Prediction of Skin Coloration of Grape Berries from Air Temperature

Abstract

To develop a method for predicting the skin color of grape berries of three cultivars of Vitis labrusca L. × Vitis vinifera L. grown in Japan, we investigated the relationship between skin color and air temperature in the grape production areas of 18 prefectures. When mean air temperature during the 40 days before harvest date was ≥24°C, the skin color ratings of ‘Kyoho’, ‘Pione’, and ‘Suzuka’ were significantly negatively correlated with temperature. Skin color ratings decreased by about 1 unit per 1°C increase; at a given mean air temperature during this period, the rating of ‘Suzuka’ was higher (by 0.7 units) than that of ‘Kyoho’, which was higher (by 1.0 unit) than that of ‘Pione’. Because an approach to predict harvest date has not been established, we developed a method to predict skin color at harvest based on air temperature after the full-flowering date. We found the times that had a strong negative correlation between the mean air temperature and the skin color rating at harvest was 43 days from 50 DAF (days after full flowering) for ‘Kyoho’, 46 days from 46 DAF for ‘Pione’, and 42 days from 52 DAF for ‘Suzuka’. We obtained a linear regression equation for the relationship between the skin color rating at harvest and the mean air temperature during the periods. If the full-flowering date is known, it is possible to predict skin color at harvest by using this equation and the predicted air temperature after full flowering. We also developed a method for predicting anthocyanin contents in berry skins at harvest using significant regressions among the skin color rating, the skin anthocyanin content, and mean air temperature.

Introduction

Climate change is already profoundly influencing fruit trees. Many phenological time-course trends have been studied in fruit trees, and most are consistent with warming temperatures (Menzel, 2003; Nemani et al., 2001). Long-term changes in fruit quality have also been reported (Sugiura et al., 2013). Grapevine phenology and grape berry quality also appear to be subject to climate change (Orduna, 2010); in particular, an increased frequency of grape berries with poor color has been observed by many grape farmers in recent years and this has become a major problem in Japan (Osada and Ohe, 2010; Sugiura et al., 2012).

High air temperatures lead to poor color grape berries (Kliewer, 1970; Kobayashi et al., 1967) by interfering with anthocyanin biosynthesis in the skins (Mori et al., 2004). ‘Kyoho’ and ‘Pione’ are black-skinned table grapes. Poor skin color of ‘Kyoho’ and ‘Pione’ is known as “red ripening”, and the grapes have low commercial value (Kobayashi et al., 1965). Therefore, measures have been developed to improve skin coloration, such as girdling (Fujishima et al., 2005; Koshita et al., 2011; Yamane and Shibayama, 2006a; Yamane et al., 2007), cluster thinning (Yamane and Shibayama, 2006a; Yamane et al., 2007), and abscisic acid treatment (Kataoka et al., 1982; Kitamura et al., 2007). However, these measures have adverse effects on grape berry production: girdling inhibits rooting (Yamane and Shibayama, 2006a) and decreases vine vitality (Kugimiya et al., 2011), cluster thinning reduces yield, and abscisic acid treatment increases production costs and reduces sugar content in berries (Habu et al., 2009; Kugimiya et al., 2011).

These measures need to be taken more than 1 month before harvest. If the skin color of grape berries at harvest could be predicted, it would be possible to minimize the adverse effects of such measures by taking them only in the years when skin coloration at harvest is projected to be poor. Determining the relationship between skin color and air temperature would make it possible to predict skin color from estimated air temperatures. In addition, the relationship could be used to assess the impact of future climate change on grape quality.

Previous studies qualitatively investigated the relationship between skin coloration of grape berries and temperature in temperature manipulation experiments (Kobayashi et al., 1965, 1967; Shinomiya et al., 2015; Tomana et al., 1979) or analyzed skin color only in certain areas (Naito et al., 1986; Yamane and Shibayama, 2006b).

‘Kyoho’ is the leading grape cultivar in Japan (Yamada and Sato, 2016), ‘Pione’ is the third most common cultivar, and ‘Suzuka’ is a new cultivar. All these cultivars are similar in appearance and have black berries at maturity. The purpose of the present study was to develop an equation for the temperature–color relationship for ‘Kyoho’, ‘Pione’, and ‘Suzuka’ grapevines cultivars based on measurements of skin color in many grape-growing regions across Japan, with the ultimate goal of helping farmers to determine whether measures to improve skin color need to be taken.

