Abstract
The carbohydrate (CHO) reserve physiology of Chardonnay grapevines growing in the cool climate of New Zealand was examined in relation to subsequent flowering and fruiting using defoliation treatments. Vines were defoliated by removing all but the four basal leaves from shoots at monthly intervals starting at four weeks postbloom. Throughout the following season, CHO reserves in the roots and trunks were measured and detailed assessments of vine flowering and yields were recorded. Previous season’s vine defoliation decreased concentrations of overwintering CHO reserves (mostly starch) in both roots and trunks, with earlier defoliation times resulting in the largest reductions. Roots were most sensitive, with early defoliation reducing starch concentrations at budburst to 1.5% DW compared with 17% DW in non-defoliated vines. Reductions in root and trunk CHO reserves were closely associated with significant decreases in inflorescence number per shoot and flower number per inflorescence (up to 50% less than in control vines). Differences in CHO concentrations in both the root and trunk were maintained to veraison, but at flowering had no effect on the percent fruit set in the first season after defoliation. Therefore, lower yields in the following season were caused by fewer inflorescences per shoot and flowers per inflorescence. Shoot growth and total vine pruning weight were also decreased in vines where CHO reserves were reduced. These findings suggest that restricted CHO reserve accumulation as a consequence of defoliation may have a negative impact on subsequent grapevine flowering and productivity, particularly so in a cool-climate environment where there is no postharvest CHO reserve accumulation period.
Marked differences in grape yields can be observed between seasons. Mareschalchi recorded weather and yields of vines in Northern Italy from 1855 to 1907 and concluded that by taking into account weather conditions and vine yield in the previous 12 months, yields in the current season could be predicted with a fair degree of accuracy (Perold 1927). However, while the impacts of temperature and light, in particular at the time of inflorescence initiation and flowering, are understood (Buttrose 1974, Ezzili 1993, May 1965, May and Antcliff 1963, Palma and Jackson 1981, Sommer et al. 2000), the role of overwintering carbohydrate (CHO) reserves in the vine is less clear. Previous studies have illustrated that root and trunk CHO reserves are the source of CHO utilized in the development of new shoots and inflorescences in the following spring (Murisier and Aerny 1994, Scholefield et al. 1978, Yang and Hori 1979, 1980). In many grapegrowing regions, leaves are retained by vines postharvest, allowing an accumulation of CHO reserves before winter (Williams 1996). In cool climates such as in New Zealand, where leaf senescence often coincides with fruit harvest, maintaining appropriate CHO reserves in the vine during the growing season is particularly important.
The accumulation of CHO in vine reserves depends on the rate of photosynthesis and the partitioning of that photosynthate between shoot, root, and fruit growth and storage (Howell 2001). Previous studies have indicated that reducing CHO production during the growing season by defoliation can reduce inflorescence number, cluster size, and vine yield in the following season (Candolfi-Vasconcelos and Koblet 1990, Koblet et al. 1993); however, these studies have not established the nature of the relationship between changes in CHO reserves and subsequent vine flowering and productivity. Low CHO reserves can also reduce winter hardiness (Hunter et al. 1995, Koblet 1996, Scholefield et al. 1977, Sommer et al. 2000). In addition, while reduced photosynthesis at flowering lowers fruit set (Ferree et al. 2001, Keller and Koblet 1994), the possible role of CHO reserves in this process is unclear. To examine and characterize the relationship between overwintering CHO reserves and the following season’s vine flowering and yield in a cool-climate environment, a defoliation trial was conducted on mature Chardonnay vines. Defoliation at three times during the season was used to produce vines with a range of CHO reserves. The impact on vine growth, flowering, and yield in the subsequent two seasons is reported and discussed.
Materials and Methods
Defoliation.
