Bahar et al.

Auto-fluorescence for analysis of ripening

353

Auto-fluorescence for analysis of ripening in Thompson Seedless and colour in Crimson Seedless table grapes _199 353..359

A. BAHAR1, T. KAPLUNOV2, Y. ZUTAHY2, A. DAUS2, S. LURIE2 and A. LICHTER2 1

2

Selcuk University, Silifke Tasucu Vocational School, Mersin, Turkey Department of Postharvest Science, ARO, The Volcani Center, PO Box 6, Bet Dagan 50250, Israel Corresponding author: Dr Amnon Lichter, fax +972 3 9683622, email: [email protected] Abstract

Background and Aims: Ripening of table grapes is routinely followed by measurement of soluble solids, acidity, colour and firmness. Non-destructive means to measure ripening can be of great value to determine optimal harvest time and to compare the effect of treatments on ripening. Methods and Results: The portable fluorescence detector (Multiplex III, Force A, France) generates 12 signals which are processed to ratios which compensate for the structural complexity of the cluster. Sampling was carried out in a Thompson Seedless vineyard in four blocks and over 9 weeks. The simple fluorescence ratio (SFR_R) (correlated to chlorophyll level) decreased exponentially with an R2 value of 0.97, while flourscence excitation ratio (FER_RG) (correlated to anthocyanin level) displayed an increasing linear trend with an R2 of 0.98. The flavonoids (FLAV) ratio which was shown to correlate to the level of flavonoids increased during the first 5 weeks of veraison and then reached a steady level. In an experiment on Crimson Seedless grapes, application of abscisic acid (ABA) resulted in a dose response using the anthocyanin (ANTH) ratio which is the log expression of FER_RG. In a further experiment on Crimson Seedless, the clusters were separated into three colour groups after veraison; ABA and Ethrel were applied, leaving untreated clusters in each colour group as control. The ANTH ratio was able to quantify a significant increase in colour for each group with respect to its control. Conclusions: The results suggest that portable fluorescence detectors may become important tools to study ripening of table grapes. Significance of the Study: This is the first report of the use of fluorescence to follow ripening of table grapes.

Abbreviations ABA abscisic acid; CV coefficient of variance; DOY day of year; FERARI Fluorescence Excitation Ratio Anthocyanin Relative Index; FL fluorescence; FLR fluorescence ratio; MP3 Multiplex III; SD standard deviation; TA titratable acidity; TSS total soluble solids. Keywords: abscisic acid, multiplex, non-destructive measurement, proximal portable fluorescence, table grape

Introduction Ripening of grapes has been studied for many years, and harvest of fruit at optimal ripening is critical for winemaking and important for marketing and storage of table grapes (Coombe 1960). Traditionally, these studies relied on simple means like the measurement of total soluble solids (TSS) with a refractometer and acidity by titration, which are the gross sum of biochemical ripening processes, and by chromameter to document fruit colour. In green grapes like Thompson Seedless, the transition from green to yellow due to breakdown of chlorophyll and accumulation of carotenoids (Curl 1964, Cerovic et al. 2008) is an important factor which has a strong influence on acceptance by consumers. In coloured varieties, like Crimson Seedless, the colour is often lacking and requires the use of growth regulators or horticultural practices to improve colour development. In both the case of green and coloured grapes, means to assess the change in colour and the efficacy of treatment to improve colour are lacking. For green grapes, the transition to yellow can be measured by a chromameter with an aperture of 8 mm, which is suitable for indidoi: 10.1111/j.1755-0238.2012.00199.x © 2012 Australian Society of Viticulture and Oenology Inc.

vidual berries, but whole cluster analysis is time consuming. For coloured grapes, wineries use wet chemistry, but the table grape industry uses it mainly for research, and the major assessment criterion is subjective visual evaluation with its inherent limitations. Therefore, objective means to assess colour in table grapes are lacking, making it difficult to convey information between growers and traders. Hence, the development of novel, non-destructive methods to assess colour in table grapes should be welcomed. Chlorophyll and many other chemical groups have specific fluorescence (FL) properties which allow quantification of their accumulation. Portable instruments which are based on this technology are commercially available and have also has been tested for analysis of wine grape maturation (Cerovic et al. 2008, Ghozlen et al. 2010, Bramley et al. 2011). The instrument that was used in this study, Multiplex III (MP3) consists of four excitation light emitting diodes (LED) and three emission sensors, which create 12 signals. The ratios between these signals in different mathematical expressions (Table 1) were correlated to FL of major chemical groups, e.g. chlorophyll (SFR_R,

354

Auto-fluorescence for analysis of ripening

Australian Journal of Grape and Wine Research 18, 353–359, 2012

Table 1. Parameters used in this study according to design and nomenclature of the manufacturer (Force A, Orssay, France).