Materials and Methods

Plant materials and vineyard measurements

We investigated the full-flowering date, harvest date, and berry skin color rating at harvest of ‘Kyoho’, ‘Pione’, and ‘Suzuka’ grapevines (Vitis labrusca L. × Vitis vinifera L.) in 18 experimental vineyards managed by Japanese prefectural agricultural research institutes (Fig. 1; Table 1) from 2011 to 2015.

Fig. 1

Map of Japan, showing the locations of the 18 experimental vineyards at prefectural agricultural research institutes where phenological events and skin color ratings were investigated. The site numbers correspond to the numbers in Table 1.

Table 1

Prefectures, cultivation conditions of vineyards, and year of investigation.

The plants were 3 (2011) to 7 (2015) years old and were trained to systems shown in Table 1. Most of the grapevines were grown under cover (colorless plastic films or glass), but some were grown in open fields. A solution of gibberellic acid (GA₃, 25 ppm) was applied to all clusters once or twice to induce seedless berries. Other cultivation conditions, including soil management and the level of cluster thinning, were as recommended in each prefectural guideline.

The full-flowering date was defined as the date when more than 80% of all florets were open in more than 80% of all flower clusters. The cluster harvest time was mainly judged by perceived flavor (Yamada, 2011). The harvest date was defined as the date when 50% of the clusters had been harvested. Ten harvested clusters were sampled randomly on the harvest date, and ten berries from each cluster were randomly selected to rate the skin color. The berry skin color rating (0 = green to 12 = black) corresponds to values in the standardized color chart (Yamazaki and Suzuki, 1980) that is used to visually assess the skin color of purple- and black-skinned grape berries.

Anthocyanin content in the berry skin

A few clusters harvested in each vineyard in 2011 were used to measure the anthocyanin content of the berry skins by the method of Shiraishi et al. (2007) with a slight modification. After rating the skin color as above, about 0.1 g of skin near the equator of each of 10 berries from the same cluster was removed, combined, blotted dry on paper towels (Crecia, Japan), macerated in 10 mL of 50% v/v aqueous acetic acid for 12 h at 4°C in the dark, and the suspension was filtered (No. 2 filter paper; Advantec, Japan). The absorbance of the filtrate (1 mL) at 520 nm was measured on a UV-260 spectrophotometer (Shimadzu, Japan) with the settings of SCALE = 20 nm·cm⁻¹ and SLIT = 2 nm. The anthocyanin content was expressed as mg of cyanidin-3-monoglucoside equivalent per gram fresh weight (g FW) of berry skin.

Climate data

For the air temperatures at each vineyard, we used the daily mean air temperatures in the AMeDAS mesh dataset (Seino, 1993), a climate dataset with 1-km resolution (each grid cell measures 45" in longitude × 30" in latitude) estimated using statistics from meteorological observation stations of the Japan Meteorological Agency (JMA) across Japan.

Results and Discussion

Comparison of skin color levels of the three cultivars

The mean phenological and color characteristics of all cultivars in the vineyards are summarized in Table 2. Although it is difficult to compare the cultivars directly owing to differences among prefectures and among years, the mean full-flowering date in late May differed by only 1 day. The harvest date and period from full flowering to harvest differed by about 1 week among the cultivars. The harvest date of ‘Suzuka’ was 4 days earlier than that of ‘Kyoho’, which was 4 days earlier than that of ‘Pione’. ‘Suzuka’ had the darkest mean skin color at harvest, followed by ‘Kyoho’ and ‘Pione’.

Table 2

Summary of the mean dates of phenological events and mean skin color rating for all vineyards and years combined.

The relationship between the skin color rating and the mean air temperature during the 40 days before harvest date for ‘Kyoho’ is shown in Figure 2A. The rating was not significantly correlated with temperature at <24°C, but decreased significantly with increasing temperature at ≥24°C. The slope of the linear regression line was −1.003; that is, the skin color rating of ‘Kyoho’ decreased by about 1 for every 1°C rise in air temperature.

