Carbon isotope composition of sugars in grapevine, an integrated indicator of vineyard water status

Journal of Experimental Botany, Vol. 53, No. 369, pp. 757 763, April 2002 Carbon isotope composition of sugars in grapevine, an integrated indicator of vineyard water status Jean-Pierre Gaudillère 1,4, Cornelius Van Leeuwen 2 and Nathalie Ollat 3 1 INRA Agronomie, BP 81, 33883 Villenave d Ornon, France 2 ENITA de Bordeaux, 1 Cours du Général de Gaulle, 33175 Gradignan cedex, France 3 INRA UREFV, BP 81, 33883 Villenave d Ornon, France Received 30 July 2001; Accepted 23 November 2001 Abstract Photosynthetic carbon isotope composition (d 13 C) was measured on sugars in mature fruits from fieldgrown grapevines. Sugar d 13 C and summer predawn leaf water potential were significantly correlated. The survey of different vineyards during four growing seasons showed that sugar d 13 C in must at harvest varied from 20ø to 26ø when conditions during berry maturation varied from dry to wet. This range allows a very sensitive detection of grapevine water status under natural conditions. However, local differences due to soil capacity to supply water to grapevines are maintained, whatever the annual water balance. Leaf nitrogen content variations of field-grown grapevines did not change d 13 C values. Genetic variability of d 13 C between 31 grapevine varieties for d 13 C was observed. Must sugar d 13 C can be used to characterize vineyards for their soil structural capacity to provide water to grapevines. It was concluded that isotope carbon composition in grapevine measured on sugars at harvest can be applied to compare the capacities of vineyard soils and canopy management to induce mild water stress in order to produce premium wines. Key words: d 13 C, genetic variability, Vitis vinifera L., water stress. Introduction Unlike other agricultural species, grapevines (Vitis vinifera L.) are generally cultivated under sub-optimum conditions, particularly mild water stress, in order to enhance quality. The water stress can be the result of particular pedoclimatic conditions (rainfall timing or available soil water in the root zone) in locations selected for high quality potential, or can be induced by vineyard management techniques (modification of canopy architecture or irrigation). Water stress reduces shoot growth which improves berry composition, either by limiting the number of sinks for carbohydrates or by improving the microclimate inside the canopy (Smart et al., 1990). Berry size is smaller when grapevines are submitted to mild water deficits, especially those occurring between flowering and veraison (beginning of maturation; Becker and Zimmermann, 1984; Hardie and Considine, 1976; Van Leeuwen and Seguin, 1994), but water deficits occurring from veraison through to harvest cause little reduction in berry size (Becker and Zimmermann, 1984). Early and late mild water deficits increase berry sugar and anthocyanin concentration (Matthews et al., 1990), partly because of reduced berry size. These impacts on vine development and berry composition enhance enological quality potential, especially for red wine production. The extent and timing of mild water deficits have been shown to be a major factor in the terroir effect (Seguin, 1983; Koundouras et al., 1999). Many studies exist on water relations in grapes. Several techniques, using soil moisture status (soil water potential or soil water content), and physiological indicators can assess environmental water supply and actual grapevine water status. However, no existing single method is convenient for assessing vine water uptake conditions over the growing season in a large number of plots at a reasonable cost. Because of the modifications induced by water stress on vine physiology and the positive effects of mild water stress on a balanced vine canopy and berry constitution, an integrative tool for the diagnosis of vine water status is required. 4 To whom correspondence should be addressed. Fax: q33 5 56 84 30 54. E-mail: gaudille@bordeaux.inra.fr ß Society for Experimental Biology 2002

758 Gaudillère et al. Stable carbon isotope uptake is discriminated by diffusion and photosynthesis at the carboxylation step (Farquhar et al., 1980). A robust model computes whole plant water use efficiency from the 13 Cu 12 C (denoted d 13 C) in primary photosynthetic products (Brugnoli and Farquhar, 2000). Basically d 13 C is determined by the gradient of CO 2 between the atmospheric and the intercellular CO 2 concentrations (C i uc a ) and the main factor which affects this ratio is water stress (Farquhar et al., 1989). In oak and pine trees, nitrogen starvation has also been shown to decrease C i uc a and d 13 C value in leaf sugars (Guehl et al., 1995). However,the water use efficiency of other species was not affected by leaf N content (e.g. Douglas fir: Mitchell and Hinkley, 1993). There is no information