How does carbohydrate supply limit flower development in grape and kiwifruit vines?

Transcription

1 How does carbohydrate supply limit flower development in grape and kiwifruit vines? Annette Claire Richardson July 2014 Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy School of Agricultural and Wine Sciences Charles Sturt University Wagga Wagga Australia

2

3 i

4 Table of Contents Table of Contents List of Figures List of Tables Certificate of Authorship Acknowledgements Abstract Abbreviations ii vii xix xxvi xxviii xxx xxxii 1 Literature review Introduction The flowering process Floral Initiation Floral organ morphogenesis How does flowering behaviour of perennials affect crop yield? Are yields of perennial fruiting crops similar? Is low flower production typical of fruiting vines? Are the flowering behaviours of grape and kiwifruit vines similar? When do carbohydrates limit flower development of grape and kiwifruit? Thesis aims and hypotheses 28 2 General materials and methods Plant material Commercial vines Potted vines Meteorological data Vine phenology Budbreak and flowering Phenological stages of development Leaf area estimation Gas exchange 42 ii

5 2.6 Carbohydrate analysis Tissue extraction Starch determination Soluble carbohydrate determination Data handling and statistical analysis Data handling Data analysis Data plots and curve fitting 45 3 Seasonal patterns of kiwifruit and grapevine development Introduction Materials and Methods Plant Material Vine phenology Gas exchange measurements Net carbon balance Sampling and carbohydrate analysis Statistical analysis Results Vine phenology Gas exchange Total carbohydrate concentrations Discussion Seasonal patterns of vine growth Carbohydrate composition Gas exchange Carbon acquisition and balance Whole vine carbon budgets Summary Carbohydrate supply during inflorescence initiation and bud development Introduction 105 iii

6 4.2 Materials and Methods Plant Material Experimental design Results Shoot growth Axillary bud and internode dry weight in winter Winter bud and internode carbohydrate content Effects on spring development Relationship between carbohydrate content, dry weight and bud performance Discussion Effects of shoot treatments on inflorescence initiation Seasonal effects on shoot growth and floral development Effects of treatments on shoot development and inflorescence production Summary Modifying carbohydrate supply during autumn: effect on the early stages of floral morphogenesis in grape and kiwifruit vines Introduction Materials and Methods Plant material Experimental design Measurements Results Total carbohydrate concentrations and content Amount of budbreak and flowering Time of budbreak and flowering Shoot growth Discussion Effect of treatments on total carbohydrate accumulation Effects of autumn shade and girdling on subsequent flower production 185 iv

7 5.4.3 Effect of girdling on other metabolites Summary Carbohydrate supply during floral sex organ development Introduction Materials and Methods Plant material Experimental design and treatments Measurements Statistical analysis Results Shoot measurements Inflorescence and flower measurements Fruit set, seed number and seed weight Yield Shoot dry weight and leaf area Shoot total carbohydrate concentrations and content Gas exchange Net carbon balance Discussion Floral morphogenesis Shoot development prior to anthesis Shoot development after anthesis Pollination and fruit set Changes in carbon supply before anthesis influences final yield Summary Conclusions and Integration Hypothesis 1: Hypothesis 2: Hypothesis 3: Hypothesis 4: Future studies 269 v

8 7.6 Overall Summary 270 References 273 Appendix 1 Kiwifruit and grape pest and disease programme 293 Appendix 2 Meteorological Data 295 Appendix 3 Leaf area determination 300 Appendix 4 Measurements of Solarshade TM characteristics 302 vi

