Abstract
Based on earlier findings showing the effectiveness of preflowering leaf removal at reducing yield in several Vitis vinifera L. genotypes, a 3-year study was carried out on Sangiovese vines to evaluate how the technique also affects vegetative growth, wood carbohydrates reserves, and specific physiological traits such as intrinsic water use efficiency (WUEi) and leaf chlorophyll fluorescence. Early defoliation (D) applied before flowering with elimination of ~80% of the leaf area as compared with a non-defoliated control (C) was confirmed as quite effective in limiting yield per vine, cluster weight, cluster compactness and rot incidence, and berry set and mass in two of three seasons. Defoliation also markedly improved relative berry skin mass regardless of season. Vine vigor (pruning weight, cane diameter, and main leaf area) was significantly reduced in D vines (2008–2009 data), whereas vine capacity as total leaf area per vine was not. The leaf-to-fruit ratio dropped dramatically after defoliation to 1 m2/kg in D vines, which recovered thereafter and had a higher ratio from veraison onward. Intrinsic WUE and tolerance to photoinhibition increased in D vines for both main and lateral leaves, which were formed after leaf stripping and which had reached full maturity by the time measurements were made. Berry sugaring was accelerated in D vines, which also showed, at harvest, higher must Brix and phenolic and anthocyanin concentrations than C vines as well as more stable anthocyanins in the wine.
In Italy, over the last decade, ~250,000 ha of vineyards have been renewed under fairly strict rules established by the European Union, which require that new plantings should have medium-to-high vine densities (≅4000 to 6000 vines/ha). For high-yielding cultivars (i.e., high node fertility associated with large clusters), such rules have given rise to the conflict that the yield limits required by law cannot be complied with unless massive shoot and/or cluster thinning is carried out (Guidoni et al. 2002, Reynolds et al. 2005), which is costly and not always effective in improving grape composition (Keller 2010). A paradox arises when, despite the low yields resulting from these crop-adjustment techniques, grape composition does not show the expected improvement simply due to excessive vegetative growth; rather, undesirable leafy flavors may occur (Chapman et al. 2004).
Identification of the best approach for grape improvement requires a better understanding of current constraints. Vine growth and yield are dependent not only on CO2 fixation capability but also on the integrated processes of carbon allocation, accumulation, and utilization. Short-term experiments on the direct consequences of canopy manipulation on yield may overlook components of the vine carbon budget, and hence misrepresent the long-term consequences.
Several articles have been published on the effects of preflowering leaf removal, either manual or mechanical, on crop regulation and final grape and wine quality (Poni et al. 2006, 2008, Intrieri et al. 2008, Diago et al. 2009). While these studies were carried out in largely different environments and on different genotypes, the results showed that significant yield reduction and improved grape composition were achieved almost systematically. This response implies that the physiological control imposed on the vine through early leaf removal is dominant over other variability factors, which is remarkable as the final outcome of any summer pruning operation is generally unpredictable because of the complexity and dynamics of the factors involved.
Yet, constancy and repeatability of the effect on crop regulation due to early leaf removal are not surprising. Carbohydrate supply at the preflowering stage is the primary regulator of subsequent fruit set (Coombe 1962) and a temporary foliar stress may reduce cell division rates during the green stage of berry growth, which negatively affects final berry size (Palliotti et al. 2009). In almost all cases, this technique has also shown significant improvement in final grape composition and greater wine appreciation. Among the several seasonally related regrowth and compensation mechanisms triggered, the most widely investigated is post-defoliation photosynthesis recovery. At the shoot or whole-canopy level (Poni et al. 2008), the defoliated vines regain a photosynthetic capacity that is similar to non-defoliated vines around the time of veraison. Subsequently, the vine might benefit from higher late season assimilation rates due to the younger foliage. Given the yield restriction due to early leaf removal, it has been shown, using a whole canopy approach, that the seasonal amount of assimilates per unit of crop made available for ripening is higher in the defoliated treatments, which has been proposed as a primary cause for the improved grape composition (Poni et al. 2008). In contrast, basically no information has been published on the effect that early leaf removal has on water use efficiency or leaf chlorophyll fluorescence parameters or on the effects that repeated use of such techniques might have on vine vigor and capacity and carbohydrate accumulation and replenishment. The main objective of the present study was to fill these gaps and to evaluate such effects in association with vineyard performance in terms of vegetative growth, yield, and grape and wine composition in the red grape cultivar Sangiovese.