Own-rooted Chardonnay vines (13-year-old Mendoza clone) were grown in the Lincoln University Vineyard, Canterbury, New Zealand, using a double-Guyot, vertical shoot-positioned (VSP) training system. Vines were defoliated on three occasions at monthly intervals starting at four weeks postbloom in the 1997/1998 season. A non-defoliated treatment was used as the control. Vine defoliation treatments removed leaves from node position five to the apex of the shoot, while retaining the basal four leaves on each shoot of the vine (~75% defoliation). The basal leaves were left to protect the clusters from sunburn and provide some leaf area to keep the vine alive. Each treatment was replicated six times in a randomized block design along a single row of grapevines in the vineyard. Shoot and lateral trimming was not performed, but leaves that developed during the course of the experiment were removed. Berry diameter and soluble solids, cluster numbers, and yield per vine were measured at harvest and pruning weight recorded once the vines were dormant.
Shoot, flower, and CHO measurements.
At the end of the defoliation season (1997/1998) the vines were pruned to two 10-node canes and two spurs. Inflorescence number was determined in two ways:
One shoot (suitable for tying down as a cane) from the prunings of each vine was cut into 10 single-node cuttings, which were then placed in trays of water in a heated glasshouse and left to grow, with unwanted leaf growth removed occasionally. Once sufficient shootgrowth had occurred, the number of inflorescences arising from each node was recorded.
The percentage budburst and shoot growth (length) on one cane per vine (~10 shoots) was measured on a twice weekly basis until close to bloom. A count of inflorescences on the developing shoots for the same cane was recorded one month after budburst.
The number of flowers per inflorescence was estimated by counting flowers on the first branch of the inflorescence and substituting that value into the following equation: Total flower number = 6.2 * (flowers on branch 1) + 8.2. The linear relationship (R2 = 0.92) between first branch and total inflorescence flower number was developed using a population of small- to large-sized inflorescences on nonexperimental vines in the same row (Bennett 2002). Four inflorescence flower number estimations per treatment vine were made by counting branch one flowers on inflorescences that arose from two prelabeled shoots per vine (two inflorescences per shoot, one shoot from each of two canes). The number of primary branches per inflorescence was also recorded as an indicator of flower cluster size.
Wood samples were taken from the trunks and roots at key phenological stages throughout the following season (1998/1999) starting at budburst (29 Sept 1998). Samples were removed from the midsection of the trunk using a 5-mm trunk corer (Kymen Pin and Implement, Kouvola, Finland). A section of root 1.0 to 1.5 cm in diameter was removed from near the base of each vine. The wood samples were handled and analyzed for CHO according to the methods described by Allen et al. (1974) and Rose et al. (1991), with minor modifications made to the time taken to heat samples and extracts and centrifuge speed (2500 rpm). Samples and extracts were heated for slightly longer periods (10 and 15 min) to allow for complete extraction and color reaction, respectively (Bennett 2002).
Fruit and pruning measurements.
Prior to vine harvest all the clusters on the labeled shoots were harvested, weighed, and berry number per cluster counted. Using estimated flower number and berry number, percentage fruit set was calculated for sample clusters. The balance of the fruit was harvested two weeks later and the weight of fruit and total cluster number per vine recorded. During dormancy vines were pruned to the standard VSP system as described earlier. The prunings were weighed and a subsample of four randomly selected lignified shoots (canes) per vine were measured for the following: length, weight, count of node number, and shoot diameter at the eighth internode.
Calculation of vine capacity (estimate of annual dry matter production) was performed using the following formula: vine capacity = 0.55(pruning weight) + 0.25(fruit weight), where 0.55 = percentage dry matter content of prunings and 0.25 = percentage dry matter content of fruit (Winkler et al. 1974). Fruit yield to pruning weight ratios (Ravaz index) were also calculated for each treatment vine (Ravaz 1930). Measurement of overwintering CHO reserves, floral and fruit components of yield, and vine yield were also recorded for the second season following original defoliation.
Statistical analysis.
All vine data from the defoliation experiment were analyzed using general ANOVA testing for polynomial significance using a GenStat statistical package (GenStat 5, release 4.; VSN, Herts, UK). Mean separations were determined using least significant difference (LSD) at the 5% level of significance. Polynomial regressions of scatterplots were also performed using GenStat.