9 List of Figures Figure 1.1: Key stages of floral development in grape buds. Reproduced from Dunn and Martin (2007) Figure 1.2. Key stages in the floral development of kiwifruit buds as described by Brundell (1975c) and Walton et al. (1997) Figure 2.1: Hayward vines on Bruno rootstocks were 24-years-old, canepruned and grown on a pergola at the Kerikeri Research Centre Figure 2.2: Shiraz vines on m470 rootstocks were five-years-old, cane-pruned and grown on a Smart-Dyson split canopy trellis at Marsden Estate Figure 2.3: Cabernet Sauvignon vines on SO4 rootstocks were 20-yearsold, cane pruned and trained on a VSP system at Ivana Wines Vineyard Figure 2.4: Hayward and Cabernet Sauvignon vines were grown in pots at the Kerikeri Research Centre for use in multi-season studies Figure 3.1: Timing of key growth stages for Hayward (A, B and C), Shiraz (D, E and F) and Cabernet Sauvignon (G) vines. Data are for the 2008/2009 season (A and D), 2009/2010 season (B and E) and the 2010/2011 season (C, F and G) Figure 3.2: Dynamics of budbreak in A, 2009/2010 and B, 2010/2011 for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ; 2010/2011 only) vines. Data are the means ± SEM, n = 16. The lines are the fit of data to the Boltzmann sigmoid function Figure 3.3: Dynamics of flowering in A, 2009/2010 and B, 2010/2011 seasons for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( 2010/2011 only) vines. Data points are the means ± SEM, n = 16. The lines are the fit of data to the Boltzmann sigmoid function Figure 3.4: Periods of root loss (die back of root tips observed at sampling) and root growth (white root tips present at sampling). Data are for Hayward vines (A, 2008/2009, B, 2009/2010 and C, 2010/2011) Shiraz (D, 2008/2009, E, 2009/2010 and F, 2010/2011) and Cabernet Sauvignon (G, 2010/2011) Figure 3.5: The development of Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots across the 2010/2011 season. A, total leaf area per vii

10 shoot, B, shoot length and C, shoot basal diameter. Data are means ± SEM, n = Figure 3.6: Changes in the total dry weight per shoot during the 2010/2011 season. Data for A, Hayward ( ), leaves ( ), stems ( ), and fruit ( ), B, Shiraz ( ), leaves ( ), stems ( ), and fruit ( ) and C, Cabernet Sauvignon ( ), leaves ( ), stems ( ), and fruit ( ) shoots. Note the different scales for each vine cultivar. Data are means ± SEM, n = Figure 3.7: Photosynthetic light responses of youngest mature leaves from A, Hayward ( ), B, Shiraz ( ) and C, Cabernet Sauvignon ( ) shoots at anthesis (December 2010). Data are mean values ± SEM, n = Figure 3.8: Rates of A, photosynthesis, B, stomatal conductance, and C, transpiration measured concurrently under ambient conditions (temperature o C, PFD µmol (photons) m -2 s -1, CO µmol CO 2 mol -1, VPD kpa) over the growing season. Data for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots are mean values for every second leaf on a shoot ± SEM, n = Figure 3.9: Changes in gas exchange with leaf position along a shoot from the base to the tip at around 80 DAB. Rates of A, photosynthesis, B, stomatal conductance and C, transpiration. Measurements were made concurrently under ambient conditions (temperature o C, PFD µmol (photons) m -2 s -1, CO µmol CO 2 mol -1, VPD kpa) at approximately 80 DAB, during the very early stages of fruit development. Data are for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots. Data are means ± SEM, n = Figure 3.10: Changes in rates of carbon acquisition (A - C), carbon accumulated as biomass (D - F) and the net carbon balance of shoots (G - I) of Hayward ( A, D, G), Shiraz ( B, E, H) and Cabernet Sauvignon ( C, F, I) shoots. Note the different scales between Hayward and other vines for each parameter. Data are means ± SEM, n = Figure 3.11: Changes in TNC in the leaves (A - C), developing axillary buds (D - F) and internodes G - I) of shoots during the 2008/09 (A, D, G), 2009/10 (B, E, H) and 2010/11 (C, F, I) growing seasons. Data are shown for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots. Note viii