on the effect of nitrogen nutrition and water use efficiency in grapevine. During berry ripening, sucrose is translocated from leaves to fruit and rapidly converted to glucose and fructose (Davies and Robinson, 1996). Therefore, the carbon isotope ratio in the sugars of mature berries should integrate leaf photosynthetic isotopic discrimination of carbon during berry ripening. Previously published results on grapes (Di Marco et al., 1977) showed that d 13 C in berry juice and water-soluble leaf extracts collected on field-grown vines were very similar during maturation. In this work, a full set of d 13 C measured in purified sugars from ripe berries is reported. The purpose of this paper is to establish, for field-grown grapevines, the relationship between d 13 C and summer water stress. The sensitivity of the response to nitrogen nutrition, water availability and genotype was studied in order to assess the use of carbon isotope composition as an integrative tool for vine water status diagnosis. Materials and methods Plant material and growing conditions All the samples were collected in rain-fed vineyards in the Bordeaux Region, on grapevines older than 5 years. Vines were managed according to the traditional local training systems (vertical shoot positioning, Guyot pruned, 5000 7000 plants ha 1 ). All vineyards were trained to produce 7000 10 000 kg of fruits ha 1. In the Bordeaux region (448519 N) grapes were collected in INRA experimental fields (Château Couhins, Cadaujac and Ferrade, Villenave d Ornon) and in the vineyard Cheval Blanc, Saint Emilion from 1997 to 2000. Berries were harvested at commercial maturity, when the Brix value of the must was higher than 20 (Smart et al., 1990). The relationships between soil water availability, leaf water potential and d 13 C in berry sugars at maturity were studied in a trial comparing three soils (sandy, clay and stony soils) in the same location (Saint Emilion), bearing three varieties (Merlot, Cabernet Sauvignon, Cabernet franc), during four growing seasons (1997, 1998, 1999, 2000). The description of soil water availability in this trial has been previously reported (Van Leeuwen and Seguin, 1994). Briefly, soil water availability at the beginning of the growing season changed with soil depth, clay content and depth of the water table. The sandy soil had a low water content, but a shallow water table provided water to the roots all through the growing season. The clay and stony soils could provide approximately 235 mm and 190 mm of water, respectively, at bud burst. Grapevine genetic variability for the d 13 C content was tested in the INRA genetic conservatory set in the Ferrade domain. The 31 varieties were grafted on the same rootstock (Fercal) in 1994. Each variety was represented by 5 10 individuals and trained collectively in the same plot. The trellis size and green pruning maintained a constant canopy area (1.3 m 2 m 2 soil) for all the varieties, from 15 June to harvest. Berries were sampled on five individuals in 1999 at the same date for all the varieties when the must Brix value of the latest variety reached 20. The canopy light interception and seasonal water balance (Table 2) was calculated from mean canopy light interception (estimated at 50%), daily rainfall and potential evapotranspiration (Penman calculation) (Riou et al., 1994). The year of sampling, 1999, was representative of a rainy season with 170 mm in July and August compared to the mean rainfall (100 mm) observed in the Bordeaux area. The effects of different soil nitrogen was studied in an experimental vineyard (Chateau Couhins, three blocks) with three soil management systems: tillage, 25% or 50% of the soil covered by Festuca elatior L. grass (Rodriguez-Lovelle and Gaudillère, 1999). Soil mineral nitrogen is made less available for grapevine by the competing grass. Water and nitrogen status assessment Predawn leaf water potential was measured every 10 d from June to September in the Saint Emilion vineyard with a pressure chamber (Scholander et al., 1965; Precis 2000 Gradignan, France) on freshly cut, healthy, primary leaves from six plants for each of the three soils and the three varieties. Primary leaves were sampled at veraison for the measurement of nitrogen content. Each leaf was sampled by punching three leaf discs (85 mm 2 ). N content (g m 2 ) of freeze-dried samples was measured using a CN analyser (NA2100 Protein, CE Instruments, San Jose, California, USA). Pruning and weighing of growth of annual wood removed, in winter, was used to estimate vine vegetative vigour. Berry juice and sugar preparation Samples were made of 200 berries harvested randomly in the vineyards just before commercial harvest, defined by the sugar content and titrable acidity of the must. Berry juice was obtained by hand pressing. Ten ml of juice were sampled, rapidly sterilized by autoclaving (120 8C, 20 min), and