SFR_G SFR_R BRR_FRF FER_RUV FLAV FER_RG ANTH NBI_G NBI_R FERARI

Description

Formula

Correlation to†

Simple fluorescence ratio Simple fluorescence Blue-to-red fluorescence Fluorescence excitation Log of FER_RUV Fluorescence excitation ratio Log of FER_RG Nitrogen balance index Nitrogen balance index Log of FRF_R

FRF_G/RF_G FRF_R/RF_R BGF_UV/FRF_UV FRF_R/FRF_UV log(FER_RUV) FRF_R/FRF_G log(FER_RG) FRF_UV/RF_G FRF_UV/RF_R Log(5000/FRF_R)

Chlorophyll Chlorophyll Complex – grape maturation Flavonols Flavonols Anthocyanins Anthocyanins Epidermal phenolics and chlorophyll Epidermal phenolics and chlorophyll Anthocyanins

†According to Ghozlen et al. (2010) and references therein. G, R and UV are excitation at the green, red and UV range, respectively. The expression before the underscore (_) marks the emission range and the expression after the underscore, the excitation range. RF, red FL; FRF, far red FL; BGF, blue green FL.

which is the far-red emission divided by the red emission, both due to red excitation) and flavonoids (FLAV, the decadic logarithm of the red to ultraviolet (UV) excitation ratio of far-red chlorophyll fluorescence). Because of the use of ratios rather than absolute signals, the results are less sensitive to variation in the shape of the cluster (Ghozlen et al. 2010). The current study had two objectives: (i) to follow the ripening of Thompson Seedless grapes; and (ii) to determine if FL can be used to measure differences in the colour of Crimson Seedless grapes.

Materials and methods Experimental setup – Thompson Seedless The experiments where performed on a Thompson Seedless (Vitis vinifera L.) vineyard in Moshav Lachish, Israel (N31°33′33; E34°51′26). The planting system was a Y-shaped trellis with planting distances of 3 ¥ 1 m on Richter 110 (Vitis berlandieri ¥ V. rupestris) rootstock. The shoots were trained to 15 buds with six to seven shoots per vine. The vines were girdled on 20 May 2011. Gibberellin was applied two times, initially at full bloom at a concentration of 15 mg/L, then again 2 days later, followed by treatment with 30, 40 and 40 mg/L, respectively, at weekly intervals. The vines were irrigated every 2 days with water of 1.57 dS/m electrical conductivity by drippers spaced at 50 cm; 40 m3/ha was applied during the season. Nitrogen was supplied in the water as 200 kg/ha of 12% ammonium sulfate from bloom to fruit set, and 1700 L/ha of 2-2-10 (NPK) solution was supplied in the same manner from fruit set to harvest. Sampling of Thompson Seedless was performed on a weekly basis from the 30 June until the 25 August 2011 in four blocks of five to six vines per block which were randomly assigned in adjacent rows. The samples for ripening indices included ten cluster sections which were selected from both sides of five vines in each block. The clusters selected for sampling were representative of the vine, and the cluster sections included at least ten berries from the lateral branches. The cluster sections were stored for 31 days at 0°C at relative humidity of 92% after the first harvest on 3 August 2011, and the fruit was then held for 3 days at 20°C at relative humidity of 85% as simulated shelf life. In the second harvest on 25 August 2011, cluster sections were harvested as in the first harvest, ten berries with the stem cap attached were removed from each cluster section, and the 100 berries were stored in open punnets for 28 days at 0°C

followed by 3 days at 20°C at a relative humidity of 92 and of 85%, respectively.