Fig. 2

The relationship between the skin color rating at harvest and the mean air temperature during the 40 days before harvest date. (A) ‘Kyoho’. Linear regression at air temperatures ≥24°C is shown. The data for cultivation in the open field and under cover were pooled for linear regression. (B) ‘Pione’ and ‘Suzuka’. Linear regressions at air temperatures ≥24°C for ‘Pione’ and ‘Suzuka’ are based on the assumption that their regression slopes equaled that for ‘Kyoho’ (= −1.003). The regression line for ‘Kyoho’ from (A) is shown for comparison. r, Pearson’s correlation coefficient (***P < 0.001; **P < 0.01).

High air temperatures inhibit color development of grape berries (Kliewer, 1970; Kobayashi et al., 1967), and high temperature during the 40 days before harvest date leads to poor color of ‘Kyoho’ berries (Naito et al., 1986; Tomana et al., 1979). Experiments with an artificial temperature treatment indicated that grape berries also cannot develop their color at a too-low temperature (Coombe, 1987; Kobayashi et al., 1965; Poudel et al., 2009). The skin color rating had no significant correlation with temperature at <24°C (the open circles in Fig. 2A). This may be caused by an increase in the number of days when skin color development advances slowly or not at all owing to low temperature.

Skin coloration data for the open field (the black squares in Fig. 2A) tended to lay above the regression line, suggesting that skin becomes darker under open-field conditions than under a colorless cover, but the data are not sufficient for a definite conclusion. Therefore, we combined also skin color ratings from both conditions in the following analyses.

The skin color ratings of ‘Pione’ and ‘Suzuka’ also significantly decreased with increasing mean air temperature during the 40 days before harvest date at ≥24°C, but not at <24°C (Fig. 2B). The slopes of the linear regressions for ‘Pione’ (−0.891) and ‘Suzuka’ (−0.802) did not differ significantly (P was 0.647 for ‘Pione’ and 0.390 for ‘Suzuka’) from the slope of ‘Kyoho’ (−1.003).

To compare skin coloration among the cultivars, we performed linear regression of the skin color ratings for ‘Pione’ and ‘Suzuka’ (Fig. 2B) based on the assumption that the slopes of their regressions equaled the slope of the regression for ‘Kyoho’. The intercepts of the regressions differed significantly among the cultivars (P was 0.000 for ‘Pione’ and 0.001 for ‘Suzuka’). This difference indicates that at the same mean air temperature during the 40 days before harvest date, the skin color rating of ‘Suzuka’ would be higher by 0.7 (= 35.79 − 35.06) units than that of ‘Kyoho’, which would be higher by 1.0 (= 35.06 − 34.03) units than that of ‘Pione’. The results suggest that ‘Suzuka’ has the darkest skin, followed by ‘Kyoho’ and then ‘Pione’.

‘Suzuka’ (Grape Fukuoka No. 15) was bred to improve skin coloration at high air temperatures (Shiraishi et al., 2013). Our results suggest that this goal has been achieved. The regression analysis (Fig. 2) suggests that the skin color rating of ‘Suzuka’ will equal that of ‘Kyoho’ if the ‘Suzuka’ vineyard is 0.7°C warmer than the ‘Kyoho’ vineyard.

Development of an equation to predict skin color

If harvest date and air temperature can be predicted, the skin color at harvest can be predicted using the regression equation in Figure 2. Because an approach to predict harvest date has not been established, we developed a method to predict skin color at harvest from the full-flowering date. Only skin color ratings at air temperatures ≥24°C were used in the following analyses.

We first estimated the timing having a strong correlation between the mean air temperature and the skin color rating at harvest (the temperature-sensitive period) for each cultivar. To identify this period, we calculated the correlation coefficients between skin color rating at harvest and mean air temperature for N (a variable number) days before the harvest date (Fig. 3). The strongest correlation was found when N was 43 days for ‘Kyoho’, 46 days for ‘Pione’, and 42 days for ‘Suzuka’. The slope and intercept of each regression equation on these dates are shown in Table 3.

Fig. 3

Changes in the coefficient of correlation (Pearson’s r) between the skin color rating at harvest and the mean air temperature for N days before the harvest date (see inset). FD, full-flowering date; HD, harvest date.

