of A and g in relation to Ψleaf at different times during dry-down in the season confirmed clear differences between the varieties (Fig. 3a–l). For Syrah, A and g decreased with ΨPD (Fig. 3b, f & j) during water deficit and the change over time in Ψleaf with increasing evaporative demand during the day maintained approximately the same amplitude (near 0.7 MPa) in comparison with the well-watered conditions at the beginning of the season (Fig. 3j). Stomatal closure therefore only exerted limited control against dehydration and could not compensate for decreases in soil water status, typical of anisohydric behaviour. For Grenache in contrast, Ψleaf did not drop significantly below the minimum Ψleaf of watered plants (Fig. 3k & l) despite decreasing ΨPD (Fig. 3d) and resulting decreases in A and g approximately similar in magnitude and diurnal pattern as those of Syrah (Fig. 3a–d). Thus, stomatal closure was capable of balancing decreasing soil water potential against diurnal evaporative demand, typical of isohydric behaviour.

Figure 3.

Diurnal time course of net photosynthesis (A) (a–d), stomatal conductance (g) (e–h), and leaf water potential (Ψleaf) (i–l) of 5 d during developing water deficit in the field for irrigated (cont.) and stressed (stress) Syrah and Grenache grapevines. The A and g data are means ± SE from four to six leaves, Ψleaf data are means ± SE of four leaves from the irrigated and six leaves from the stressed plants. Each measurement was conducted on a different plant. Dates and symbols are as follows: 30 June (○), 15 July (▪), 11 August (▵), 23 August (•), 30 August (–).

Leaf specific conductance

Tyree & Sperry (1988) suggested that stomatal regulation may function to keep transpirational flow in tune with the plants hydraulic capacity. Therefore the leaf specific hydraulic conductance, Kl, was estimated from the relationships of single leaf transpiration rate to the soil–xylem water potential difference (in the present case soil water potential was assumed to be equivalent to predawn water potential). Figure 4 shows the relationship of Ψleaf to E for 4 d during the water deficit. The slope of this relationship is an estimate of Kl (mmol MPa−1 m−2 s−1). In stressed vines Kl decreased with increasing water deficit (Fig. 4a–d). To some extent a decrease in Kl was also apparent in the control vines at the end of the season (Fig. 4c & d). Large differences in Kl were not observed between the two varieties on the 4 d shown, but stomatal regulation after mid-morning maximum conductance, as indicated by the arrows in Fig. 4, occurred at higher Ψleaf and lower E in Grenache, especially when the stress became severe (Fig. 4c & d). Since the ΨPD was lower for Syrah than for Grenache on all days during the stress, Kl at similar levels of stress (i.e. ΨPD) was less for Grenache than for Syrah (Fig. 5) which could imply a causal relationship between Kl and stomatal conductance. After re-watering (24 h), Kl did not immediately recover to prestress values for both varieties (Fig. 5 ellipse), suggesting a longer lasting impairment of water transport, which was independent of predawn water potential.

Figure 4.

Relationship between leaf water potential, Ψleaf, and transpiration rate, E, of control and stressed Syrah and Grenache vines plotted for 4 d during a continuous water deficit. Direction of arrows indicate stomatal influence after mid-morning maximum conductance. Lines are linear regression lines. Arrows were added by hand.

Figure 5.

Mid-morning leaf specific hydraulic conductance, Kl., as a function of ΨPD for irrigated (control) and water-stressed (stress) Grenache and Syrah vines between June and September Some previously stressed vines were re-watered on 22 August (recovery). The encircled symbols represent values from recovery plants 24 h after re-watering (23 August).

When diurnal data of g from control, stressed and recovery plants were plotted as a function of Ψleaf and Kl for different days during the water deficit no relationship was found between g and Ψleaf(Fig. 6a, c & e), yet a close and common relationship for all treatments and both varieties existed between g and Kl (Fig. 6b, d & f). These data suggested that differences in stomatal behaviour resembling isohydric and anisohydric strategies were related to plant hydraulics. The controlling mechanism for g was uniquely related to Kl for stressed and unstressed plants alike and functional also on a diurnal time scale when environmental variables and leaf water relations underwent large fluctuations. Furthermore, this relationship existed despite large differences in leaf area and leaf area indices (LAI) between varieties and treatments (Table 1) which probably strongly influenced whole-plant transpiration rates. It should be noted that the cultivar with the highest stomatal sensitivity had the smallest leaf area, thus soil water exploitation was slower (see Fig. 1). Nevertheless it needs to be pointed out that E, Kl, and g are not completely independent variables and that some auto-correlation is involved in the shown relationships since E could not be estimated by an independent technique.