Results
Current season’s fruit yield.
Defoliation had no effect on clusters per vine or mean cluster weight, despite a tendency for reduced berry size in early defoliations. A small but significant linear decrease in yield, fruit soluble solids, and winter pruning weight was induced by early defoliation treatments (Table 1⇓).
Effect of vine defoliation on current season’s vine yield components, juice soluble solids, and pruning weight (1997/1998).
Root and trunk CHO reserves.
Changes in both starch and sugar concentrations in the trunk and roots of non-defoliated control vines were observed during the season (Figures 1⇓ and 2⇓). Root starch and soluble sugar concentrations decreased rapidly after budburst during the period of rapid shoot development to bloom (0–80 days after bud-burst; DABB). These changes were matched by a decrease in trunk soluble sugar concentration. In all cases, concentrations returned to budburst levels by leaf fall. In contrast, trunk starch concentrations changed little during the season. In the earliest defoliation treatment, starch concentrations in the trunks and roots were very low and remained at these levels until 50% bloom (80 DABB), progressively increasing in concentration until they were similar to non-defoliated vines at 50% veraison (155 DABB) (Figures 1a⇓, 2a⇓). In all vines, trunk and root CHO concentrations reached their highest levels by the time of leaf fall (230 DABB) (Figures 1⇓ and 2⇓).
Effect of time of vine defoliation in the previous season on (A) starch, (B) soluble sugar, and (C) total nonstructural CHO concentration in trunks (1998/1999). Bars represent LSD at 5% level of significance.
Effect of time of vine defoliation in the previous season on (A) starch, (B) soluble sugar, and (C) total nonstructural CHO concentration in roots (1998/1999). Bars represent LSD at 5% level of significance.
Return bloom and yield components.
Defoliation treatments caused a reduction in inflorescence number per shoot in the following season, with the earliest defoliation nearly halving the number of inflorescences per shoot compared with no defoliation (Table 2⇓). The effect was consistent along the cane (Figure 3⇓), regardless of whether or not the leaves at basal node positions were retained in the previous season. Similar results were obtained from both the single-node cutting and field evaluation (Table 2⇓), although inflorescence number per shoot appeared to be slightly underestimated using the single-node cutting technique.
Effect of vine defoliation in the previous season on floral and fruit components of yield (1998/1999).
Effect of vine defoliation in the previous season on inflorescence number per shoot for node positions one to ten on the cane (from single-node cuttings).
Defoliation also reduced inflorescence size by reducing the number of primary branches per inflorescence and the number of flowers on the branches (Table 2⇑). In contrast, fruit set in the season following defoliation was higher in early defoliated vines than no defoliation, although that did not compensate for the lower flower and berry number, and cluster weights were lower where vines had been defoliated (Table 2⇑). Smaller clusters on previously defoliated vines, together with fewer clusters per vine, resulted in vine yields as low as 31% of non-defoliated vines.
Yield component and CHO relationships.
To examine the close association between CHO reserves and subsequent vine productivity in response to previous season’s defoliation CHO, inflorescence, flower, and yield data were integrated using polynomial regression analysis. Root starch concentration at budburst related to inflorescence number per shoot, estimated flowers per inflorescence, and vine yield in a linear manner (Figure 4⇓) (p < 0.001). Early defoliated vines with a root starch concentration of ~2% DW displayed reduced fruitfulness and yield (1.1 inflorescences/shoot, 149 estimated flowers/inflorescence, and 1.8 kg fruit/vine) compared with non-defoliated vines, which had a root starch concentration of 17% DW and significantly higher fruitfulness and yield (2.2 inflorescences/shoot, 258 estimated flowers/inflorescence, and 5.9 kg fruit/vine). Trunk starch concentrations also related to inflorescence number per shoot, estimated flower number per inflorescence, and vine yield in a linear fashion, with coefficients of determination of 0.66 (p < 0.001), 0.74 (p < 0.001), and 0.71, (p < 0.001), respectively (Bennett 2002). The interpretation of the regression relationships in regard to crop development is discussed later.