Table 3

Relationships between the skin color rating at harvest and the mean air temperature for N days before harvest. Values represent the strongest (most negative) correlation coefficient for each cultivar (based on the data in Fig. 3).

Next, we calculated the correlation coefficients between skin color rating at harvest and mean air temperature during 43, 46, and 42 days beginning from a variable number (M) of DAF (days after full flowering) for ‘Kyoho’, ‘Pione’, and ‘Suzuka’, respectively (Fig. 4). The correlation was strongest when M was 50 for ‘Kyoho’ and 46 for ‘Pione’.

Fig. 4

Changes in the coefficient of correlation (Pearson’s r) between the skin color rating at harvest and the mean air temperature for N (= 43 for ‘Kyoho’, 46 for ‘Pione’, or 42 for ‘Suzuka’) days starting from M days after the full-flowering date (see inset). FD, full-flowering date; HD, harvest date.

The correlation coefficient for ‘Suzuka’ had two peaks (52 and 65 DAF). When M was 65, the end of the strong correlation was 106 DAF, which was considerably later than the mean harvest date (87.6 ± 11.3 days; Table 2). Therefore, we adopted 52 as M to predict the skin color of ‘Suzuka’. The slope and intercept of each regression equation on these dates are shown in Table 4. These regression equations can be used to predict the skin color rating at harvest.

Table 4

Relationships between the skin color rating at harvest and the mean air temperature for N (=43 for ‘Kyoho’, 46 for ‘Pione’, or 42 for ‘Suzuka’) days from M days after the full-flowering date, where M was the value when the correlation coefficient for each cultivar was most negative (based on the data in Fig. 4).

These results suggest that the temperature-sensitive period for predicting the skin coloration with temperature was 43 (=N) days from 50 DAF (=M) for ‘Kyoho’. The end of the durations was similar to the mean harvest date (92.5 DAF; Table 2).

Previous studies (Barnuud et al., 2014; Yamane and Shibayama, 2006b) suggested that the temperature-sensitive period for skin coloration of grape berries begins from soon after veraison. In ‘Kyoho’ and ‘Pione’, veraison is 45 to 50 DAF (Wang et al., 1997; Yakushiji et al., 2001a, b). The M values for both cultivars (M = 50 and 46 DAF, respectively) were similar to the veraison.

If the full-flowering date is known, the skin color rating at harvest can be estimated from the predicted air temperature after full-flowering and the regression equations in Table 4. Recently, the JMA started releasing forecasts of air temperature for up to 28 days in each prefecture through its official website (http://www.data.jma.go.jp/gmd/risk/probability/guidance/index_k1.php, December 17, 2016). After 28 days, normal values (average values for the past 30 years) of daily mean air temperature are available. A method for predicting daily mean air temperatures across Japan at a 1-km resolution based on the forecasts distributed by the JMA and climatic normal values has also been developed (Ohno et al., 2016). However, because all mean air temperatures during the temperature-sensitive periods used for regression calculations were ≥24°C, the equations may not provide reliable results at lower temperatures during the temperature-sensitive period.

The root-mean-square errors (RMSE) of the skin color rating between the estimated and measured values ranged from 0.73 to 1.09 color units (Table 4), indicating that skin coloration can be predicted with an uncertainty of about 1 unit if prediction of the air temperature is accurate. One of the chief sources of error is solar radiation, high levels of which deepen the skin color of grape berries (Kliewer, 1970; Naito et al., 1984; Shinomiya et al., 2015; Smart et al., 1988; Tarara et al., 2008). In the present study, the amount of solar radiation varied among the vineyards, years, and growing conditions (under cover or open field). The next step in our research will be to improve the ability to predict skin color by accounting for solar radiation.

Variability in cultivation conditions is another source of error. In particular, the level of cluster thinning affects skin color (Yamane et al., 2007). ‘Suzuka’ had the lowest correlation coefficient (Figs. 3 and 4) and the largest RMSE, possibly because of variability in cultivation conditions such as the thinning level and the criteria used to determine harvest time, as it is a new cultivar. Future work is needed to accurately predict the skin color of ‘Suzuka’.