Figure 6.

Stomatal conductance, g, as a function of leaf water potential, Ψleaf (a, c, e), and leaf specific conductance, Kl (b, d, f) for control, stress, and recovery treatments of Grenache and Syrah vines during 3 d of increasing water stress in the field. Plants of the recovery treatment had been re-watered twice during the 4 d prior to the measurements on 23 August. Water potential measurements were conducted immediately after measurements of leaf gas exchange.

Table 1.  Leaf area parameters of irrigated and non-irrigated Grenache and Syrah vines. Measurements were conducted at the end of the growing season in 1994 (September). A one factor analysis of variance was applied to compare the two controls and the two stress treatments
 GrenacheSyrah
ControlStressControlStress
  • **

    P < 0.01,

  • *

    P < 0.05, and NS, not significant P > 0.05.

LA/vine (m2)7.54**4.87*12.385.59
Projected LAI (m2 m−2)3.81**2.95 NS 7.033.18
LAI (m2 m−2)2.91**1.88* 4.792.16
LA/vine lost (m2)0.54 NS1.44 NS 0.481.43

The observed differences in stomatal sensitivity between the two cultivars during drought were apparently linked to differences in hydraulic architecture somewhere in the water-conducting pathway of the plants. To test this, the hydraulic properties of excised internodes (stem portions) and petioles were analysed. The Kh of internodes was not different between cultivars irrespective of physiological age (LPI) (Fig. 7a), but was two- to three-fold higher for Grenache leaf petioles than for Syrah (Fig. 7b). Leaf area specific conductivity, LSC, was about 30% higher for Grenache internodes> LPI 10 (Fig. 7c), which is the developmental stage when internodes and corresponding leaves reach their final size (Schultz & Matthews 1988b). LSC of leaf petioles of Grenache was four- to six-fold higher than the LSC of Syrah petioles (Fig. 7d). When shoots of comparable leaf area (Grenache, 1701 ± 274 cm2; Syrah, 1758 ± 183 cm2) were left to dry in air, Syrah attained its peak acoustic emission rate early, declining before Grenache reached its peak. Despite of the differences in timing, peak AE rates occurred at about the same Ψleaf for both varieties suggesting different rates of drying. However, Grenache emitted acoustic signals at higher rates for longer time periods than Syrah from both internodes (up to about 1.5 h) and petioles (up to about 1 h) (Fig. 8), yet maintained higher Ψleaf values than Syrah (Fig. 8b), which was probably related to a better control of stomatal conductance. There was no apparent difference in the timing of the onset of cavitation events between petioles and internodes.

Figure 7.

Hydraulic conductivity, Kh (a, b), and leaf area specific conductivity, LSC (c, d), of internodes (a, c), and leaf petioles (b, d) at different locations (LPI) on the shoots of Grenache and Syrah plants. The Kh was calculated from pressure–flow experiments on excised plant parts. The LSC for internodes was calculated by relating Kh of a stem section to the leaf surface area of all leaves distal to that section. For petioles LSC was calculated relating Kh to the leaf surface area supplied by that petiole.

Figure 8.

Acoustic emission (AE) rate after excision of fully developed shoots of Grenache and Syrah vines over time for (a) internodes and (b) petioles. AE-sensors were connected to three internodes per shoot and variety (shoot leaf area: Grenache, 1701 ± 274 cm2; Syrah, 1758 ± 183 cm2) and three additional sensors were connected to three petioles per shoot (leaf blade area: Grenache, 64.8 ± 26.1 cm2; Syrah, 71.4 ± 10.5 cm2). Data are means of a replicated experiment over 2 d. Error bars denote SE for three sensors per organ per shoot replicated once (n = 6). Ψleaf was measured occasionally over the course of the experiment (b).