stored at room temperature to prevent mono potassium tartaric salt precipitation until the preparation of the samples for 13 C measurements. Each sample contained approximately 200 g l 1 hexose. The hexose fraction was purified by anion and cation exchange. One ml of juice was gently pushed onto a 1 ml column filled with ion exchange resin (Bio-Rad; Mixed bed resin AG501-X8). Isotopic analysis Ten ml of purified must containing approximately 1 mg sugars or 5 ml of berry juice were put in a tin capsule, dried and oxidized under oxygen. Carbon isotope content was measured using a continuous flow isotope ratio mass spectrometer (Europa Scientific Ltd., Crewe, UK) as described earlier (Avice et al., 1996). The conventional expression was used with reference

Carbon isotope composition of grapevine 759 to the Pee Dee Belemnite standard (Farquhar et al., 1989): d 13 Rs Rb C ¼ 1000 Rb where Rs is the 13 Cu 12 C ratio of the sample and Rb is the 13 Cu 12 C of the PBD Standard. The isotopic composition of atmospheric CO 2 gives a d 13 C equal to 8ø (Farquhar et al., 1989). Statistical comparisons were done with Excel (Microsoft) and Systat 10 (SPSS Inc.) software. Analytical precision was estimated with replicates of a single sample. One single measure on a sample can estimate the mean d 13 C value better than 0.15 unit of isotope composition (-0.65% of the measure) with a confidence interval of 95%. The field sampling error was measured in 1998 and 1999 on 12 samples collected in the same plot. One single measure on juice made with 200 berries can estimate the mean of the d 13 C value of the plot "0.7 unit with a probability of 95% or "3% of the measure. The d 13 C measured in the must and the purified sugars were significantly linearly correlated (R 2 ¼ 0.995, 66 samples). Sugar d 13 C content could therefore be assessed on must at maturity without any preliminary treatments. However, in this study all the d 13 C values were obtained on purified sugar from mature berry must. four growing seasons. The linear relationship between Y min and d 13 C is highly significant (P-0.001). A general relationship can be drawn (Fig. 3) with data collected from the three red vine varieties (Cabernet Sauvignon, Merlot and Cabernet franc, in 1997, 1998, 1999, and 2000) and a highly significant linear correlation (P-0.001) was obtained. ANOVA analysis of the four seasons, three soils and three varieties data showed a significant year, soil and variety effect with significant soil 3 year and variety3 year interactions (Table 1). The 1998 year and the stony soil combined to cause the driest conditions. Table 2 shows the mean d 13 C measured on all the samples from the three plots in the Saint Emilion vineyard from 1997 to 2000. Mean water balance from May to August was calculated as the difference between rainfall and half the potential evapotranspiration (Penman calculation). The ranking of the d 13 C values matches well with that of the summer mean water balance (July and August), but not the May to August values. Results Relationship between predawn leaf water potential during fruit maturation and d 13 C measured at harvest Leaf water status was measured on three soils types and three varieties (Cabernet Sauvignon, Merlot, Cabernet franc) from the vineyard Cheval Blanc, Saint Emilion, in 1997, 1998, 1999, and 2000. Figure 1 shows an example of the seasonal predawn leaf water potential measured in summer 1998 for the Merlot variety. Minimum predawn leaf water potential measured in August (Y min ) was used as a water stress indicator for each crop. Figure 2 shows the relationship between Y min and d 13 C for the Merlot variety sampled on the three soils and over the Genetic variability of grapevine carbon isotope content The ANOVA analysis of d 13 C recorded on the Saint Emilion vineyard reveals a significant variety effect and a significant variety 3 year interaction (Table 1). A pairwise comparison with the Bonferroni test showed that the Merlot variety response was significantly different (P-0.01) from the two other varieties (mean d 13 C ¼ 23.63). Cabernet Sauvignon and Cabernet Franc were not different (mean d 13 C ¼ 23.25 and 23.29, respectively). Variability for d 13 C was examined on a large set of 31 varieties. Samples were collected in 1999. All berries were harvested at the same date. Sugar d 13 C varied significantly according to the varieties (Table 3). Fig. 1. Vine leaf predawn water potential during the growing season in 1998. Saint Emilion vineyard, Merlot variety, grown on three soils with different water reserves (mean of six replicates "SE). Fig. 2. The relationship between minimum predawn leaf water potential recorded in August and d 13 C. Grapevine, Merlot variety, grown on sandy, clay and gravely soils in 1997, 1998, 1999, and 2000. Each point represents one soil type and one growing season (R 2, significant at 99% confidence level).