Potential regression relationships between root starch concentrations at budburst and (A) inflorescences per shoot, (B) estimated flowers per inflorescence, and (C) vine yield (1998/1999).
Vegetative growth and vine capacity.
Defoliation had no effect on percent budburst or shoot number per vine (Table 3⇓). At both 18 and 25 DABB there were no significant differences in shoot length between the four defoliation treatments (Figure 5⇓). Shoot growth slowed after this time for defoliated vines, but differences in cane length and node number were not significant by the end of the growing season (Table 3⇓). In contrast to cane length and node number, cane fresh weight and internode diameter were markedly reduced by defoliation (Table 3⇓).
Effect of vine defoliation in the previous season on budburst, final shoot growth parameters, vine balance, and capacity (1998/1999).
Effects of time of vine defoliation in the previous season on shoot length from budburst to prebloom (1998/1999). Bars represent LSD at 5% level of significance.
Ravaz index and vine capacity were also reduced by defoliation. The earliest defoliation treatment produced an index of 1.3 compared with 2.4 for non-defoliated vines (Table 3⇑). Vine capacity was halved by earliest defoliation; later defoliations resulted in less reduction in capacity (Table 3⇑). Root and trunk starch concentrations at bud-burst (29 Sept 1998) were shown by linear regression to relate strongly (R2 = 0.72, p ≤ 0.001 and R2 = 0.80, p ≤ 0.001, respectively) with vine capacity (Bennett 2002).
Return bloom and yield components in season 2.
A small but significant linear reduction in inflorescences per shoot, inflorescence flower number, fruit set, berry number, cluster weight, and yield in response to defoliation two seasons previously was still evident (Table 4⇓). Yield components in the 1999/2000 (Table 4⇓) season were considerably lower than those of the previous season (1998/ 1999) (Table 2⇑) but were similar to those of the 1997/1998 season (Table 1⇑).
Effect of vine defoliation two seasons previously on floral and fruit components of yield (1999/2000).
Discussion
Carbohydrate reserves.
Overwintering trunk starch concentrations of 11.4% and 2.5% DW have been reported in mature Carignane vines in California (Winkler and Williams 1945) and Pinot noir vines in Europe (Koblet et al. 1993), respectively. The present study found a similar concentration of 10% DW in mature non-defoliated Chardonnay vines. The wide range of starch concentrations reported to date most likely indicates that CHO reserves are influenced by grape variety and viticultural region (climate and vine management).
The significant reduction in trunk CHO reserves in response to previous season’s defoliation is consistent with published research. Both Candolfi-Vasconcelos and Koblet (1990) and Koblet et al. (1993) reported a 2.0 to 3.0% DW reduction in overwintering trunk starch concentration following partial vine defoliation. The larger reduction in starch concentration in the present study (4.8% DW decrease) (Figure 1a⇑) can be attributed to the more severe defoliation treatment that was used. Results also revealed that root CHO reserves at budburst (29 Sept 1998) were more sensitive to previous season’s defoliation than trunk CHO reserves, with the first defoliation treatment causing a 15.5% DW decrease in starch concentration (Figure 2a⇑).
Although a large reduction in leaf area (photosynthate supply) was the most significant cause of impaired CHO reserve accumulation in roots and trunks, it is possible that starch concentrations in the roots and trunks of the earliest defoliation (four weeks; Figures 1a⇑, 2a⇑) were also affected by remobilization of CHO reserves to ripening fruits during the season of defoliation. Candolfi-Vasconcelos et al. (1994) and Koblet et al. (1993) found that ripening fruit on severely defoliated vines initiated the remobilization of CHO reserves to allow for fruit maturation. Some evidence for fruit-induced remobilization, in the current study, is that berry soluble solids (Brix) in early defoliated vines was not significantly reduced compared with later defoliations during the season of defoliation treatment (Table 1⇑).