DISCUSSION

The analysis of seasonal and diurnal leaf gas exchange revealed that Grenache exhibits near-isohydric and Syrah anisohydric stomatal behaviour with respect to a slowly developing water deficit. It was demonstrated that a common and close coupling between g and leaf specific hydraulic conductance exists, which can fully account for the observed differences in stomatal behaviour on a seasonal and diurnal basis. It is hypothesized that the differences in water-conducting capacity of stems and especially petioles, may be at the origin of this behaviour, with the cultivar with the highest hydraulic conductance being more sensitive to cavitation and thus inducing stomatal closure at higher leaf water potentials.

Categorizing two varieties of the same species as isohydric and anisohydric is new. Intraspecific differences in stomatal sensitivity have been related to faster leaf area development with a faster soil water depletion and consequently earlier stomatal closure (Borel et al. 1997; Tardieu & Simonneau 1998) or to a higher gmax in the absence of stress (Davies & Gowing 1999; Oren et al. 1999). Both arguments do not apply to the described grapevine cultivars. Soil exploitation of water was faster for Syrah, which maintained a larger leaf area, yet stomata remained less sensitive than those of Grenache. Additionally, both varieties did not differ significantly in gmax in the absence of stress.

Whole-plant versus single organ hydraulics

Syrah could sustain transpirational flow despite a larger leaf area, through less sensitive stomates which implies a higher water transport capacity in the soil–plant–atmosphere continuum, and/or possibly a reduced sensitivity to xylem embolisms (Tyree & Sperry 1989). If the hydraulic capacity is high enough or the threshold for cavitation resides at a lower Ψleaf, g would start responding at higher values of the vapour pressure deficit between leaf and air (VPD), or E (Mott & Parkhurst 1991). A co-ordination between liquid flow conductivity and the vapour phase conductance, limited either by the stomata or the external boundary layer has been shown for several species (Meinzer & Grantz 1990; Meinzer et al. 1995; Schultz & Matthews 1997; Meinzer et al. 1999) and has been extended to demonstrate a link between hydraulic supply and photosynthesis (Brodribb & Field 2000; Bond 2000; Brodribb, Holbrook & Gutiérrez 2002). This link can also be invoked from the relationships of seasonal and diurnal changes of g and Kl (Figs 3, 4 & 6) in this study.

The diurnal measurements of Kl in the field could be subject to errors since they were conducted under non-steady-state conditions under which the Ohm's law analogy would not hold (Hubbard et al. 2001). These errors could arise if changes in plant capacitance over time contribute significantly to E (for an example see Schultz & Matthews 1997). The possible contribution to E of water stored in different organs was therefore estimated. From pressure–volume isotherms of the two varieties in the same experiment, the amount of water lost by dehydration was calculated for a drop in Ψleaf of 0.7 MPa between predawn and mid-day for stressed plants (Fig. 4). This drop resulted in a decrease in relative water content of about 10%, which was equivalent to a contribution of 13.3 g m−2 for Grenache and 12.5 g m−2 for Syrah for a period of about 4 h in the morning at the height of the stress. These values represent between 4% (Syrah) and 6.5% (Grenache) of the measured E (in mmol m−2 s−1) at mid-day. For control plants the error is about 2%. The small capacitance in grape was also evident from the rapid onset of cavitation events when shoots were left to dry in air (Fig. 8). For this reason, whole branches instead of single shoots are usually used in excision experiments to ensure that the internal water reservoir will buffer the decrease in Ψleaf (Salleo et al. 2001).

In addition, one could assume a contribution of the trunk volume in a field situation, which has been shown to undergo contraction–expansion cycles in grapevines from 0.02 mm under well-watered conditions to a maximum of 0.1 mm for a 6 h period under water stress (Myburgh 1996). For a trunk diameter of 3.5 cm and a trunk height of 60 cm in relation to the leaf areas shown in Table 1, changes in capacitance would then account for an additional 2.4–7.6% error in Kl calculation for stressed plants and somewhat less than half these values for control plants (some compensation occurs due to the larger leaf areas). Taken together, these errors would not have substantially changed the shape of the curves in Fig. 6, nor would they have influenced the similarities for varieties and stress levels in the dynamics of Kl.