760 Gaudillère et al. Table 2. Mean plot water balance during the growing season (May August) and the fast berry growth phase (July August), and mean d 13 C measured on sugars from berries harvested at maturity in the Saint Emilion vineyard (three varieties, and two soils) Water balance was calculated according to the equation: wrainfall 0.5 3(Potential evapotranspiration)x in mm of water). R 2 is correlation between d 13 C and the calculated water balance from meteorological data; figure in brackets is probability of no correlation. Year Water balance (mm) d 13 Cø May August July August 1997 206 28.5 24.1"0.2 1998 115 102 22.2"0.5 1999 42 28 24.1"0.3 2000 16 28 23.2"0.5 R 2 0.74 (42%) 0.99 (0.3%) Table 3. d 13 C measured on grape sugar at maturity from 31 grapevine varieties grafted on the same rootstock (Fercal) and cultivated at Ferrade domain in 1999 The four groups were defined by the K-means clustering method, for four groups. Group Variety d 13 C (ø) Mean d 13 C (ø) Standard deviation Fig. 3. Minimum predawn leaf water potential and d 13 C relationship. (A) Year effect, all varieties mixed. (B) Variety effect (Merlot, Cabernet Sauvignon and Cabernet Franc), all years mixed (R 2, significant at 99% confidence level). Table 1. ANOVA analysis of the year, soil and genotype effect on the d 13 C of sugars in mature berries from three grapevine varieties grown on three different soils at Saint Emilion in 1997, 1998, 1999, and 2000 Source Sum of squares Degrees of freedom F-ratio Probability Year 22.24 3 251.5-0.1% Soil 37.99 2 644.4-0.1% Variety 1.04 2 17.7-0.1% Soil 3 Year 4.08 6 23.0-0.1% Variety 3Soil 0.17 4 1.5 25.8% Year 3Variety 0.95 6 5.3 0.6% Error 0.35 12 A k-means clustering test for four classes which minimized the sums of square within each group was applied to partition the data set (Systat 10). It showed a significant varietal effect on d 13 C in sugars. For the same growing conditions, Riesling, Petit Verdot, Malbec, and Viognier had a high similar d 13 C value. Conversely Carignan, Chenin and Muscat de Hambourg showed the lowest d 13 C value of the variety set. The other varieties were ranked in two clusters. 1 Riesling 21.6 21.9 0.33 Petit Verdot 21.7 Malbec 22.2 Viognier 22.3 2 Cabernet franc 22.5 22.9 0.22 Pinot meunier 22.6 Grenache noir 22.6 Chardonnay 22.8 Aramon 22.9 Mourverdre 23.0 Tannat 23.0 Muscadelle 23.0 Pinot gris 23.1 Merlot 23.2 Gamay 23.2 3 Sémillon 23.3 23.5 0.16 Ugni Blanc 23.4 Cabernet Sauvignon 23.4 Chasselas 23.4 Sauvignon blanc 23.5 Muscat Ottonel 23.5 Sylvaner 23.6 Pinot noir 23.6 Sirah 23.7 Aligoté 23.7 Colombard 23.7 Guewurtztraminer 23.7 Folle blanche 23.8 4 Carignan 24.1 24.5 0.4 Chenin 24.5 Muscat de Hambourg 24.9 Soil nitrogen supply and d 13 C Low leaf nitrogen content showed nitrogen starvation in grassed plots (Table 4) and this was correlated with reduced pruned wood and yield. However, there was no effect of grapevine nitrogen nutrition on d 13 C in berry sugars.