Materials and Methods
Plant material and experimental layout
This study was carried out from 2007 to 2009 in a nonirrigated vineyard in central Italy, near Perugia (42°58′N; 12°24′E, elevation 405 m asl, loamy soil type). The vineyard was an 8-year-old planting of Vitis vinifera L. cv. Sangiovese (clone VCR30 grafted to 420A rootstock) trained to a vertical shoot-positioned, spur-pruned cordon trellis with a bud load of ~10 nodes per meter of row length. Vine spacing was 2.5 m × 1.0 m (inter- and intrarow) and the cordon was trained 90 cm aboveground with three pair of surmounting catch wires for a canopy wall extending ~1.2 m above the cordon. Pest management was carried out according to local standard practices and shoots were mechanically trimmed when most started to outgrow the top wire. Depending on the year, shoot trimming was performed between 167 and 177 days of year (DOY).
Ten adjacent rows were selected to build a complete randomized block design, with each row designated as a block. Within each row, four uniform vines were tagged and randomly assigned in pairs to hand-remove all basal leaves on all shoots up to and including node 2 above the distal cluster (D = defoliated vines) or the leaves were left (C = control vines). Each year, the defoliation treatment was applied on the same vines at stage H (“flowers separated” according to Baggiolini 1952), which occurred on 14, 16, and 22 May in 2007, 2008, and 2009, respectively. The weather conditions during the study were monitored by an automatic meteorological station located near the vineyard.
Gas exchange and chlorophyll fluorescence measurements
In 2007, beginning one day before defoliation (DOY 134) single leaf gas exchange readings in both D and C vines were taken in the morning (0900 to 1100 hr) on clear days at varying intervals until harvest using a portable, open system, LCA-3 infrared gas analyzer (ADC BioScientific Ltd., Herts, UK). The system had a broad leaf chamber with a 6.25 cm2 window, and all readings were taken at ambient relative humidity with an air flow adjusted to 350 mL min−1. For each treatment, 12 fully expanded leaves at nodes 3, 8, and 14 above the distal cluster on a main shoot and 12 fully expanded leaves at node 3 from laterals were sampled under saturating light (PAR > 1400 μmol m−2 s−1). The assimilation rate (A), transpiration rate (E), and stomatal conductance (gs) were calculated from inlet and outlet CO2 and H2O relative concentrations. Intrinsic water use efficiency (WUEi) was then calculated as A/gs.
On the same leaves from node 14 above the distal clusters and laterals, chlorophyll fluorescence was measured with a light-weight portable continuous excitation fluorimeter (Handy-PEA, Hansatech Instruments, Norfolk, UK). Dark adaptation was achieved by covering the sample area to be analyzed with a small, lightweight leaf clip for at least 20 min. The small shutter plate of the clip was then opened and the dark-adapted leaf tissue was exposed to an actinic light flash (650 nm wavelength, intensity > 3000 μmol m−2 s−1). The instrument provides the Fv/Fm ratio, which is a widely accepted indicator of the maximum efficiency of photosystem II (PSII), where Fm is the fluorescence maximum over the induction curve. Fv (termed variable fluorescence) was calculated as the difference between Fm and Fo, where Fo is the ground fluorescence (Strasser et al. 1995). The area above the fluorescence curve between Fo and Fm (Area), which indicates the pool size of plastoquinone on the reducing size of PSII, was also automatically calculated.