Measurement of E inside gas exchange cuvettes may be problematic in calculating Kl if the natural leaf or canopy boundary layer conductances are substantially different from the conditions in the cuvette (Meinzer et al. 1995). Grapevine canopy conductance in the field has been shown to be highly coupled to stomatal conductance (Lu et al. 2003). Boundary layer conductance in a field situation has been measured to be in the range of 0.4–0.9 mol m−2 s−1 in a study with individual leaf replicas embedded in the canopy at intermittent wind speeds of 1.5–3.0 m s−1 (Daudet et al. 1998). Much larger wind speeds usually prevail in the study area (Trambouze 1996) which are in the range of the wind speed in the cuvette (4 m s−1). Thus boundary layer conductance is expected to be higher than the above values (Grace 1989) but not expected to vary between cultivars because of similar leaf size and shape.

Whereas mid-morning Kl showed a higher whole-plant hydraulic capacity for Syrah during the water deficit at comparable ΨPD, the data from excised stem segments and petioles showed the inverse (Fig. 7). This apparent contradiction may form the basis of the control mechanism for g in the isohydric and anisohydric strategies of different grape cultivars. Grenache with a high LSC, especially in the petioles, will maintain a higher water potential at similar ΨPD than Syrah (Tyree 1997). This high hydraulic capacity may lower the Ψleaf threshold for the formation of embolisms. This could be related to larger vessel diameters (Tyree & Sperry 1989; Lo Gullo et al. 1995; Lovisolo & Schubert 1998), although specifically for the species in question, in vivo observations of cavitation were not correlated with vessel diameter (Holbrook et al. 2001). These cavitation events may then act as a signal for stomatal closure (Sperry & Pockman 1993; Salleo et al. 2000; Nardini et al. 2001) or stomatal closure would occur to prevent cavitation (Tyree & Sperry 1988). Both possibilities would occur earlier, namely at higher Ψleaf for Grenache than for Syrah, in the first case maintaining homeostasis in Ψleaf. This behaviour would be compatible with that of many other species in which typical midday xylem pressure is close to the threshold pressure for cavitation (Tyree & Sperry 1989; Cochard et al. 1996; Nardini et al. 2001). Although one of these mechanisms might explain the sequence of events, the cavitation data in the present study on shoots drying in free air are no prove that this actually occurs. Nevertheless, the data provide some indication that Grenache is more sensitive to the formation of embolisms at the beginning of dry-down (Fig. 8), and that this may occur at a higher Ψleaf. Clearly more data are needed to link the regulation of g to cavitation for grapevines under field conditions and identifying the origin and the location of the closure-inducing signal.

In the present study, the differences in Kh and LSC between the two varieties were much larger for leaf petioles than for stem segments indicating substantial differences in vessel anatomy, yet the pattern of cavitation was similar (Figs 7 & 8). Zwieniecki et al. (2000) found decreases in in situ petiole specific conductivity associated with increasingly more negative Ψleaf in maple and tulip tree, but not in grape (Vitis labrusca L.). In their study, petioles had conductivity values ranging from 12 to 100 g s−1 m−1 MPa−1. In comparison, the data presented here would convert to 41.9 g s−1 m−1 MPa−1 (± 11.5 SE, n = 25) for Syrah and 127.7 g s−1 m−1 MPa−1 (± 20.7 SE, n = 24) for Grenache. The decrease in conductivity of petioles from the other two species were attributed to cavitation and the mid-afternoon increase to embolism repair under tension. Recovery of specific hydraulic conductance could be inhibited by the application of mercury derivatives suggesting the involvement of aquaporins (Zwieniecki et al. 2000), which have been shown to modify hydraulic conductance (Maurel & Chrispeels 2001). In this context the stability of the grape petiole specific conductivity values with decreasing Ψleaf would indicate the absence of cavitation or a perfect balance between embolism formation and repair (Zwieniecki et al. 2000). Holbrook et al. (2001) observed in vivo that refilling embolized vessels in stems of grape did not occur when plants transpired actively which would contradict the repair hypothesis for this species, at least for stems, yet it seems possible that differences in these mechanisms may also exist between cultivars.