Crimson Seedless Crimson Seedless (Vitis vinifera L.) vines grafted on Richter 110 rootstock were treated with 200 or 400 mg/L S-abscisic acid (S-ABA, Protone, Valent BioScience, Walnut Creek, CA, USA) to enhance colour development (Peppi et al. 2008) with 0.02% Triton X100 (Sigma, Rehovot, Israel) as a surfactant on 8 July 2011, which was approximately a week after the beginning of veraison. The experiment was carried out in a randomised block design with each block composed of three vines and in four replications. Control blocks were sprayed only with the surfactant. Cluster fluorescence as described below was measured on 21 July (3 weeks after veraison) on ten clusters per block. Another experiment on Crimson Seedless grapes was made by labelling clusters according to three colour categories according to the proportion of berries which were at the stage of colour change from green to red: G – ‘green’ (10% or less); GR – ‘green-red’ (about 20%) and R – ‘red’ (about 50%). The clusters were sprayed on 15 August 2011 (about 2 weeks after veraison) with 300 mg/L S-ABA and 480 mg/L of Ethrel (Bayer CropScience, Monheim, Germany). Cluster FL was measured after 17 days on 20 to 25 clusters for each colour group.

Measurement of fruit maturity Ten berries with stem cap attached were removed from each of the ten cluster sections for each replicate, the berries mixed and then 50 berries were taken for measurement of weight, TSS and titratable acidity (TA), with the latter calculated as tartaric acid equivalents and titrated to pH 8.2. Juice was prepared by maceration of the berries through a juice blender (Centrifugeuse, Lyon, France) and filtration through four layers of cheese cloth. TSS and TA were determined using a digital refractometer (Atago, Tokyo, Japan) and an automatic titrator (Metrohm, Herisau, Switzerland), respectively. From the remaining 50 berries, 20 berries were randomly sampled for determination of berry firmness. Firmness was measured by placing the 20 berries with intact cap stems on the turntable of a small fruit firmness analyser (Firmtech II, BioWorks, Wamego, Kansas, USA). The instrument measured force deformation compression with the load cell set at 350 g, equipped with a flat probe of 15 mm. Results are given in units of g/mm. © 2012 Australian Society of Viticulture and Oenology Inc.

Auto-fluorescence for analysis of ripening

Berryweight (g)

(a)

Weight

500

8.0

450

7.5

400

7.0 350 6.5 300

6.0

250

5.0

180

200

190

(b)

Statistical analysis was performed on a JMP 7.0.1 platform (SAS Institute Inc.) by one-way analysis of variance (ANOVA) with the least squares (LS)-means Tukey’s Honesty Significant Difference (HSD) protocol at a P-value ⱕ 0.05 or student – t-test as indicated.

Results Analysis of ripening of Thompson Seedless grapes The sampling of Thompson Seedless started during the rapid increase in fruit weight as the berries gained 2 g in 2 weeks (Figure 1). In the next 4 weeks the berries gained another 10 to 15% of weight, while during the last 2 weeks, berry weight started to decline. One week prior to the beginning of sampling, mean berry weight and firmness were 3.5 and 345 g/mm, respectively. On 30 June (DOY = 180) firmness was only 281 g/ mm, suggesting that veraison, marking the beginning of softening, occurred just prior to the beginning of sampling. From then on, fruit firmness increased steadily for 7 weeks after which it started to decline. Total soluble solids accumulated rapidly until the fifth time point (day of year, DOY = 208) after which the rate of accumulation was slightly lower (Figure 1). Given the fact that in the last 2 weeks of sampling, fruit weight started to decline, it is assumed that sugars were concentrated at this stage due to water loss. Acidity declined during this period (the first time point was not sampled for acidity) and both acidity and TSS could be described by simple second-order exponential equations with R2 values of 0.99. It should be pointed out that measurement at the sixth time point was not included due to a sampling issue. The commercial harvest in this vineyard spanned between DOY 215 to 230.

Changes in cluster fluorescence during ripening Cluster FL was measured in the vineyard over nine time points, and the FL signals were computed to 10 FL ratios (Table 1). SFR_G and SFR_R are ratios which represent chlorophyll FL excited by green or red light, respectively. These ratios created an overlapping line (Figure 2a) suggesting that they are interchangeable, and the steady decrease in their value matched the transition in grape berry colour from green to yellow during this period. The R2 value of the SFR ratios was 0.91, and the % © 2012 Australian Society of Viticulture and Oenology Inc.