Carbohydrate reserve recovery ceased during the harvest/leaf fall period of the season following defoliation (Figures 1⇑ and 2⇑). In the roots of defoliated vines, recovery was accomplished by much higher rates of starch accumulation during the bloom to veraison period compared with non-defoliated vines (see slope of lines in Figure 2a⇑). The increased rate of starch accumulation in early defoliation treatments may have been aided by the considerably lighter crop loads carried by these vines (Table 2⇑). In trunks, recovery was accomplished by a combination of both starch and soluble sugar accumulation (Figures 1a and b⇑). In warm climates such as California, where post-harvest leaf photosynthesis occurs for several weeks, Williams (1996) has shown that starch and sugar continue to accumulate in the trunks of Thompson Seedless vines until leaf fall. In contrast, the findings of this study have shown in the cool climate of Canterbury, New Zealand there is no postharvest accumulation of CHO reserves in trunks and roots, as harvest and leaf fall often coincide (Figures 1⇑ and 2⇑). Therefore, it is suggested that, unlike grapevines growing in warm climates, CHO reserve accumulation must compete with ripening fruit for available photosynthates during the summer.
Yield components and vegetative growth.
As has been reported previously (Candolfi-Vasconcelos and Koblet 1990, Hunter and Visser 1990, Mansfield and Howell 1981, May et al. 1969), we have seen in this study a reduction in inflorescence number per shoot in response to previous season’s defoliation (Table 2⇑). However, both Mansfield and Howell (1981) and May et al. (1969) noted that complete shoot defoliation was required to reduce inflorescence number. Those findings appear to contrast with results presented here, where inflorescence number was reduced at all node positions along the cane (Figure 3⇑) regardless of the foliated/defoliated status of these nodes in the previous season. Such results suggest that a whole-vine effect, in response to previous season’s defoliation, was mediating the reduction in inflorescence number. A whole-vine effect may have also been responsible for the significant decrease in flowers per inflorescence, berries per cluster, and cluster weight observed in defoliated vines (Table 2⇑). Candolfi-Vasconcelos and Koblet (1990) reported that previous season’s vine defoliation reduced berries per cluster and cluster weight; however, they could not determine the cause or ascertain whether a partial or whole-vine effect was responsible. Results presented here (Table 2⇑, Figures 1⇑ and 2⇑) clearly showed that a reduction in flowers per inflorescence was the main mechanism for reducing berries per cluster and thus cluster weight and that a decline in CHO reserves may have been the whole-vine effect responsible for the reduction in flowers per inflorescence.
The preflowering impediment to shoot growth in response to the previous season’s defoliation and reduced CHO reserves (Figure 5⇑) did not continue through to end-of-season cane length (Table 3⇑). However, a significant reduction in cane fresh weight and diameter was associated with a decrease in pruning weights and, ultimately, vine capacity (Table 3⇑), suggesting a close link between available CHO reserves at the start of the season and vine growth and productivity throughout the growing season. This is important in a cool-climate environment because the opportunity for both compensatory growth and CHO reserve accumulation is limited by cooler temperatures and a shorter growing season.
Relationship between CHO reserves and yield components.
The proposed link between CHO reserves, subsequent vine fruitfulness, and productivity is not new (Thomas and Barnard 1937, Candolfi-Vasconcelos and Koblet 1990, Hunter et al. 1995, May 1965, Scholefield et al. 1977, Sommer et al. 2000). Unlike previous work, the present study has revealed positive relationships between starch concentrations at budburst in both roots (Figure 4⇑) and trunks (Bennett 2000) and the number of inflorescences per shoot and flowers per inflorescence subsequently formed in the spring. Although these findings provide evidence to support previous propositions, the interpretation of the regression relationships needs to be carefully considered in regard to the development of the crop from its very inception in the year previous to fruiting. The exact mechanism(s) of effect involved in the link between defoliation, CHO reserves, and floral components may be considered under two scenarios: (1) a direct effect during the season of defoliation, and (2) a carry-over effect into the season after defoliation.