To which signal from which organ stomata may respond seems species dependent. Salleo et al. (2001) found that stomata in Laurus nobilis L. responded to stem cavitation, although cavitation in leaves occurred at higher Ψleaf. In the same experiment the leaf conductance of Ceratonia siliqua L., which is termed more drought resistant, decreased in coincidence with the cavitation threshold in the leaf and Cochard et al. (2002) demonstrated for Juglans regia that stomata acted to prevent cavitation in the leaf petiole as being the ‘Achilles heel’ of the sap pathway. Thus, the degree of vulnerability to cavitation may not be related directly to the control of g. Other studies have tried to correlate changes in hydraulic conductance with local changes in the leaf in order to get closer to the controlling mechanism of photosynthesis and stomatal conductance during water deficit (Aasamaa, Söber & Rahi 2001, Nardini et al. 2001; Zwieniecki et al. 2002).

Stomatal closure in the present study probably responded to a decrease in leaf specific hydraulic conductance or vice versa but other controlling factors cannot be entirely excluded. Most of the evidence in support for a hormonal control or even a combination of chemical and hydraulic messages come from studies on herbaceous plants (e.g. Tardieu & Davies 1993; Tardieu & Simonneau 1998). Fuchs & Livingston (1996) argued that there may be fundamental differences between herbaceous and woody plants, which by virtue of their larger size, may be less reliant on slow-moving root signals. In favour of an exclusive hydraulic control in Vitis, irrespective whether stomatal regulation resembled isohydric or anisohydric behaviour, are the data showing a strong correlation between g and Kl for control, recovery and stressed plants throughout different days during the water deficit (Fig. 6). In contrast, Correia et al. (1995) in their study on grapevine found conductance to be strongly related to [ABA] coming from the roots only in the morning when g was maximum, but correlation was lacking during the course of the day. Nevertheless, there are other possible mechanisms, by which [ABA] may transfer the closing signal to stomates (Hartung, Sauter & Hose 2002). Lovisolo et al. (2002) recently showed with experiments on potted grapevines grown in a split-root system, that [ABA] controlled g without changes in whole-plant hydraulic conductance during water deficit (about −0.2 MPa Ψsoil). To draw conclusions from this for plants grown in the field is somewhat difficult, since the stress level in the present experiment was much more severe (up to −1.4 MPa ΨPD) and g was not at all affected by water deficits of the magnitude used in the study by Lovisolo et al. (2002) (see Fig. 2).

Differences in rooting depth between the two varieties may have played a role in conductance capacity and sensitivity to embolisms (Jackson, Sperry & Dawson 2000) and may have resulted in or contributed to stomatal closure through a hydraulic signal (Fuchs & Livingston 1996; Lovisolo & Schubert 1998) but was not assessed in this study.

CONCLUSION

In this paper some evidence is presented that near-isohydric and anisohydric behaviour in grapevine cultivars of different geographical origin under water stress can be explained by the differences in Kl both on a seasonal and a diurnal basis. This is the first time that anisohydric and isohydric behaviour has been described to occur within the same species. Although the exact controlling mechanisms for stomata remain to be elucidated, differences in hydraulic architecture in the green shoot, specifically in the petioles between the two varieties suggest that stomata may react to prevent embolism at different levels of Ψleaf or that cavitation may be the signal for stomatal closure.

ACKNOWLEDGMENTS

I thank Gérard Bruno, Eric Lebon, and Francois Champagnol for their help with irrigation and setting up the experiment in France and Professor Mark Matthews, U.C. Davis, for his comments on the manuscript. Antigone Chouzouri helped with data treatment of the ultrasound acoustic emission system. Financial support for this study was provided in part by the Deutsche Forschungsgemeinschaft to H.S, the EU contract AIR 1743 to INRA Montpellier and the ‘Forschungsschwerpunkt Wassersressignale und Haltbarkeit’ of the Forschungsanstelt Geisenheim.

Received 11 November 2002; received in revised form 14 February 2003; accepted for publication 3 April 2003

Ancillary