200

TSS

210

220

230

240

Acidity

24

14 y = –0.0022x2 + 1.0766x–110.97 R² = 0.9924

22

12 10

20

8

18

6

16

4 y = 0.0045x2 – 2.0851x + 244.75 R²=0.994

14

Statistical analysis

Firmness

8.5

5.5

TSS (° Brix)

Analysis of cluster FL was carried out using a portable fluorometer (Multiplex III, Force A, France) (Ghozlen et al. 2010). The Thompson Seedless vineyard was sampled on the southern facing side of the row, while Crimson Seedless was sampled on the western facing side. Measurement was performed on ten clusters of uniform shape in each block. The clusters touched the 8 cm diameter aperture of the instrument, and the instrument was held horizontally and firmly to minimise movement during measurement. The measurements were all carried out by the same person and always at around 8 am for Thompson Seedless, while Crimson Seedless was sampled around 10 am. In the laboratory, cluster sections were measured by spreading the clusters on a table with a white background, and placing the MP3 above the clusters in a vertical position. For the second harvest of Thompson Seedless, on 25 August, measurements were carried out by randomly sampling ten detached berries from the 100 berries, which were contained in open punnets. The ten berries were placed on a black metal plate supplied with the instrument, and analysis was repeated ten times, each with a different group of berries. This process was repeated for each of the four replications representing the blocks in the vineyard.

2

12

180

Acidity (gL–1)

Measurement of fluorescence

355

Firmness (g mm–1)

Bahar et al.

0

190

200 210 220 Day of the year

230

240

Figure 1. Ripening indices of Thompson Seedless grapes during ripening. The sampling was performed between DOY 181 to 237 by removing five berries from ten clusters in four replications in a random block design. The results are an average of four replications and standard deviation values are presented as bars. coefficient variance (CV) was close to 10%, making them reliable markers for transition from green to yellow grapes (Table 2). FER_RUV is the ratio of far-red FL excited by red or UV, and the log value of this expression is the FLAV ratio and both correlate to accumulation of flavonols in the tissue (Figure 2a,d). These values increased steadily until the sixth time point and thereafter remained at the same level for the next 3 weeks in parallel to the time of the commercial harvest. Note the higher variation in the values of FER_RUV with respect to other ratios (Figure 2a). FER_RG is optimised for anthocyanin and is the ratio of far-red emission excited by red or green, and ANTH is the log of FER_RG (Figure 2b,c). These ratios had high R2 values (0.98) and low %CV values of 10 and 5, respectively (Table 2). Nitrogen balance index (NBI_G and NBI_R) ratios are both far-red FL excited by UV divided by red emission excited by green or red light, respectively (Figure 2b). These ratios are thought to integrate multiple FL signals derived from both chlorophyll and phenolics (Ghozlen et al. 2010). They had a similar trend between them and a similar trend to the SFR ratios. The R2 values of the NBI ratios are also similar to the SFR ratios, but the %CV is more than double. The Fluorescence Excitation Ratio Anthocyanin Relative Index (FERARI) value is the only parameter which is not a ratio, and consequently it can be seen that SD values are large, the trend line is not linear, and the R2 value is relatively low (Figure 2d and Table 2). It does show a similar trend to the ANTH ratio, but the %CV is 35%, which makes it inferior to other parameters for analysis of clusters that are complex objects. The BRR_FRF ratio, which is the blue-green to far-red FL excited by UV light, had the highest R2 value (0.98), but the %CV value (38%) was high, making it a less reliable marker under the experimental setup used in this

356

Auto-fluorescence for analysis of ripening

(a)

SFR_G

7

SFR_R

Australian Journal of Grape and Wine Research 18, 353–359, 2012

study (Figure 2d and Table 2). For three ratios, BRR_FRF, FER_RG and ANTH, a linear trend line had similar or higher R2 value compared to a second order polynomial trend line (Table 2). For the other ratios (excluding FERARI), the average R2 value of the second order trend line was 0.96 compared to 0.91 for a linear trend line. The second order correlation between SFR_R and TSS or FER_RG and TSS reached R2 values of 0.97 and 0.98, respectively (Figure 3).