Prediction of the anthocyanin content in berry skins

The skin color rating and the anthocyanin content in the berry skins at harvest were significantly correlated in all three cultivars (Fig. 5). Because the relationship between skin color and anthocyanin content varies among cultivars (Shiraishi et al., 2007; Watanabe and Shiraishi, 1991), we calculated the linear regression separately for each cultivar. The anthocyanin contents in the berry skins at harvest could be predicted by using a combination of the regressions shown in Figure 5 and the regressions to predict the skin color rating (Table 4). For instance, an increase in 1°C would decrease the skin color rating at harvest by 0.96 (= 0.96 × 1) units in ‘Kyoho’ and the decrease in the skin color rating is comparable to the decrease in anthocyanin content by 0.67 (= 0.96 × 0.696) mg·g⁻¹FW. The same increase in temperature would decrease the anthocyanin content by 0.66 mg·g⁻¹FW in ‘Pione’ and 0.48 mg·g⁻¹FW in ‘Suzuka’.

Fig. 5

The relationships between the skin color rating and the anthocyanin content in the grape berry skins. Linear regressions are shown. r, Pearson’s correlation coefficient (***P < 0.001; *P < 0.05).

Conclusions

We developed a method to predict the skin color rating of grape berries and the anthocyanin content of their skins from the mean air temperature during a certain period. This method will help grape growers and researchers to determine whether to take measures to improve skin coloration and to assess the effect of future climate changes on grape quality.

Acknowledgements

We thank the researchers of the Ibaraki Agricultural Center, Horticultural Research Institute; Tochigi Prefectural Agricultural Experiment Station; Ishikawa Agriculture and Forestry Research Center, Sand Dune Agricultural Experiment Station; Yamanashi Fruit Tree Experiment Station; Nagano Fruit Tree Experiment Station; Mie Prefecture Agricultural Research Institute, Iga Branch; Kyoto Prefectural Agriculture, Forestry and Fisheries Technology Center, Tango Agriculture Research Division; Research Institute of Environment, Agriculture and Fisheries, Osaka Prefecture; Nara Fruit Tree Research Center; Tottori Prefectural Agriculture and Forest Research Institute, Sand Dune Agricultural Research Center; Shimane Agricultural Technology Center; Okayama Prefectural Technology Center for Agriculture, Forestry and Fisheries, Research Institute for Agriculture; Tokushima Prefectural Agriculture, Forestry and Fisheries Technology Support Center, Fruit tree Experiment Station; Kagawa Prefectural Agricultural Experiment Station, Fuchu Branch; Ehime Research Institute of Agriculture, Forestry and Fisheries, Fruit Tree Research Center; and Kagoshima Prefectural Institute for Agricultural Development, Hokusatsu Branch for providing the high-quality data from each vineyard.