11 different scales for leaves, axillary buds and internode TNC concentrations. Data points are the means ± SEM, n = Figure 3.12: Changes in TNC concentration in inflorescences and berries of three vines across the 2010/2011 season. Data are shown for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) berries. Data are means ± SEM, n = Figure 3.13: Changes in TNC concentration in wood during A, 2008/2009, B, 2009/10 and C, 2010/11 seasons and fine roots over D, 2008/2009, E, 2009/10 and F, 2010/11 seasons. Data are shown for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) vines. Data are means ± SEM, n = Figure 3.14: Changes in starch ( ) and total sugar concentrations ( ) in the leaves (A - C), developing axillary buds (D - F) internodes G - I) and fruit (J - L) of shoots during the 2010/11 growing seasons. Data are shown for Hayward (, ), Shiraz (, ) and Cabernet Sauvignon (, ) shoots. Note different scales for fruit concentrations. Data points are the means ± SEM, n = Figure 3.15: Changes in starch and total sugar concentrations in the wood (A - C), and root (D - F) of vines during the 2010/11 growing seasons. Data are shown for Hayward (starch, total sugars ), Shiraz (starch, total sugars ) and Cabernet Sauvignon (starch,total sugars ). Data points are the means ± SEM, n = Figure 3.16: Changes in total shoot carbohydrate content of A, Hayward leaves ( ) and stems ( ), B, Shiraz leaves ( ) and stems ( ) and C, Cabernet Sauvignon leaves ( ) and stems ( ) across the 2010/2011 season. Data are mean values ± SEM, n = Figure 3.17: Changes in carbohydrate content of shoots (solid symbols) and fruit (open symbols) throughout the growing season for Hayward shoots ( ), Shiraz shoots ( ) and Cabernet Sauvignon shoots ( ). Data are means ± SEM, n = Figure 4.1 Pot grown vines were used in this experiment with Hayward vines in the foreground (row one) and Cabernet Sauvignon vines in rows three - five) ix

12 Figure 4.2: Girdling treatments applied to shoots. Examples A, a girdled Cabernet Sauvignon shoot exposed to full sun, B, a girdled and shaded Cabernet Sauvignon shoot and C, a girdled and defoliated Hayward shoot Figure 4.3: Growth of shoots at each stage of development for Cabernet Sauvignon A - C, 2009/2010 and D - F, 2010/2011; Hayward G - I, 2009/2010 and J - L, 2010/2011. Shoot treatments include untreated, girdled, girdled and defoliated and girdled and shaded. The lines are the fit of data to the Boltzmann sigmoid function. Data are means ± SEM, n = 11. Note the differences in scale between Cabernet Sauvignon and Hayward shoots. For Cabernet Sauvignon 50% anthesis occurred DAB and veraison DAB, and for Hayward 50% anthesis occurred DAB and seed maturation DAB Figure 4.4: A, Effects of vine cultivar (averaged over shoot girdling treatments) and stage of shoot development treatment on the dry weight of dormant buds compared with dry weight of untreated buds in winter Vines are Cabernet Sauvignon and Hayward. B, Effects of shoot girdling treatments (averaged over vine cultivar) and stage of development treatment on the dry weight of dormant buds compared with the dry weight of untreated buds in winter Shoot treatments include untreated, girdled, girdled and defoliated and girdled and shaded. Data are means ± SEM, n = Figure 4.5: A, Effects of shoot girdling treatments (averaged over vine cultivar and stage of development) on the dry weight of shoot internodes compared with untreated shoot internodes in winter B, effects of stage of shoot development treatment (averaged over vine cultivar and shoot girdling treatment) on the dry weight of internodes compared with the dry weight of untreated internodes in winter Data are means ± SEM, n = Figure 4.6: Effects of treatments applied at different stages of development in 2009/2010 on the carbohydrate content of A, Cabernet Sauvignon buds and C, adjacent internodes (untreated, girdled, girdled and defoliated and girdled and shaded ); B, Hayward buds and D, adjacent internodes (untreated, girdled, girdled and defoliated and girdled and shaded ) x