Yield and grape composition
During ripening, the soluble solids concentration in the must (Brix) was periodically assessed on 80-berry samples (four samples per treatment and measurement date) using a temperature-compensating refractometer (RX-5000; Atago Co. Ltd., Tokyo, Japan). Harvest dates were 12, 17, and 14 Sept in 2007, 2008, and 2009, respectively. Experimental vines were individually picked, and crop weight, including both primary and secondary clusters, and cluster number per vine were recorded. Incidence of bunch rot was assessed as percentage of affected clusters per total, where affected clusters were designated as those having a minimum of 10 berries showing visual symptoms. Compactness was visually estimated on 100 clusters per treatment (five from each vine replicate) using OIV code 204 (OIV 1983). Number 1 indicates “berries in grouped formation with many visible pedicels” and number 9 indicates “misshapen berries.” The same clusters were used to calculate the ratio of total berry weight/main rachis length.
Each year, a 20-berry sample/vine (400 berries/treatment) was randomly collected from the harvested primary (basal) clusters, and skin, pulp, and seed weights of each berry were measured and the number of seeds per berry was counted. After individual weighing, the berries were sliced in half with a razor blade and the seeds and flesh were removed from each berry half with a small metal spatula without rupturing any pigmented hypodermal cells. Seeds were then carefully separated by hand from the flesh. Both skins and seeds were rinsed in deionized water, blotted dry, and weighed. The remaining flesh was used to make a juice sample to measure the soluble solids (Brix), pH, and titratable acidity (TA). TA was measured with a Titrex Universal Potentiometric Titrator (Steroglass S.r.l., Perugia, Italy); 0.1 N NaOH was used to obtain a pH 8.2 end point, expressed as g/L tartaric acid equivalent. The total anthocyanin and phenolic contents were determined according to Iland (1988) on a 50-berry sample per vine replicate randomly collected from the harvested primary clusters. Total anthocyanins and phenols are expressed as mg per berry and per kg of fresh berry mass.
Bud fertility, vine growth, and carbohydrate storage
In 2008, 2009, and 2010 at the end of May, bud fruitfulness was estimated as number of clusters per shoot. In 2007 and 2008, at 25 days before leaf removal, defoliation, flowering, veraison, and harvest, five fruiting shoots per treatment were taken from extra vines planted in rows adjacent to the experimental plots and the total leaf area per shoot, with the contribution of primary and lateral leaves kept separate, was measured with a surface area meter (model AAM-7; Hayashi-Denko, Tokyo, Japan). Total leaf area per vine was then estimated on the basis of mean shoot area assessment and shoot counts per vine. Concurrently, five clusters per treatment were removed from the same shoots used to estimate the total leaf area and immediately weighed to obtain a seasonal trend of the leaf area-to-yield ratio.
In 2008 and 2009, 10 wood samples per treatment were taken from mature canes about one month before budbreak from node 3 (five samples) and node 14 (five samples). Alcohol-soluble sugars and starch concentrations were determined according to a colorimetric method (Loewus 1952) using the anthrone reagent (Merck, Darmstadt, Germany). Absorbance readings at 620 nm wavelength were taken using a LKB Ultrospec III spectrophotometer (Pharmacia, Uppsala, Sweden).
Microvinification and wine analysis
In 2007 and 2008, experimental wines were produced on a microvinification scale. At harvest, grapes from all D and C vines were harvested manually and transported to the experimental winery in 20-kg plastic boxes. For each treatment, the total harvested grape mass was divided into two lots (30 and 45 kg per lot according to year) and mechanically crushed, destemmed, transferred to 50-L stainless-steel fermentation containers, sulfited with 35 mg/L SO2, and inoculated with 35 mg/L of a commercial yeast strain (Lalvin EC1118; Lallemand, Ontario, Canada). Wines were fermented for 14 to 18 days on the skin and punched down twice daily, with the fermentation temperature ranging from 20 to 28°C. After alcoholic fermentation, wines were pressed at 0 Brix and inoculated with 30 mg/L Oenococcus oenii (Lalvin Elios 1 MBR; Lallemand). After completion of malolactic fermentation, the samples were racked and transferred to 30-L steel containers and 25 mg/L SO2 was added. Two months later, the wines were racked again, bottled into 750-mL bottles, and closed with cork stoppers. After eight months, wines were analyzed for alcohol, TA, and pH (Iland et al. 1993). Wine color intensity (OD420+OD520), color hue (OD420/OD520), and total phenol and anthocyanin concentrations were determined by spectrometry. Total phenols were quantified according to Ribéreau-Gayon (1970) by measuring the absorbance at 280 nm of wine diluted 1:100 with distilled water. Anthocyanins were analyzed as reported elsewhere (Ribéreau-Gayon and Stonestreet 1965), and total tannins were quantified by precipitation with methyl-cellulose (Montedoro and Fantozzi 1974). Individual anthocyanin analysis was performed using HPLC (Mazza et al. 1999). All determinations were carried out in duplicate.