FER_RUV

6

FLR

5 4 3 2 1 180

190

200 FER_RG

(b) 2.0

210

220

NBI_G

230

240

Storage of grapes may have a substantial effect on their quality, and the question in the current study was if it had any specific effect on fruit fluorescence. The data in Figure 4 show that in an early harvest (3 August), there was a significant reduction in the SFR_G but not of NBI_G ratios of the clusters after 31 days at cold storage and 3 days at 20°C. A second harvest and storage was performed 3 weeks later (25 August) with detached berries. The initial SFR_G was lower than the final value of the first harvest, and there was no further reduction in this value after storage.

NBI_R

1.6

FLR

1.2 0.8 0.4 0.0 180

Measurement of the effect of S-ABA on fruit fluorescence 190

200

210

220

FLAV

230

240

230

240

ANTH

FLR

(c) 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 180

190

(d) 0.5

200

210

220

FERARI

BRR_FRF

FLR

0.4 0.3 0.2 0.1 0.0 180

Changes in cluster and berry fluorescence during storage

190

200

210

220

230

240

Day of the year

Figure 2. Changes in fluorescence ratio (FLR) of Thompson Seedless grapes during ripening. The sampling was performed between DOY 181 to 237 on ten clusters in four replications. (a) SFR-G, SFR_R and FER_RUV. Note that curve lines of SFR_G and SFR_R overlap; (b) FER_RG, NBI_G and NBI_R; (c) Flavonoids (FLAV) and anthocyanin (ANTH); and (d) FERARI and BRR_FRF. Statistical data for the curves are presented in Table 2. FLR is the fluorescence ratio (except FERARI) which is not a ratio. SD values of 40 measurements for each time point are presented as error bars.

Crimson Seedless grapes can suffer from poor colour development in hot climates and this can be alleviated by the use of S-ABA (Peppi et al. 2008). In the experiment shown in Figure 5, the clusters were treated with 200 or 400 mg/L of S-ABA, and measurements of fruit FL were taken 3 weeks later. The data clearly show a dose-dependent response for the effect of S-ABA on the ANTH ratio but no significant effect on the FLAV or SFR ratios. The difference between the effect of the 200 and 400 mg/L treatments was maintained in a second measurement performed 1 month later (not shown). Both S-ABA and Ethrel are known to improve colour of Crimson Seedless (Peppi et al. 2008). The ability to differentiate between the treated and untreated clusters by cluster FL was tested in another experiment. Clusters were divided into three colour groups of ‘green’ (G – beginning of colour change); ‘green-red’ (GR – 20% colour change) and ‘red’ (R – 50% colour change). The clusters were treated with Ethrel and S-ABA at 480 and 300 mg/L, respectively, and measurements of fruit FL were taken after 3 weeks. Analysis of the SFR_R ratio, which correlates to chlorophyll level, showed significant difference for the control clusters between the G or GR categories and the R category (Figure 6). In the treated clusters, there was a significant difference in the SFR_R ratio between the G and R categories. Comparison between the untreated control and treated clusters by colour category showed a significant difference in the SFR_R ratio only for the GR category, however, combined analysis of all groups showed a significant difference between untreated and treated clusters. An opposite trend was observed for the ANTH ratio: for the untreated clusters, the R category differed from the GR and G groups. For the treated clusters, there was a significant difference between each category. A comparison between the untreated and treated clusters showed there was a significant difference for each category. These results show that the ANTH ratio can separate colour groups better then the SFR_R (chlorophyll) ratio.

Discussion The use of portable proximal FL for wine grapes was boosted by the release of commercial equipment. The MP3 technology was shown to be able to measure anthocyanin accumulation in congruence with wet chemistry (Ghozlen et al. 2010). Chlo© 2012 Australian Society of Viticulture and Oenology Inc.

Bahar et al.

Auto-fluorescence for analysis of ripening

357

Table 2. Statistical evaluation of fluorescence ratios obtained from measurement of Thompson Seedless grape clusters.