Table 4. Leaf nitrogen content at veraison, dry weight of winter pruned wood, yield and d 13 C in berry sugars at harvest (cv. Merlot grafted on Fercal) Grapevines were grown in the INRA experimental domain of Chateau Couhins in 1999 in a long-term soil tillage trial testing soil nitrogen availability effects and grapevine N nutrition. Soil treatments were, respectively, tillage, and 25% and 50% of the soil covered by Festuca grass. Treatment effects on nitrogen in leaves, growth and d 13 C were tested by analysis of variance and Bonferroni test. Different letters indicate means differ with 99% confidence, nine repetitions for each soil treatment. Soil management Discussion Leaf N (g m 2 ) Pruned wood (g dw plant 1 ) Precision, and ecological use of d 13 C Grape (kg plant 1 ) d 13 C Tillage, bare soil 1.81 a 259 a 3.38 a 25.7 a Grass cover, 25% area 1.67 b 181 b 3.21 a 25.8 a Grass cover, 50% area 1.42 c 128 c 2.83 c 25.8 a This work has shown that carbon isotope composition can be measured with good precision and allows a comparison of the effect of water stress response on grapevines. Basically d 13 C is related to the ratio of intercellular and atmospheric CO 2 concentration (C i uc a ) and water use efficiency and it is well documented that water stress changes water use efficiency and carbon isotopic composition. While low nitrogen nutrition has been shown to reduce C i uc a in pine and oak leaves and increase the d 13 C value (Guehl et al., 1995), in grape no significant nitrogen effect was observed on d 13 C, at least in the range of variation of leaf nitrogen content observed in the trial (Table 4). Therefore it can be suggested that most of the d 13 C change observed in field-grown grapevines was related to water stress. Grape sugar is mainly composed of glucose and fructose, produced after sucrose hydrolysis by invertase during grapevine maturation (Davies and Robinson, 1996). Sugars are stored from the beginning of maturation (veraison) to full maturity, reaching at that time a concentration higher than 1 M l 1. This period of development lasts approximately 45 d from the beginning of August to the end of September depending on variety and location. Sugar carbon at harvest is a large carbon pool without turnover. It can be easily sampled to measure d 13 C of primary photosynthetic products during maturation. Carbon remobilization would interfere by providing carbon previously stored. Under normal growing conditions, however, remobilization is unlikely as it has been shown that a significant amount of carbon is only remobilized to supply berries after major leaf removal (Candolfi-Vasconcelos and Koblet, 1990). A significant correlation between d 13 C measured in sugars of ripe berries, and plant water status measured by minimum predawn leaf water potential was demonstrated for field-grown grapevines. However, occasional predawn Carbon isotope composition of grapevine 761 leaf water potential is not a reliable measure of the longterm water status of field-grown plants submitted to uneven rainfall. The integration of the water stress cannot simply be assessed by computing the mean of several changing predawn leaf water potentials during a period of interest because plant stress response is not additive or linearly related to the environmental stress intensity. Moreover, it has been shown that predawn leaf water potential did not reveal moderate water stress as did stem water potential (Choné et al., 2001). The scatter of data around the relationship between Y min and d 13 C reported in Figs 2 and 3 can be ascribed to some extent to the inaccuracy of predawn leaf water potential as a measure of mean plant water status. Therefore, d 13 C composition in berry sugars is proposed as a potent indicator of grapevine water status during maturation, not only for ecophysiological research but also for practical application in commercial vineyards to improve training systems. All the data collected over four seasons in various locations showed a range in sugar d 13 C from 20ø to 26ø. This range is large considering the precision of the measure (error -3% of the measure), and allows very discriminating comparisons to be made between season or soil condition effects on grapevine water status. Differences in water availability for grapevine between years and soils was clearly and consistently reflected in sugar d 13 C measured at harvest. Variation of d 13 Cin relation to site characteristics has been demonstrated for annual species (Robinson et al., 2000) and for perennials (Saurer et al., 1997; Brooks et al., 1997). This measure can be used to quantify the general ability of vineyards to supply water and the degree of stress in different vineyards. Water stress cannot be controlled precisely in rain-fed vineyards but, according to the level of stress expected, in order to get well-matured berries grapevine management can be adapted to fit grapevine water demand to the local water supply. Canopy leaf area can be adjusted by the planting density, the height of trellising and pruning to fit the canopy structure to the soil and the local mean rainfall. Genetic variability of carbon isotope composition Grapevine is an isohydric species (Tardieu and Simonneau, 1998). In the field, grapevine stomata close early to control water loss and midday leaf water potential (Smart and Coombe, 1983). At the leaf level, the Farquhar theory predicts a high water use efficiency related to lower net photosynthetic activity (Jones, 1983) and water stress-tolerant genotypes. According to Schultz, grapevine varieties with a low water use efficiency have a lower capacity to control midday leaf potential (Schultz, 1997). Genetic variability of d 13 C has been reported for different perennials and annual species (Sun et al., 1996; Handley et al., 1997). Genetic variability was observed