Statistical treatment
Vine performance data were subjected to a two-way analysis of variance using the SigmaStat software package (Systat Software, San Jose, CA). Year was considered as a random variable and the error term for the treatment factor was the year × treatment interaction mean square. Treatment comparison was performed each year by the t-test at p ≤ 0.05. Visual ratings of bunch compactness were subjected to square root transformation before analysis. Gas exchange and chlorophyll fluorescence parameters, seasonal evolution of total soluble solids, and leaf-area-to-yield ratio data are shown as means ± standard error.
Results
Environmental conditions
Heat accumulation values calculated as growing degree days (GDD) from 1 Apr to 30 Sept for each of the three seasons were considerably higher than the historical 55-year average extending from 1951 to 2006 (Table 1). The warmest season was in 2009 with 1963 GDD; the T mean and T max during the summer was higher than in 2008 and, with the exception of T max in July, was also higher than in 2007. The rainfall summation over the same period was lowest in 2007 (204 mm) with no rain in April, compared with the fairly wet springs in 2008, 2009, and the 55-year average. Despite such weather trends and the absence of irrigation, no visual symptoms of water stress or leaf yellowing were observed throughout the trial seasons.
Gas exchange and chlorophyll fluorescence
Post-defoliation leaf assimilation (A) rates measured on the D vines on primary leaves of defoliated shoots at nodes 3, 8, and 14 above the distal cluster showed no significant differences compared to the C leaves (Figure 1B, C, D), whereas the lateral leaves from the D vines had significantly higher A rates except for the end season data set (Figure 1E). Intrinsic WUE was enhanced in both the upper (i.e., nodes 8 and 14) primary leaves and the lateral leaves of D vines, especially at DOY 210 and 234 following the onset of veraison. A milder response was observed in basal primary leaves (Figure 1F, G, H, I). Regression analyses of gs versus A for different leaf positions and data pooled over DOY 198, 210, and 234 (Figure 2) indicate that lateral leaves on the D vines had higher A rates than leaves on the C vines regardless of the degree of stomatal opening (Figure 2D). For main leaves located at nodes 8 and 14 above the distal cluster, such an effect was manifest for a gs range from 100 to 150 mmol m−2 s−1 (Figure 2B, C).
Growing degree days (GDD, base 10°C) and rainfall from 1 Apr to 30 Sept at the experimental site in 2007, 2008, and 2009; mean values over 55 years (1951–2006); and mean (T mean) and maximum (T max) air temperatures (°C) recorded in July, August, and September 2007, 2008, and 2009.
Seasonal trends of air vapor pressure deficit (VPD) and photosynthetic active radiation (PAR) (A), assimilation rate and intrinsic water use efficiency in primary leaves at nodes 3 (B and F), 8 (C and G), and 14 (D and H) above the distal cluster and in medial leaves from lateral shoots (E and I) recorded in 2007 in control and defoliated Sangiovese vines. Bold arrow indicates timing of defoliation. Data are means ± SE (n = 12).