R2 value† Average ratio‡ Average SD§ % CV¶

SFR_G

SFR_R

BRR_FRF

FER_RUV

FLAV

FER_RG

ANTH

NBI_G

NBI_R

FERARI††

0.979 2.045 0.202 9.9

0.966 2.017 0.196 9.7

0.977 0.038 0.015 38.3

0.951 3.848 0.937 24.4

0.943 0.567 0.103 18.2

0.978 1.697 0.088 5.2

0.986 0.227 0.022 9.8

0.963 0.962 0.215 22.3

0.972 0.580 0.149 25.7

0.718 0.233 0.083 35.6

†R2 values were calculated from the second order polynomial trend lines of the kinetic curves (Figure 2) by Excel. ‡The average ratio for each parameter was generated from the 9 time points of the kinetic curves (Figure 2). Each time point is an average of 40 independent measurements. §Average SD is the average of the 9 standard deviation values which correspond to the 9 time points of the kinetic curves (Figure 2). Each SD value corresponds to 40 independent measurements. ¶%CV is the coefficient of variation – the average SD divided by the average ratio and multiplied by 100. ††FERARI is not a ratio.

FLR (SFR_R; NBI_G)

(a) TSS (°Brix)

22 y = –4.63x2 + 9.07x + 19.01 R2 = 0.97

20 18 16 14

1.8

A

SFR_G 1.6

B

1.4 1.2

NBI_G

A a

A

a a

1.0

a

12 1.6

1.7

1.8

1.9

2.0

2.1

2.2

2.3

2.4

2.5

FLR (SFR_R)

0.8 Before

After

Before

After

TSS (° Brix)

(b) First harvest-clusters

22

Second harvest-berries

2

y = –28.70x + 116.06x – 95.47 R2 = 0.98

20 18 16 14 12 1.5

1.6

1.7

1.8

1.9

2.0

FLR (FER_RG)

Figure 4. Changes in cluster or berry fluorescence (FL) during storage. The first harvest was carried out on DOY 215 and the second harvest on DOY 237. Storage was carried out at 0°C for 31 or 28 and 3 days at 20°C for the first and second harvest, respectively. Clusters or berries were measured in the laboratory on ten clusters (first harvest) or ten berries in ten replications (second harvest) both in four replications. FLR is the FL ratio. Statistical analysis was carried out by student t-test for each harvest and parameter separately. Different letters indicate significant differences between the treatments.

Figure 3. Correlation between average total soluble solids (TSS) and flourescence ratios (FLR). (a) Correlation of TSS to SFR_R. (b) Correlation of TSS to FER_RG. The correlation coefficients and the second order polynomial trend lines are indicated. Each data point is an average of 40 FL measurements and four TSS measurements from a pool of ten clusters.

rophyll degradation achieved less attention in wine grapes, but SFR was shown to correlate well with TSS (Ghozlen et al. 2010), and in other crops, with nitrogen balance and chlorophyll (Tremblay et al. 2012). In the current study on Thompson Seedless grapes, TSS continued to accumulate during the whole sampling time, but in the last 2 weeks, this rise could also be attributed to the reduction in berry weight which was accompanied by reduction in firmness. Such decrease in weight is unlikely because of water backflow in Thompson Seedless and possibly as a result of transpiration from the berry (Tilbrook and Tyerman 2009). The correlation coefficient between TSS and the major parameter in the MP3 system that represents chlorophyll degradation, the average SFR, has a value of 0.97 compared to 0.85 as a combined value for three wine grapes (Ghozlen et al. 2010). Interestingly, that study showed that the reduction in SFR_R during ripening was similar between the green and the coloured wine grapes suggesting that it can be used as a general marker to follow ripening. The FER_RG which © 2012 Australian Society of Viticulture and Oenology Inc.

Figure 5. Measurement of cluster FL of Crimson Seedless grapes after treatment with S-ABA. The measurements were carried out in the vineyard 13 days after treatment on ten clusters in four replications. FLR is the FL ratio. Statistical analysis was carried out by Tukey–Kramer test for each parameter separately. Different letters indicate significant differences between the treatments.