13 in winter Data are means ± SEM, n = 3. Note the differences in scale between bud and internode carbohydrate content Figure 4.7: Time of budbreak in spring for Cabernet Sauvignon (A 2010 and C 2011, untreated, girdled, girdled and defoliated, girdled and shaded and Hayward (B 2010 and D 2011 untreated, girdled, girdled and defoliated, girdled and shaded canes. Budbreak percentage was calculated from the cumulative number of buds that broke at each date and final budbreak of each cane. Data are means ± SEM, n = for each shoot girdling treatment (averaged across stage of development treatments) compared with untreated shoots. The lines are the fit of data to the Boltzmann sigmoid function Figure 4.8: The effects of the stage of development that shoot treatments were applied in 2009/2010 compared with untreated shoots, on floral productivity of buds of Cabernet Sauvignon (untreated treated ) and Hayward (untreated treated ) in spring A, the number of inflorescences produced per bud; B, floral budbreak (%) and C, inflorescences per floral shoot. Data are means ± SEM, n = Figure 4.9: The main effects of shoot girdling treatments (averaged over vine cultivar and stage of development treatment) (A, C and E) and the stage of development treatments (averaged over vine cultivar and shoot girdling treatment) (B, D and F) that were applied to shoots in 2010/2011 on floral productivity of buds in spring 2011/2012. A, and B, the number of inflorescences produced per bud; C, and D, floral budbreak; E, and F, inflorescences per floral shoot. Data are means ± SEM, n = Figure 4.10: The effects of shoot girdling treatments (averaged over stage of development treatments) in the previous season on the number of inflorescences produced in each 20% of nodes from the base (0%) to the apex (100%) of canes. Treatment means are shown for Cabernet Sauvignon canes in A, 2010 and C, 2011 (untreated, girdled, girdled and defoliated and girdled and shaded ); and for Hayward canes in B, 2010 and D, 2011 (untreated, girdled, girdled and defoliated and girdled and shaded ). Data are means ± SEM, n = Figure 4.11: The relationship between bud dry weight (A - D), internode dry weight (E - H), bud carbohydrate content (I - L) and internode carbohydrate xi

14 content (M - P), on inflorescence number per bud, total budbreak (%), floral budbreak (%) and inflorescence number per floral shoot. Data are for Cabernet Sauvignon ( ) and Hayward ( ) shoots in winter and spring of Figure 5.1 Autumn shade and girdling treatments applied to canes. A, shaded Hayward vine, B, shaded Cabernet Sauvignon vine, C, girdled Hayward cane and D, girdled Cabernet Sauvignon cane Figure 5.2: TNC concentrations in shaded and exposed canes (averaged across vine cultivar and girdling treatment) in A, winter 2009, C, winter 2010, and in girdled and intact canes (averaged across vine cultivar and shading treatment) in B, winter 2009 and D, winter Data are means ± SEM, n = Figure 5.3: TNC concentrations in roots in response to light treatments (averaged across vine cultivar) in A, winter 2009 and B, winter Data are means ± SEM, n = Figure 5.4: TNC content in roots and cordons of Cabernet Sauvignon and Hayward in response to light treatments. A, roots winter 2009, B, roots winter 2010, C, cordon winter 2009 and D, cordon winter Note the difference in scale between A, B, and C, D. Data are means ± SEM, n = Figure 5.5: TNC content of exposed and shaded canes (averaged across vine cultivar and girdling treatment) in A, winter 2009, C, winter 2010, and in girdled and intact canes (averaged across vine cultivar and shading treatment) in B, winter 2009 and D, winter Data are means ± SEM, n = Figure 5.6: The effect of shade treatments on the number of inflorescences produced per winter bud in A, 2009/2010 (averaged across girdling treatment), C, 2010/2011 (averaged across girdling treatment) and E, 2011/2012 (averaged across vine cultivar and girdling treatments). The effect of girdling treatments on the number of inflorescences produced per winter bud in B, 2009/2010 (averaged across vine cultivar and shade treatments), D, 2010/2011 (averaged across shade treatments) and F, 2011/2012 (averaged across vine cultivar and shade treatments). Data are xii