In D vines, from one month before to one month after veraison, the Fv/Fm ratios in leaves formed after the defoliation treatment were significantly higher than those in C vines; the values tended to converge at the preharvest reading (Figure 3). When chlorophyll fluorescence parameters were analyzed for two dates during veraison (DOY 198 and 210) under high transpiration demand (VPD and PAR higher than 2.2 kPa and 1900 μmol photons m−2 s−1, respectively; Figure 1A), the Fv/Fm value declined in both primary and lateral leaves of the C vines. This behavior could be attributed to a significant drop in Fm rather than to an increase in Fo (Table 2). Moreover, the Area parameter, which indicates the pool size of plastoquinone, increased significantly in the primary and lateral leaves (+39 and +31% for data pooled over DOY 198 and 210, respectively) of D vines compared with the homologous leaves of C vines (Table 2).
Yield component and grape composition
While bud fruitfulness in the following season (estimated as clusters per shoot) was not affected by early leaf removal, defoliation reduced yield per vine from a minimum of 24% in 2007 to a maximum of 39% in 2008 because of smaller clusters (−35% compared with C vines over the three years) (Table 3). Berry mass and berry number per cluster responded differently to the leaf removal. In 2007, only berry mass was drastically limited (−32%); in 2008 both components had a significant reduction, and in 2009 only berry number per cluster was affected (−30%). Overall, these combined effects were great enough to significantly reduce cluster compactness each season which, in turn, resulted in fewer incidence of botrytis rot (Table 3). Relative skin mass was significantly increased in D vines each season, whereas seed number and mass per berry were significantly reduced in 2008 and 2009 (Table 3).
Relationship between stomatal conductance (gs) and assimilation rate (A) for primary leaves at nodes 3 (A), 8 (B), and 14 (C) above the distal clusters and for leaves from lateral shoots (D) recorded in 2007 in control and defoliated Sangiovese vines. Within treatment, regressions were calculated for data pooled over DOY 198, 210, and 234. Linear model fitted to the leaf data from the 3rd node above distal clusters for control and defoliated treatments gave y = 4.75 + 0.016 x, R2 = 0.63, and y = 2.96 + 0.026 x, R2 = 0.86, respectively; from the 8th node above clusters gave y = 1.94 + 0.048 x, R2 = 0.82, and y = 4.14 + 0.045 x, R2 = 0.71, respectively; from the 14th node above the clusters gave y = 3.76 + 0.034 x, R2 = 0.33, and y = 4.11 + 0.032 x, R2 = 0.31, respectively; and from lateral shoots gave y = 3.37 + 0.027 x, R2 = 0.79, and y = 4.22 + 0.036 x, R2 = 0.87, respectively.
Seasonal trends of the Fv/Fm ratio in primary leaves at node 14 (A) above the distal cluster and in medial leaves from lateral shoots (B) recorded in 2007 in control and defoliated Sangiovese vines. Data are means ± SE (n = 12).
Early defoliated vines had significantly higher soluble solids concentration (Brix) at harvest each year (Table 4); the three-year average sugar gain was 2.7 Brix higher than in C vines. Juice titratable acidity and pH were not affected. Dynamics of soluble solids accumulation by the berry for 2007 and 2008 showed a faster build up in D vines (Figure 4). Regardless of season, the total concentrations of anthocyanins and phenolics (mg/kg) were significantly increased in D vines, while treatment differences diminished or were cancelled out when data were expressed on a per berry basis (Table 4).
Vine growth and replenishment of carbohydrate reserves
Vine vigor expressed as one-year-old pruning weight was reduced in D vines by 16% in 2007 and 22% in 2008, which correlates with a decrease in cane diameter (Table 5). Conversely, vine capacity as final total leaf area did not differ between treatments, because of compensating growth of lateral shoots in D vines (+53% in 2007 and +37% in 2008 compared with C vines). In both years, in D vines, starch concentration of wood samples taken from basal and apical cane zones was ~25 to 35% higher than in C vines; conversely, soluble sugars were either unchanged (node 14) or reduced (node 3) in the early season defoliated canes (Table 5). Estimated seasonal leaf area-to-yield ratios showed an expected drop to ~1 m2 kg−1 upon defoliation (Figure 5); thereafter, a substantial recovery was recorded until flowering, while from veraison onward, leaf area-to-yield ratios of D vines outweighed those estimated on C vines.