358

Auto-fluorescence for analysis of ripening

Figure 6. Measurement of cluster fluorescence of Crimson Seedless grapes after treatment with S-ABA and Ethrel. Clusters were assigned to three colour categories of G – ‘green’ (beginning of colour change); GR – ‘green-red’ (20% colour change) and R – ‘red’ (50% colour change). The clusters were treated by Ethrel and S-ABA at 480 and 300 mg/L, respectively on 15th of August, and measurements of fruit FL were taken after 3 weeks in the vineyard. Statistical analysis was carried out for each parameter separately using the Tukey–Kramer test. Different letters indicate significant differences between the treatments.

is considered an anthocyanin marker also increased steadily during ripening similarly to the ANTH_RG (the log value of FER_RG) for Chardonnay (Ghozlen et al. 2010). However, in a previous study with the white Chardonnay wine grapes, the change of the ANTH ratio was minor during ripening (Cerovic et al. 2008). One interpretation is that the latter finding is because of the accumulation of colourless proanthocyanins during ripening, but the other is that the berries become translucent during ripening (Cerovic et al. 2008) and, together with the degradation of chlorophyll, the phenolic FL becomes more pronounced (Ghozlen et al. 2010). Flavonols, which are influenced to a large extent by exposure to sunlight (Lenk et al. 2007), are represented by FER_RUV or FLAV, which is its log value (Table 1). In the present study, a steady increase occurred until the sixth time point (DOY = 215) followed by a plateau during the next three measurements. These results are inconsistent with the data reported by Cerovic et al. (2008) and hence require further validation. Interestingly, the change in slope in the FER_RUV was the time in which commercial selective harvest started, and hence it would be interesting to determine if this change in slope could be a marker for harvest maturity. The best timing for harvest of Thompson Seedless is dictated mainly by TSS and colour but also from commercial demands, which may require early selective harvest or storage on the vines. For cold storage and overseas transport, it is necessary to harvest at specific stages that are currently defined by threshold TSS and subjective measures. Early or late harvest will compromise the grapes because of higher sensitivity to postharvest treatments or pathogens, respectively (Crisosto et al. 2002). Grapes are non-climacteric, and it was of interest in this study to determine if their FL changed significantly during storage. According to the first harvest, the answer was positive (Figure 4) with respect to the SFR_G, but when the assay was performed after cold storage of detached berries in the late harvest, there was no difference in FL. This result could be due to the lower initial FL of the berries in the second harvest. It should be noted, however, that in the first harvest, the rachis suffered from severe browning after storage, and until further experiments can be carried out, it is not known what contribution the rachis made to this difference.

Australian Journal of Grape and Wine Research 18, 353–359, 2012

Colour can be a significant issue for Crimson Seedless in hot climates, and much effort has been invested in improving its colour as well as the colour of other varieties (Peppi et al. 2008). Treatment with ethylene, and more recently, in combination with abscisic acid, have led to good results, but these studies rely mainly on laborious wet chemistry or measurement of colour at the berry level with a chromameter with relatively low correlation coefficients between the parameters (Peppi et al. 2007). In many vineyard trials, subjective evaluation is also practised. Colour in many table grape varieties is generally less intense than in wine grapes, and it is not uncommon to find clusters in the vineyard with variable colour as well as internal variability in the colour. This requires integration of both colour intensity and uniformity in both evaluation and data analysis. Portable proximal FL as configured in the MP3 system seems like a promising solution to the problem: the window of 8 cm in diameter and the quick processing at less than a second means that a significant proportion of the cluster can be sampled and that hundreds of clusters can be sampled in 1 h at a rate similar to subjective evaluation. According to studies in wine grapes, the FER_RG ratio or its log ANTH ratio are obvious choices at relatively low anthocyanin concentration, but at the high anthocyanin concentration, it is the FERARI value which is less skewed by interference from chlorophyll (Ghozlen et al. 2010). There are complex models to account for this conclusion, but the fact is that the ANTH ratio increases exponentially until about 0.2 mg/cm2 of anthocyanins, followed by plateau or decrease, while the FERARI index has a reasonable linear behaviour. However, since the anthocyanin level in major table grape varieties such as Red Globe is much lower than that of major coloured wine grapes such as Merlot (Du et al. 2012), FER_RG or ANTH may be considered as reliable comparative markers. Unlike coloured wine grapes in which the absolute level of anthocyanins is considered important to the quality of the wine, in table grapes, colour is a matter of consumer preference, and, therefore, it is important to have tools to determine the relative efficacy of the treatments to improve colour. The results obtained in this study are not intended to compare the efficacy of specific treatments, and also they do not represent an extensive validation of the use of the MP3 device for colour assessment of table grapes. They do demonstrate, however, that this system can differentiate between treatments in a dose-dependent manner, and also it can separate the efficacy of the treatments based on the colour maturity of the clusters at the time of application. The rapid sampling time and data processing of the MP3 system were recently utilised in realtime measurements on a conveyer of a wine grape harvester in Australia (Bramley et al. 2011) demonstrating the robustness of the system. In summary, portable proximal fluorescence can be an efficient tool for analysis of the quality of table grapes in the vineyard and after harvest.