15 means ± SEM, n = In A, C and D data are for Cabernet Sauvignon and Hayward. Data are means ± SEM, n = Figure 5.7: The effect of shade treatments (averaged across vine cultivar and girdling treatments) on total budbreak A, 2009/2010 and C, 2011/2012 and floral budbreak B, 2009/2010 and D, 2011/2012. Data are means ± SEM, n = Figure 5.8: The effect of girdling treatments on total budbreak (A, 2009, C, 2010 and E, 2011) and floral budbreak (B, 2009, D, 2010 and F 2011). A, B, E and F averaged across vine cultivar and shade treatment. C and D averaged across shade treatment, and data are for Cabernet Sauvignon and Hayward. Data are means ± SEM, n = Figure 5.9: Effects of shade and girdling treatments on inflorescence numbers per floral shoot. A, spring 2009, B, spring 2010 and C, spring Data are for Cabernet Sauvignon and Hayward vines. Data are means ± SEM, n = Figure 5.10: Time of budbreak for Hayward (A, spring 2009 B, spring 2010 and C, spring 2011) and Cabernet Sauvignon (D, spring 2009, E, spring 2010 and F, spring 2011) canes. Budbreak (%) is the number of buds that broke at each date as a percentage of the final number of buds that broke. Treatments are Hayward exposed and shaded, and Cabernet Sauvignon exposed and shaded with intact canes represented by a solid line and girdled canes by a dashed line. Data points are means ± SEM, n = Figure 5.11: Time of flowering for Hayward (A, spring 2009 B, spring 2010 and C, spring 2011) and Cabernet Sauvignon (D, spring 2009, E, spring 2010 and F, spring 2011) canes. Flowering (%) is the number of flowers open at each date as a percentage of the total flower number at the end of anthesis. Treatments are Hayward exposed and shaded, and Cabernet Sauvignon exposed and shaded, with intact canes represented by a solid line and girdled canes by a dashed line. Data are means ± SEM, n = Figure 5.12: Effects of girdling treatments (averaged across shade treatments) on inflorescence weight in A, spring 2009 and B, spring xiii

16 Data are from Cabernet Sauvignon and Hayward canes. Data are means ± SEM, n = Figure 5.13: The effect of shade on canes in autumn of the previous growing season (2009/2010) on the growth of shoots between budbreak and anthesis in the spring of the 2010/2011 season. Data for light exposure treatments are averaged across vine cultivar and girdling treatment. Data are means ± SEM, n = Figure 6.1: Spring shoots and applied treatments. A, Shiraz shoot, B, Hayward shoot, C, shaded Shiraz shoot pair and D, girdled Hayward shoot Figure 6.2: Effects of vine cultivar and light (averaged across girdling treatments) on the dimensions of shoots at anthesis. A, length 2009/2010, B, length 2010/2011; C, number of nodes 2009/2010, D, number of nodes 2010/2011; E, basal diameter 2009/2010; F, basal diameter 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = Figure 6.3: Effects of vine cultivar and shoot girdling treatment (averaged across light treatments) on shoot dimensions at anthesis in 2010/2011. A, shoot length; B, number of nodes. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = Figure 6.4: Effects of vine cultivar and light treatment (averaged across girdling treatment) on flower numbers per shoot A, Shiraz 2009/2010; B, Hayward 2009/2010; C, Cabernet Sauvignon 2010/2011; D, Hayward 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ Figure 6.5: Effects of vine cultivar and light treatment (averaged across girdling treatment) on fruit set (%) and retention in A, 2009/2010; B, 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ Figure 6.6: Effects of vine cultivar and light treatment (averaged over girdling treatment) on seed number per berry of A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet xiv

17 Sauvignon 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between Shiraz/Cabernet Sauvignon and Hayward data. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ Figure 6.7: Effects of vine cultivar and light treatment (averaged over girdling treatment) on mean seed dry weight of A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between Shiraz/Cabernet Sauvignon and Hayward data. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ Figure 6.8: Effects of vine cultivar, light and girdling treatment on fruit yield per shoot in A, 2009/2010; B, 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 12, 2009/2010, n = 16, 2010/ Figure 6.9: Effects of vine cultivar and light treatment (averaged over girdling treatment) on berry number per shoot. A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Note the differences in scale between Shiraz/Cabernet Sauvignon and Hayward berry numbers. Data are means ± SEM, n = 12, 2009/2010, n = 16, 2010/ Figure 6.10: Effects of vine cultivar and girdling treatment (averaged across light treatment) on berry numbers per shoot in 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale for Shiraz and Cabernet Sauvignon compared to Hayward. Data are means ± SEM, n = 12, 2009/2010, n = 16, 2010/ Figure 6.11: Effects of vine cultivar, light and girdling treatment on mean berry weight. A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale for Shiraz and Cabernet Sauvignon compared to Hayward. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ xv