Chlorophyll fluorescence parameters in leaves at node 14 of primary shoots and in leaves from laterals in Sangiovese vines subjected to early defoliation (D) or control (C) at 198 and 210 day of year (DOY) in 2007.
Wine characteristics
For each year, wines from D vines had higher alcohol, total anthocyanin, phenolic, and total tannin concentrations than C vines, as well as more intense color and better color hue, whereas no changes were found in total acidity and wine pH (Table 6). Different anthocyanins in D vines showed an expected significant increase compared with C vines, especially malvidin-3-glucoside (+19 mg/L in 2007 and +28 mg/L in 2008) and petunidin-3-glucoside (+5.3 mg/L in 2007 and +8 mg/L in 2008) (Table 6).
Discussion
Seasonal post-defoliation assessment of A rates of main leaves formed either before (node 3 above distal cluster) or after (node 8 and 14 above distal cluster) leaf removal showed negligible photosynthetic compensation, confirming the variability of such a response in both intensity and duration (Hofäcker 1978, Candolfi-Vasconcelos and Koblet 1991, Poni and Giachino 2000). In the present study, the much greater photosynthetic recovery found in newly formed lateral leaves of D vines likely relates to their more advanced maturity, as can be inferred from total leaf area per vine partitioning showing that approximately one-half of the foliage was accounted for by lateral leaves as compared with one-third in C vines. In Cabernet Sauvignon, a three-fold variation in A rates was measured in lateral leaves at veraison, depending on the degree of vigor (Poni and Giachino 2000).
Yield components and cluster and berry characteristics recorded over 2007–2009 in Sangiovese vines subjected to early defoliation (D) or control (C).
Grape composition at harvest recorded over 2007–2009 in Sangiovese vines subjected to early defoliation (D) or control (C).
Seasonal trends of total soluble solids accumulation in control and defoliated Sangiovese vines for 2007 and 2008. Data are means ± SE (n = 4).
Vegetative growth, vine balance indices, and cane wood reserves recorded over 2007–2008 in field-grown Sangiovese vines subjected to early defoliation (D) or control (C).
Early leaf removal enhanced WUEi in both main and lateral leaves that were formed after leaf stripping and that had reached full maturity by the time measurements were made. As such, these findings are different from those reported elsewhere (Candolfi-Vasconcelos et al. 1994), where there were no differences in the assimilation to transpiration ratio measured on main and lateral leaves maintained on severely defoliated Pinot noir shoots. This discrepancy could be due to either the fairly late timing of leaf removal applied by these authors (pea size, six weeks after flowering) or the fact that all readings were mainly taken on leaves fully grown by the time of leaf removal. In the present study, despite a similar chronological age, it is apparent that readings taken postveraison on apical or lateral leaves of defoliated shoots showing higher WUEi than corresponding leaves sampled on non-defoliated shoots would suggest the validity of a “transpiration stream” partitioning hypothesis (Flore and Lakso 1991). According to this hypothesis, a severe source reduction (by shading or leaf removal) would allow the currently or newly developing leaves to benefit from an increased root supply of water, nutrients, and hormones, generating in turn, when mature, a more efficient leaf tissue. Interestingly, main leaves inserted at node 3 above distal clusters (i.e., already expanded by the time defoliation was performed) had higher A rates in D vines than in C vines on readings taken DOY 198 at high gs (350 to 500 mmol m−2 s−1). As such performance diminished the two following measuring dates, it is likely that it was still an effect of photosynthesis compensation, which is typically temporary (Candolfi-Vasconcelos and Koblet 1991).
Seasonal evolution of leaf area-to-yield ratio in control and defoliated Sangiovese vines for 2007 and 2008. Data are means ± SE (n = 20).
Wine composition and anthocyanin content recorded over 2007–2008 vintages in Sangiovese vines subjected to early defoliation (D) or control (C). Analyses were performed at the end of May (8-month-old young wines).