762 Gaudillère et al. among different grapevine varieties d 13 C. The range of variation, from 22 to 24.5, is large and could be used to study the genetic determinism of d 13 C in response to environmental stress. Muscat de Hambourg, Chenin, Carignan, Riesling, and a few other varieties showed very different d 13 C values, indicating significant variability of the stomatal control in Vitis vinifera species (Table 3). Drought tolerance of grapevine varieties have been reported previously (Düring and Scienza, 1980; Schultz, 1997). Riesling and Grenache are classified as water stress-resistant, while Syrah is poorly resistant to water stress. This hierarchy is in accordance with the hypothesis that high d 13 C values would be related to water stress tolerance reported for these varieties (Table 2). The extensive comparison of three varieties, grown on three soils over four years showed significant genetic differences and year interaction in the d 13 C value (Table 1). Merlot was shown to have lower d 13 C value than Cabernet Sauvignon and Cabernet franc. Surprisingly the genetic effect was significant when the season was rainy, in 1997 and 1999 (data not shown). It is concluded that grapevine varieties have different genetic capacities for response to water limitation, and this study is the first large survey of stress response variability. A detailed ranking of grapevine varieties for stress response will require a multi-year survey to take account of the interactions. It was tempting to presume that this behaviour would be related to the geographic origin of the varieties. However, there was no simple relationship between the d 13 C ranking of varieties studied and the climate of their geographical origin (Ambrosi et al., 1997). Conclusion Berry sugar d 13 C composition correlated well with summer grapevine water status. It was less sensitive to the genotype and not affected by nitrogen nutrition. This marker integrates conditions during the ripening phase and allows a precise comparison of mild water stress conditions. d 13 C is well suited to characterize the mean water regime during berry maturation in relation to the training, the soil and the water supply rate (rain or irrigation). This indicator will be very useful to adjust grapevine water demand anduor irrigation to obtain the sought-after mild water stress for grape quality. Grapevine genotype appears as a secondary but significant factor contributing to variations in d 13 C. Acknowledgements Grants from INRA (Agraf) and the Conseil Régional d Aquitaine supported this work. We thank A Ourry, R Dewar, C Molot and workers of the INRA experimental domains for their excellent assistance. References Ambrosi H, Dettweiler-Münch E, Rühl EH, Schmid J, Schumann F. 1997. Guide des ce`pages: 300 ce`pages et leurs vins. Stuttgart: Eugen Ulmer Gmbh & Co, 320 pp. Avice JC, Ourry A, Lemaire G, Boucaud J. 1996. Nitrogen and carbon flows estimated by 15 N and 13 C pulse-chase labeling during regrowth of alfalfa. Plant Physiology 112, 281 290. Becker N, Zimmermann H. 1984. Influence de divers apports d eau sur des vignes en pots, sur la maturation des sarments, de développement des baies et la qualité des vins. Bulletin de l Office International de la Vigne et du Vin 57, 573 683. Brooks JR, Flanagan LB, Ehleringer JR. 1997. Carbon isotope composition of boreal plants: functional groups of life forms. Oecologia 110, 301 311. Brugnoli E, Farquhar GD. 2000. Photosynthetic fractionation of carbon isotopes. In: Leegood RC, Sharkey TD, von Caemmerer S, eds. Photosynthesis: physiology and metabolism. Dordrecht: Kluwer Academic Publishers, 399 434. Candolfi-Vasconcelos MC, Koblet W. 1990. Yield, fruit quality, bud fertility and starch reserves of the wood as a function of leaf removal in Vitis vinifera L. evidence of compensation and stress recovering. Vitis 29, 199 221. Choné X, Van Leeuwen C, Dubourdieu D, Gaudillère JP. 2001. Stem water potential is a sensitive indicator for grapevine water status. Annals of Botany 87, 477 483. Davies C, Robinson SP. 1996. Sugar accumulation in grape berries. Cloning of two putative vacuole invertase cdnas and their expression in grapevine tissues. Plant Physiology 111, 275 283. Di Marco G, Grego S, Tricoli D, Turi B. 1977. Carbon isotope ratios ( 13 Cu 12 C) in fractions of field-grown grapes. Physiologia Plantarum 41, 139 141. Düring H, Scienza A. 1980. Drought resistance of some Vitis species and cultivars. In: Olmo HP, ed. 3rd international symposium on grape breeding. Davis, USA: University of California, 179 190. Farquhar GD, Ehleringer JR, Hubick KT. 1989. Carbon isotope discrimination and photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology 40, 503 537. Farquhar GH, von Caemmerer S, Berry JA. 1980. A biochemical model of photosynthetic CO 2 assimilation in leaves of C 3 species. Planta 149, 78 90. Guehl J-M, Fort C, Ferhi A. 1995. Differential response of leaf conductance, carbon isotope discrimination and wateruse efficiency to nitrogen deficiency in maritime pine and pedunculate oak plants. New Phytologist 131, 149 157. Handley LL, Robinson D, Scrimgeour CM, Gordon D, Forster BP, Ellis RP, Nevo E. 1997. Correlating molecular markers with physiological expression in Hordeum, a developing approach using stable isotopes. New Phytologist 137, 159 163. Hardie W, Considine J. 1976. Response of grapes to waterdeficit stress in particular stages of development. American Journal of Enology and Viticulture 27, 55 61. Jones HG. 1983. Plants and microclimate, a quantitative approach to environmental plant physiology. Cambridge: Cambridge University Press, 264 293. Koundouras S, Van Leeuwen C, Seguin G, Glories Y. 1999. Influence de l alimentation en eau sur la croissance de la vigne, la maturation des raisins et les caractéristiques des vins en zone méditerranéenne (exemple de Némée, Grèce, cépage Saint-Georges, 1997). Journal International des Sciences de la Vigne et du Vin 33, 143 160.