18 Figure 6.12: Effects of vine cultivar and light treatment (averaged across girdling treatment) on the dry weight of shoot components at anthesis. A, leaf 2009/2010; B, leaf 2010/2011; C, stem 2009/2010; D, stem 2010/2011; E, inflorescence 2009/2010; F, inflorescence 2010/2011; G, total shoot weight 2009/2010; H, total shoot weight 2010/2011. E, F and G, show light treatment means averaged across vine cultivar and girdling treatment. In A, B, C, D and H data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between the shoot components. Data are means ± SEM, A, C and F n = 16, E and G n = 32, B, D and H n = Figure 6.13: Effects of vine cultivar and light treatment (averaged over girdling treatment) on the dry weight of shoot components at harvest A, leaf 2009/2010; B, leaf 2010/2011; C, stem 2009/2010; D, stem 2010/2011; E, fruit 2009/2010; F, fruit 2010/2011; G, total shoot weight 2009/2010; H, total shoot weight 2010/2011. B and G, show light treatment means averaged across vine cultivar and girdling treatment. In A, C - F and H data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between the shoot components. Data are means ± SEM, F and H n = 4, D n = 8, A, C, and E n = 16, B n = 24, G n = Figure 6.14: Effects of light treatment (averaged over vine cultivar and girdling treatment A, B, and D) on the TNC concentration of shoot components in 2010/2011. A, leaf at anthesis; B, leaf at harvest; C, stem at anthesis; D, stem at harvest. In C, the bars represent data from Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, C n = 4, A, B and D n = Figure 6.15: Effects of vine cultivar and light treatment (averaged over girdling treatment A and D) and of light treatment (averaged over vine cultivar and girdling treatment B and C) on the TNC content of shoot components in 2010/2011. A, leaf at anthesis; B, leaf at harvest; C, stem at anthesis; D, stem at harvest. In A, and D the bars represent data from Shiraz, Hayward and Cabernet Sauvignon. Note the different scale in D. Data are means ± SEM, A and D n = 8, B and C n = xvi

19 Figure 6.16: Effects of girdling treatment (averaged across vine cultivar and light treatment) on A, the TNC concentration and B, the TNC content of stems at harvest in 2010/2011. Data are means ± SEM, n = Figure 6.17: Photosynthetic light responses for the youngest mature leaves on shaded ( ) and exposed ( ) leaves from A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots at anthesis. Data for are means ± SEM, n = Figure 6.18: Effects of shade and girdling treatments on mean photosynthetic rates of shaded ( ) and exposed ( ) leaves of A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots in the 2010/2011 season. Solid lines represent intact shoots and dashed lines girdled shoots. Treatments were removed from Shiraz vines 70 DAB, Cabernet Sauvignon vines 72 DAB and Hayward vines 62 DAB. Data were collected from every second leaf on each shoot and the mean rates calculated per shoot. Data are means ± SEM, n = Figure 6.19: Effects of shade and girdling treatments on stomatal conductance of shaded ( ) and exposed ( ) leaves of A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots in the 2010/2011 season. Solid lines represent intact shoots and dashed lines girdled shoots. Treatments were removed from Shiraz vines 70 DAB, Cabernet Sauvignon vines 72 DAB and Hayward vines 62 DAB. Data were collected from every second leaf on each shoot and the mean rate calculated per shoot. Data are means ± SEM, n = Figure 6.20: Effects of shade and girdling treatments on mean transpiration rates of shaded ( ) and exposed ( ) leaves of A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots in the 2010/2011 season. Solid lines represent intact shoots and dashed lines girdled shoots. Treatments were removed from Shiraz vines 70 DAB, Cabernet Sauvignon vines 72 DAB and Hayward vines 62 DAB. Data were collected from every second leaf on each shoot and the mean rate calculated per shoot. Data are means ± SEM, n = Figure 6.21: Effects of treatments on the net carbon acquisition A, at anthesis and B, fruit harvest; carbon in shoot biomass C, at anthesis and D, fruit harvest; and the shoot net carbon balance E, at anthesis and F, fruit xvii