Scenario 2 proposes that the carry-over effect is mediated by low overwintering CHO reserves. Low levels of CHO reserves impair inflorescence formation and development during and after budburst such that inflorescence number and size are reduced. Scenario 1 proposes that a reduction in CHO supply (photoassimilate) at the time of inflorescence initiation (during imposition of the defoliation treatments) directly impacts on inflorescence initiation, with the result that inflorescence number is reduced. Scenario 1 has been proposed by May (1965), Smart et al. (1982), and Sommer et al. (2000), but never clearly illustrated. These authors suggest that reduced CHO supply to latent buds (nodes), as a consequence of subtending leaf removal, during the summer initiation period reduces inflorescence primordia initiation and hence inflorescences per shoot in the following season. If this proposition is correct, then the expectation is that a reduction in inflorescence number is restricted to nodes that were defoliated in the previous season. However, results presented in Figure 3⇑ show a reduction occurred at all nodes regardless of the presence or absence of subtending leaves in the previous season. The single-node data presented in Table 2⇑ cannot be used as a definitive indicator of inflorescence number, given that they consistently underestimated actual vine fruitfulness. Given these results, attention is refocused on a whole-vine carry-over effect occurring (scenario 2). Evidence for such an effect is presented in Figure 4a and b⇑, where the close relationship between overwintering CHO reserves and grapevine fruitfulness suggests that inflorescence formation and development during the budburst period is dependent on an adequate level of CHO reserves. That makes sense in terms of vine developmental physiology, as both inflorescence branching and flower formation are determined over the budburst period (Barnard and Thomas 1933, May 2000). Previous work by Scholefield et al. (1977) has indicated that inflorescence size (flower number) may be dependent on overwintering CHO reserves, based on their findings that harvest pruning of Sultana vines in the previous season prematurely killed off 60% of a vine’s leaf area and resulted in fewer flowers per inflorescence in the following spring.
Carry-over effects into season 2.
Yield reductions, albeit small, in early defoliated vines two seasons after defoliation treatment were primarily attributed to a decrease in inflorescence and flower numbers; however, a significant decline in fruit set also contributed (Table 4⇑). The cause of the reduced fruit set is unclear, but less favorable weather conditions during the 1999/2000 season (Bennett 2002) may have been responsible, as fruit set in general was considerably lower than in the previous season (Table 2⇑). The small reduction in inflorescence and flower number could not be related to CHO reserves, but a link to the previous season’s shoot (current seasons’ canes) vigor (weight and diameter; Smart and Robinson 1991) was found. Significant (p ≤ 0.01) linear relationships illustrated that between 50% and 60% of the variation in inflorescence and flower number was attributed to shoot vigor (Bennett 2002). Such findings suggest CHO reserve stress in the previous season may lead to lighter and thinner canes that display reduced fruitfulness in the fruiting season.
Conclusions
It is proposed that overwintering trunk and root CHO reserves play important roles in determining the fruitfulness, productivity, and capacity of grapevines, in particular, the number of inflorescences and the number of flowers on each inflorescence formed. The mechanisms by which inflorescence development is altered are unclear at this stage, but data suggest that they may be linked to both inflorescence initiation and flower formation phases over two seasons. Additionally, shoot growth and the quality of replacement fruiting canes and their subsequent fruitfulness may be affected by the overwintering CHO reserve status of the vine. Thu, CHO reserve stress (depending on the severity) may have both direct and indirect impacts on grapevine flowering and yield over two seasons. More studies should be conducted to further validate the relationship between CHO reserves and subsequent vine flowering and productivity. Experiments in which overwintering root and trunk CHO reserves can be altered by other means, such as vine shading, trunk girdling, and photosynthesis inhibition, should be tested. With a restricted period of CHO reserve accumulation evident in cooler climates, vine management practices and stress events like excessive defoliation, early frost, or disease are likely to have a greater impact on the ability of grapevines to accumulate sufficient overwintering CHO reserves for the new season’s reproductive and vegetative growth.
- Received June 2005.
- Revision received June 2005.
- Revision received September 2005.
- Copyright © 2005 by the American Society for Enology and Viticulture
Literature Cited
Vol 56 Issue 4
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