Literature Cited

•  Barnuud,  N. N.,  A.  Zerihun,  M.  Gibberd and  B.  Bates. 2014. Berry composition and climate: Responses and empirical model. Int. J. Biometeorol. 58: 1207–1223.
•  Coombe,  B. G. 1987. Influence of temperature on composition and quality of grapes. Acta Hortic. 206: 23–35.
•  Fujishima,  H.,  M.  Shiraishi,  S.  Shimomura and  Y.  Horie. 2005. Effect of girdling on berry quality of ‘Pione’ grapevine. Hort. Res. (Japan) 4: 313–318 (In Japanese with English abstract).
•  Habu,  T.,  F.  Kishida,  T.  Konishi,  K.  Kusumi,  M.  Matsuda and  A.  Kitajima. 2009. Improvement in coloring technique of grape berries by applying fertilizer containing natural abscisic acid. II. Bull. Exp. Farm Kyoto Univ. 18: 41–43 (In Japanese).
•  Kataoka,  I.,  A.  Sugiura,  N.  Utsunomiya and  T.  Tomana. 1982. Effect of absicisic acid and defoliation on anthocyanin accumulation in Kyoho grapes (Vitis vinifera L. × V. labruscana Bailey). Vitis 21: 325–332.
•  Kitamura,  H.,  M.  Nakayama,  H.  Kondou,  Y.  Nishikawa,  M.  Koshioka and  S.  Hiratsuka. 2007. Effect of abscisic acid treatment at variable stages on the expression of proper and deep color in ‘Aki Queen’ grape berry skins. Hort. Res. (Japan) 6: 271–275 (In Japanese with English abstract).
•  Kliewer,  W. M. 1970. Effect of day temperature and light intensity on coloration of Vitis vinifera L. grapes. J. Amer. Soc. Hort. Sci. 95: 693–697.
•  Kobayashi,  A.,  T.  Fukushima,  N.  Nii and  K.  Harada. 1967. Studies on the thermal conditions of grapes. VI. Effects of day and night temperature on yield and quality of Delaware grapes. J. Japan. Soc. Hort. Sci. 36: 373–379.
•  Kobayashi,  A.,  H.  Yukinaga and  T.  Itano. 1965. Studies on the thermal conditions of grapes. III. Effects of night temperature at the ripening stage on the fruit maturity and quality of Delaware grapes. J. Japan. Soc. Hort. Sci. 34: 26–32 (In Japanese with English abstract).
•  Koshita,  Y.,  T.  Yamane,  H.  Yakushiji,  A.  Azuma and  N.  Mitani. 2011. Regulation of skin color in ‘Aki Queen’ grape: Interactive effects of temperature, girdling, and leaf shading treatments on coloration and total soluble solids. Sci. Hortic. 129: 98–101.
•  Kugimiya,  N.,  Y.  Imai,  S.  Kawada,  S.  Kiyohara and  M.  Ueyama. 2011. Methods for improving the coloration of the grape variety ‘Pione’. Bull. Oita Pref. Agric. For. Fish. Res. Cent. 1: 89–101 (In Japanese with English abstract).
•  Menzel,  A. 2003. Plant phenological anomalies in Germany and their relation to air temperature and NAO. Climatic Change 57: 243–263.
•  Mori,  K.,  S.  Sugaya and  H.  Gemma. 2004. Regulatory mechanism of anthocyanin biosynthesis in ‘Kyoho’ grape berries grown under different temperature conditions. Environ. Control Biol. 42: 21–30.
•  Naito,  R.,  H.  Yamamura,  K.  Ikegami,  S.  Ohi and  K.  Mishima. 1984. Effects of light intensity around cluster on the coloration and pigmentation in ‘Kyoho’ grape. Bull. Fac. Agric. Shimane Univ. 18: 8–15 (In Japanese with English abstract).
•  Naito,  R.,  H.  Yamamura and  K.  Murata. 1986. Effects of temperature and solar radiation after veraison on coloration of ‘Kyoho’ grapes. Bull. Fac. Agric. Shimane Univ. 20: 1–7 (In Japanese with English abstract).
•  Nemani,  R. R.,  M. A.  White,  D. R.  Cayan,  G. V.  Jones,  S. W.  Running,  J. C.  Coughlan and  D. L.  Peterson. 2001. Asymmetric warming over coastal California and its impact on the premium wine industry. Climate Res. 19: 25–34.
•  Ohno,  H.,  K.  Sasaki,  G.  Ohara and  K.  Nakazono. 2016. Development of grid square air temperature and precipitation data compiled from observed, forecasted, and climatic normal data. Climate in Biosphere 16: 71–79 (In Japanese with English abstract).
•  Orduna,  R. M. 2010. Climate change associated effects on grape and wine quality and production. Food Res. Int. 43: 1844–1855.