20 harvest in 2010/2011. In each case, the bars represent data from Shiraz, and Cabernet Sauvignon and Hayward. Data are means ± SEM, n = xviii

21 List of Tables Table 1.1. The range in yield and fruit set (the percentage of flowers that produce mature fruit) for common perennial fruiting crops Table 1.2: Key stages of floral morphogenesis in grape buds using the Eichorn-Lorenz (E-L) and BASF, Bayer, Ciba-Geigy and Hoechst (BBCH) models. Reproduced from Coombe (1995) and May (2004) Table 1.3: Key stages of floral morphogenesis in kiwifruit buds (Brundell 1975c; Polito and Grant 1984; Hopping 1990) Table 2.1: Details of the location, site characteristics, vine material and cultural techniques used on the three commercial orchards/vineyards used Table 2.2: Details of the location, site characteristics, vine material and cultural techniques used on the Hayward and Cabernet Sauvignon vines grown in pots for multi-season studies Table 2.3: Key phenological stages described using the modified E-L (Eichorn-Lorenz) and BBCH (BASF, Bayer, Ciba-Geigy and Hoechst) growth stage models for Shiraz and Cabernet Sauvignon grapevines reproduced from Lorenz et al. (1994); Coombe (1995); May (2004). Key BBCH stages for Hayward kiwifruit (Salinero et al. 2009) and a newly developed E-L type system for Hayward Table 3.1: The date and DAB for key growth stages of Hayward and Shiraz vines (2008/ /2011) and Cabernet Sauvignon vines (2010/2011) Table 3.2: Budbreak and floral production of vines. Data are mean values ± SEM, n = 16 for Hayward and Shiraz vines (2009/2010 and 2010/2011 seasons) and for Cabernet Sauvignon vines (2010/2011 season) Table 3.3: Parameters describing photosynthetic light responses for Hayward, Shiraz and Cabernet Sauvignon leaves measured at anthesis, where P max is the light saturated maximum rate of photosynthesis, Rs is the rate of respiration, φ app is the apparent (CO 2 limited) photon yield and PFDsat is the photon flux density at which photosynthesis was light saturated according to Greer and Halligan (2001). Data for each vine cultivar are means ± SEM, n = xix

22 Table 3.4: Gas exchange parameters for Hayward, Shiraz and Cabernet Sauvignon leaves averaged over the growing season. Environmental conditions ranged between: temperature o C, PFD µmol (photons) m -2 s -1, CO µmol CO 2 mol -1, VPD kpa). Data are mean values ± SEM. Hayward n = 1120, Shiraz n = 922 and Cabernet Sauvignon n = Table 3.5: Coefficients and statistics from fitting a multiplicative model of stomatal conductance and internal CO 2 concentration to photosynthetic data. The percentage of error mean square (EMS) accounted for by each of the terms and the interaction between terms, are presented. The number of measurements used to develop the model for each cultivar (n) is also shown Table 3.6: Estimated whole vine carbon allocations determined from vine dry weight data from Field (2013) for Shiraz and Cabernet Sauvignon vines and data from Smith et al. (1992) for Hayward vines. Total non-structural carbon content per vine component was calculated using dry weight data and TNC concentrations from immediately prior to fruit harvest (23 March 2011 Shiraz and Cabernet Sauvignon and 27 April 2011 Hayward ). Total structural carbon is the difference between total carbon and total nonstructural carbon. Percentages each component contributed to the total vine are shown in brackets Table 4.1: Calendar date and days after budbreak (DAB) of each stage of development (E-L stage) that shoot treatments were applied to Cabernet Sauvignon and Hayward shoots, during the 2009/2010 and 2010/2011 growing seasons Table 4.2: Effects of treatments applied during different stages of inflorescence initiation in the 2009/2010 season on the mid-point (50%) and duration of growth of Cabernet Sauvignon and Hayward shoots calculated from fitting the Boltzmann sigmoid function. Data are means ± SEM, n = Table 4.3: The effects of shoot treatments applied during different stages of inflorescence initiation development in the 2010/2011 season on the midpoint and duration of growth of Cabernet Sauvignon and Hayward shoots xx