Even if no apparent drought was observed during the summers of 2007, 2008, and 2009, a reversible photoinhibition phenomenon occurred in the younger leaves of C vines around veraison, with the Fv/Fm yield dropping below 0.75 due to a large reduction in Fm rather than an increase in Fo. This behavior indicates a non-photochemical quenching of xanthopylls, thus a diminished nonradiative energy dissipation, rather than photodamage of PSII (Krause and Weis 1991). Greater tolerance to photoinhibition shown by the younger leaves of D vines could be attributable to changes in total carotenoids, since these pigments surely have a positive role in acquiring tolerance as indicated by several studies on xanthophylls (Leipner et al. 1997, Niyogi et al. 1997), the major components of total carotenoids. The reduction in Fv/Fm of leaves from C vines was also associated with a significant decrease in the plastoquinone pool (as indicated by lowering Area parameter) and a lower efficiency of the PSII centers, as previously reported for Tempranillo leaves (Maroco et al. 2002).
Although early leaf removal resulted in stripping off ~80% of the total shoot leaf area of the vines, the fruitfulness of the basal nodes in the following season, expressed as clusters per shoot, was not affected, as also shown in previous early defoliation studies, despite a sudden drop in the net CO2 exchange rate of the whole canopy up to 75% compared with the control vines (Poni et al. 2008). Based on previous studies (Perez and Kliewer 1990, Sanchez and Dokoozlian 2005), it is feasible that the expected decline in bud induction due to the unfavorable source balance might have been fully counterbalanced in D vines by increased local light exposure to basal nodes retained with the following winter pruning.
The extent and repeatability of yield per vine reduction due to defoliation adds to earlier findings for other Vitis vinifera varieties such as Trebbiano (Poni et al. 2006) and Tempranillo (Diago et al. 2009), supporting a robust physiological control of such techniques on yield components such as percentage of fruit set and/or berry size. Although the relative impact of defoliation on fruit set and berry size varied considerably across seasons, the overall effect on cluster morphology was toward a drastic reduction in compactness, which is an important step toward acquired tolerance to rot. Yet, in examining berries per cluster, the marked decrease recorded in D vines in years 2 and 3 (Table 3) suggests that this parameter might have been buffered by existing reserves the first season of treatment application, while reiteration of D treatment on the same vines has likely triggered a cumulative source-limitation leading to heavier effects on fruit set. Although an increase in relative skin mass as a response to preflowering leaf removal had preliminarily been reported for cv. Barbera (Poni and Bernizzoni 2010), the magnitude and consistency of such modification in this study were remarkable. Most interestingly, a relative skin mass increase of at least 3.6% was maintained each season in D vines regardless of variation in total berry mass, ranging from 1.6 g in 2007 to 2.06 g in 2009. It is therefore confirmed (Roby and Matthews 2004) that differential growth of various berry organs is preeminent toward expectations driven from the geometrical law stating that the surface-area-to-volume ratio of approximately spherical berries decreases with berry size according to 3/R, where R is the radius of the sphere, and that larger berries can still maintain a quite favorable relative skin mass growth. These findings strengthen previous hypotheses that the early abrupt and long-lasting cluster microclimatic change caused by leaf removal would favor both cell division over the first part of the green stage of berry growth and berry skin thickening as a long-term adaptation mechanism (Poni et al. 2009).
Regardless of season, final soluble solids and total phenolic and anthocyanin concentrations of berries were decidedly higher in D vines, and the dynamics of berry sugar accumulation for 2007 and 2008 showed a faster build-up in D vines beginning from veraison. This last feature correlates well with the seasonal changes in the leaf-to-fruit ratio, which in both years confirmed veraison as a switching date beyond which more leaf area per unit of crop was made available for ripening in the D vines. However, advanced ripening in the D treatment also reflects a different vegetative growth pattern than in C vines. Controlled vigor usually aids more efficient and/or anticipated solute accumulation in the berry due to lessened competition with vegetative sinks (Winkler et al. 1974). As determined by the weight of one-year-old wood removed by winter pruning as well as cane diameter data, in both years, vine vigor was significantly reduced in the D vines. The causal mechanism is again linked to the dynamic changes of the leaf-to-fruit ratio which, at flowering, was still halved in the D vines as compared with C vines. Interestingly, such vigor containment had no adverse impact on vine capacity as represented by final total leaf area per vine (Winkler et al. 1974), indicating that the early defoliation technique has potential for more efficient vegetative growth at no expense to the total pool of carbohydrates available for ripening.