Matthews M, Ishii R, Anderson M, O Mahony M. 1990. Dependence of wine sensory attributes on wine water status. Journal of Science of Food and Agriculture 51, 321 335. Mitchell AK, Hinkley TM. 1993. Effect of foliar nitrogen concentration on photosynthesis and water use efficiency in Douglas fir. Tree Physiology 12, 403 410. Riou C, Pieri P, Le Clech B. 1994. Consommation d eau de la vigne en conditions hydriques non limitantes. Formulation simplifiée de la transpiration. Vitis 33, 109 115. Robinson D, Handley LL, Scrimgeour CM, Gordon DC, Forster BP, Ellis RP. 2000. Using stable isotope natural abundance (d 15 N and d 13 C) to integrate the stress responses of wild barley (Hordeum spontanaeum C. Koch.) genotypes. Journal of Experimental Botany 51, 41 50. Rodriguez-Lovelle B, Gaudillère JP. 1999. Nitrogen availability and carbon partitioning in grapevine. In: Offler T, Thorpe M, eds. International conference on assimilate transport and partitioning, Newcastle, Australia. Hamilton: Lloyd Scott Enterprise Pty Ltd. Saurer M, Aellen K, Siegwolf R. 1997. Correlating d 13 C and d 18 O in cellulose in trees. Plant, Cell and Environment 20, 1543 1550. Scholander PF, Hammel HJ, Bradstreet A, Hemmingsen EA. 1965. Sap pressure in vascular plants. Science 148, 339 346. Schultz HR. 1997. Physiological mechanisms of water use efficiency in grapevine under drought conditions. In: Bravdo Carbon isotope composition of grapevine 763 BA, ed. 5th international symposium on grapevine physiology. Jerusalem, Israel: ISHS, 115 136. Seguin G. 1983. Influence des terroirs viticoles sur la constitution de la qualité des vendanges. Bulletin de l Office International de la Vigne et du Vin 56, 3 18. Smart RE, Coombe BG. 1983. Water relations in grapevines. In: Koslowski TT, ed. Water deficits and plant growth, Vol. VII. New York: Academic Press, 136 196. Smart RE, Dick JK, Gravett IM, Fisher BM. 1990. Canopy management to improve grape yield and wine quality principles and practices. South African Journal of Enology and Viticulture 11, 1 17. Sun Z, Livingstone NJ, Guy RD, Ethier GJ. 1996. Stable carbon isotopes as indicators of increased water use efficiency and productivity in white spruce (Picea glauca (Moench) Voss) seedlings. Plant, Cell and Environment 19, 887 894. Tardieu F, Simonneau T. 1998. Variability among species of stomatal control under fluctuating soil water status and evaporative demand: modelling isohydric and anisohydric behaviours. Journal of Experimental Botany 49, 419 432. Van Leeuwen C, Seguin G. 1994. Incidences de l alimentation en eau de la vigne, appréciée par l état hydrique du feuillage, sur le développement de l appareil végétatif et la maturation du raisin (Vitis vinifera variété Cabernet franc, Saint- Emilion, 1990). Journal International des Sciences de la Vigne et du Vin 28, 81 110.