•  Osada,  K. and  Y.  Ohe. 2010. Evaluating impacts of climate change on farm management—The case of grape farming in Yamanashi. J. Rural Problem 179: 266–269 (In Japanese with English abstract).
•  Poudel,  P. R.,  R.  Mochioka,  K.  Beppu and  I.  Kataoka. 2009. Influence of temperature on berry composition of interspecific hybrid wine grape ‘Kadainou R-1’ (Vitis ficifolia var. ganebu × V. vinifera ‘Muscat of Alexandria’). J. Japan. Soc. Hort. Sci. 78: 169–174.
•  Seino,  H. 1993. An estimation of distribution of meteorological elements using GIS and AMeDAS data. J. Agric. Meteorol. 48: 379–383 (In Japanese).
•  Shinomiya,  R.,  H.  Fujishima,  K.  Muramoto and  M.  Shiraishi. 2015. Impact of temperature and sunlight on the skin coloration of the ‘Kyoho’ table grape. Sci. Hortic. 193: 77–83.
•  Shiraishi,  M.,  R.  Shinomiya and  M.  Awamura. 2013. Agronomic characteristics of well-colored grape, ‘Fukuoka No. 15’ for global warming. Hort. Res. (Japan) 12 (Suppl. 2): 301 (In Japanese).
•  Shiraishi,  M.,  M.  Yamada,  N.  Mitani and  T.  Ueno. 2007. A rapid determination method for anthocyanin profiling in grape genetic resources. J. Japan. Soc. Hort. Sci. 76: 28–35.
•  Smart,  R. E.,  S. M.  Smith and  R. V.  Winchester. 1988. Light quality and quantity effects on fruit ripening for Cabernet Sauvignon. Am. J. Enol. Vitic. 39: 250–258.
•  Sugiura,  T.,  H.  Ogawa,  N.  Fukuda and  T.  Moriguchi. 2013. Changes in the taste and textural attributes of apples in response to climate change. Sci. Rep. 3: 2418.
•  Sugiura,  T.,  H.  Sumida,  S.  Yokoyama and  H.  Ono. 2012. Overview of recent effects of global warming on agricultural production in Japan. JARQ 46: 7–13.
•  Tarara,  J. M.,  J.  Lee,  S. E.  Spayd and  C. F.  Scagel. 2008. Berry temperature and solar radiation alter acylation, proportion, and concentration of anthocyanin in Merlot grapes. Am. J. Enol. Vitic. 59: 235–247.
•  Tomana,  T.,  N.  Utsunomiya and  I.  Kataoka. 1979. The effect of environmental temperature on fruit ripening on the tree. II. The effect of temperatures around whole vines and clusters on the coloration of ‘Kyoho’ grapes. J. Japan. Soc. Hort. Sci. 48: 261–266 (In Japanese with English abstract).
•  Wang,  S.,  G.  Okamoto and  K.  Hirano. 1997. Vine growth and fruit development of ‘Pione’ grapes planted in root-restricted buried and raised beds. J. Japan. Soc. Hort. Sci. 66: 253–259 (In Japanese with English abstract).
•  Watanabe,  Y. and  S.  Shiraishi. 1991. Relationship between skin color and anthocyanin of grapes. Sci. Bull. Fac. Agric. Kyushu Univ. 46: 33–37 (In Japanese with English abstract).
•  Yakushiji,  H.,  K.  Morinaga and  S.  Kobayashi. 2001a. Promotion of berry ripening by 2, 3, 5-triiodobenzoic acid in ‘Kyoho’ grapes. J. Japan. Soc. Hort. Sci. 70: 185–190.
•  Yakushiji,  H.,  N.  Sakurai and  K.  Morinaga. 2001b. Changes in cell-wall polysaccharides from the mesocarp of grape berries during veraison. Physiol. Plant. 111: 188–195.
• Yamada, M. 2011. Cross breeding in woody fruit crops (In Japanese). Yokendo, Tokyo.
•  Yamada,  M. and  A.  Sato. 2016. Advances in table grape breeding in Japan. Breed. Sci. 66: 34–45.
•  Yamane,  T. and  K.  Shibayama. 2006a. Effect of trunk girdling and crop load levels on fruit quality and root elongation in ‘Aki Queen’ grapevines. J. Japan. Soc. Hort. Sci. 75: 439–444.
•  Yamane,  T. and  K.  Shibayama. 2006b. Effects of changes in the sensitivity to temperature on skin coloration in ‘Aki Queen’ grape berries. J. Japan. Soc. Hort. Sci. 75: 458–462.
•  Yamane,  T.,  J.  Kato and  K.  Shibayama. 2007. Coloration of ‘Aki Queen’ grapes in different production areas and improvement of coloration by girdling and cluster thinning. Hort. Res. (Japan) 6: 441–447 (In Japanese with English abstract).
•  Yamazaki,  T. and  K.  Suzuki. 1980. Color charts: Useful guide to evaluate the fruit maturation. I. Colorimetric specification of color charts for Japanese pear, apple, peach, kaki and citrus fruits. 1980. Bull. Fruit Tree Res. Stn. A7: 19–44 (In Japanese with English abstract).