Acknowledgements We wish to thank Yuval Sadan, Shahaf Dudai and Guy Rozenfeld from Moshav Lachis and Hovav Weksler for their assistance in the experiment. We also acknowledge the advice of Laurent Florin and Marc Pastor from Force A, Orsay France on operation of the Multiplex III instrument. This research was partly funded by the Plant Council of Israel – the table grape board. A scholarship for the work of Dr Askin Bahar was supplied by the Turkish Higher Education Council. This is contribution 633/12 from the ARO, The Volcani Center. © 2012 Australian Society of Viticulture and Oenology Inc.

Bahar et al.

References Bramley, R.G.V., Le Moigne, M., Evain, S., Ouzman, J., Florin, L., Fadaili, E.M., Hinze, C.J. and Cerovic, Z.G. (2011) On-the-go sensing of grape berry anthocyanins during commercial harvest: development and prospects. Australian Journal of Grape and Wine Research 17, 316–326. Cerovic, Z., Moise, N., Agati, G., Latouche, G., Benghozlen, N. and Meyer, S. (2008) New portable optical sensors for the assessment of winegrape phenolic maturity based on berry fluorescence. Journal of Food Composition and Analysis 21, 650–654. Coombe, B. (1960) Relationship of growth and development to changes in sugars, auxins, and gibberellins in fruit of seeded and seedless varieties of Vitis vinifera. Plant Physiology 35, 241–250. Crisosto, C.H., Garner, D. and Crisosto, G. (2002) High carbon dioxide atmospheres affect stored ‘Thompson Seedless’ table grapes. Hortscience 37, 1074–1078. Curl, A.L. (1964) The carotenoids of several low carotenoid fruits. Journal of Food Science 29, 241–245. Du, B., He, B.J., Shi, P.B., Li, F.Y., Li, J. and Zhu, F.M. (2012) Phenolic content and antioxidant activity of wine grapes and table grapes. Journal of Medicinal Plant Research 6, 338–3387. Ghozlen, N.B., Cerovic, Z.G., Germain, C., Toutain, S. and Latouche, G. (2010) Non-destructive optical monitoring of grape maturation by proximal sensing. Sensors 10, 10040–10068.

© 2012 Australian Society of Viticulture and Oenology Inc.

Auto-fluorescence for analysis of ripening

359

Lenk, S., Buschmann, C. and Pfündel, E.E. (2007) In vivo assessing flavonols in white grape berries (Vitis vinifera L. cv. Pinot Blanc) of different degrees of ripeness using chlorophyll fluorescence imaging. Functional Plant Biology 34, 1092–1104. Peppi, M.C., Fidelibus, M.W. and Dokoozlian, N.K. (2007) Application timing and concentration of abscisic acid affect the quality of ‘Redglobe’ grapes. Journal of Horticultural Science and Biotechnology 82, 304–310. Peppi, M.C., Fidelibus, M.W. and Dokoozlian, N.K. (2008) Timing and concentration of abscisic acid applications affect the quality of ‘Crimson Seedless’ grapes. International Journal of Fruit Science 7, 71–83. Tilbrook, J. and Tyerman, S.D. (2009) Hydraulic connection of grape berries to the vine: varietal differences in water conductance into and out of berries, and potential for backflow. Functional Plant Biology 36, 541– 550. Tremblay, N., Wang, Z. and Cerovic, Z. (2012) Sensing crop nitrogen status with fluorescence indicators. A review. Agronomy for Sustainable Developmentent 32, 451–464.

Manuscript received: 13 December 2011 Revised manuscript received: 5 June 2012 Accepted: 19 June 2012