Notably, improved grape and wine composition achieved with defoliation took place at no detriment to replenishment of reserves in storage organs. One month before budbreak in 2008 and 2009, D vines had starch concentration in both basal and apical zones of the canes that were 32 to 36% and 25 to 30% higher than amounts recorded in C vines, respectively, whereas soluble sugars significantly decreased only at basal node 3. While this latter finding might reflect the possible need of the defoliated vine to use readily available reserve pools, higher late season leaf-to-fruit ratio and A rates of D vines were conducive to higher carbohydrate availability per unit of crop in the postveraison period in the D vines, in turn generating more efficient reserve replenishment compared with non-defoliated vines.
D vines resulted in both greater potential for anthocyanin accumulation in the berry and better capability of carrying the color into the wine over time. This finding is especially significant, as the Sangiovese cultivar typically suffers from a fairly low ability to develop and accumulate anthocyanins compared with other red cultivars (Mattivi et al. 2002). Furthermore, color accumulation in Sangiovese berries is highly sensitive to crop level and seasonal weather conditions (Palliotti and Cartechini 1998), therefore increasing year-to-year variability. Different reports have shown reduced anthocyanin accumulation in the skin of grape berries under high temperature, which results from anthocyanin degradation and the inhibition of mRNA transcription of the biosynthetic genes of the anthocyanins (Spayd et al. 2002, Yamane et al. 2006, Mori et al. 2007). A practical guideline drawn from these studies is that leaf stripping should be used cautiously to avoid berry overheating and sunburning. Results shown in the present study suggest that such recommendation should be more precisely calibrated toward timing of leaf removal, as full stripping of basal leaves at preflowering actually enhanced color and phenolic accumulation on a concentration basis. The critical stage of development for a response of color build-up to temperature has not been established, although an in vitro study has indicated that a key period might be the two weeks following the first visual indicator of veraison (Yamane et al. 2006). Albeit no seasonal monitoring of cluster microclimate upon defoliation was performed, it is likely that by the time that onset of veraison occurred, clusters borne on D shoots benefitted from new partial leaf cover due to the development of laterals from basal and midshoot nodes.
Determinations made at 8 months of wine aging of chiefly stable anthocyanins such as malvidin-3-glucoside (+22% in 2007 and +43% in 2008 compared with C vines), petunidin-3-glucoside (+34% in 2007 and +62% in 2008), and delphinidin-3-glucoside (+18% in 2007 and +78% in 2008) support the higher aging capability of the Sangiovese wines produced from D vines. This capability is likely due to a higher degree of copigmentation and/or lower degradation sensitivity (Boulton 2001).
Conclusions
Early defoliation applied before flowering with elimination of ~80% of the leaf area present on the shoots was shown to be a reliable technique. This reliability is especially important where legal limiting of yield is associated with traits of high vegetative vigor, high node fruitfulness, and large and compact clusters. Seasonal dynamic changes in the crop load brought about by early leaf removal increased canopy efficiency, especially during the veraison ripening period, and induced better performance of WUE and tolerance to photoinhibition. Such features, together with the ability to control vigor at no detriment to vine capacity and wood carbohydrate replenishment, are the physiological bases for improved grape and wine composition regardless of year-to-year variability.
Acknowledgments
Acknowledgments: The authors thank the Regione dell’Umbria for partial financial support and Mary Frances Traynor for language revision.
- Received August 1, 2010.
- Revision received December 1, 2010.
- Accepted January 1, 2011.
- Published online December 1969
- © 2011 by the American Society for Enology and Viticulture
Literature Cited
Vol 62 Issue 2
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