23 calculated from fitting the Boltzmann sigmoid function. Data are means ± SEM, n = Table 4.4: Effect of shoot girdling and stage of development treatments in 2009/2010 season on the dry weight of Cabernet Sauvignon and Hayward buds and internodes compared with that in untreated shoots during winter Data are means ± SEM, n = Table 4.5: Effect of shoot girdling and stage of development treatments in 2009/2010 on the carbohydrate content of Cabernet Sauvignon and Hayward buds and internodes during the winter of Data are means ± SEM, n = Table 4.6: Effects of shoot girdling and stage of development treatments in the 2009/2010 season on the performance of Cabernet Sauvignon and Hayward buds compared with that of untreated shoots during spring 2010/2011. Data are means ± SEM, n = Table 4.7: The effects of shoot girdling treatments and stage of development treatment in the 2010/2011 season on the performance of Cabernet Sauvignon and Hayward buds during spring 2011/2012 compared with that of untreated shoots. Data are means ± SEM, Cabernet Sauvignon n = 7, Hayward n = Table 4.8: The statistics of fitting the additive general linear model to bud performance data. The P value and r 2 for the model and the F value for each dependent variable are presented for each attribute tested Table 5.1: The effects of shoot treatments applied during autumn (February - June/July) 2009 on the TNC concentration of Cabernet Sauvignon and Hayward tissues during winter of the same season (August 2009). Data are means ± SEM, n = Table 5.2: The effects of shoot treatments applied during autumn (February - June/July) 2010 on the TNC concentration of Cabernet Sauvignon and Hayward tissues during winter of the same season (August 2010). Data are means ± SEM, n = Table 5.3: The effects of shoot treatments applied during autumn (February - June/July) 2009 on the TNC content of Cabernet Sauvignon and Hayward tissues during winter (August 2009). TNC concentrations from Tables 5.1 and 5.2 were combined with bud and cane fresh biomass, cane length and xxi

24 DM% data from carbohydrate analyses to calculate bud and cane TNC content. Biomass data from Smith et al. (1992) for Hayward and Field (2013) for Shiraz together with TNC concentrations were used to calculate cordon and root TNC content. Data are means ± SEM, n = Table 5.4: The effects of shoot treatments applied during autumn (February - June/July) 2010 on the TNC content of Cabernet Sauvignon and Hayward tissues during winter (August 2010). TNC concentrations from Tables 5.1 and 5.2 were combined with bud and cane fresh biomass, cane length and DM% data from carbohydrate analyses to calculate bud and cane TNC content. Biomass data from Smith et al. (1992) for Hayward and Field (2013) for Shiraz together with TNC concentrations were used to calculate cordon and root TNC content. Data are means ± SEM, n = Table 5.5: The effect of treatments applied to vines in February 2009 on the performance of Cabernet Sauvignon and Hayward buds during spring Data are means ± SEM, n = Table 5.6: Effects of treatments applied to vines in February 2010 on the performance of Cabernet Sauvignon and Hayward buds during spring Data are means ± SEM, n = Table 5.7: The effects of treatments applied to vines in February 2011 on the performance of Cabernet Sauvignon and Hayward buds during spring Data are means ± SEM, n = Table 5.8: The effects of treatments applied during reserve storage in autumn on the mid-point and duration of budbreak of Cabernet Sauvignon and Hayward canes during the following spring, calculated from fitting the Boltzmann sigmoid function to data. Data are presented for three seasons and are means ± SEM, n = Table 5.9: The effects of treatments applied during reserve storage in autumn on the mid-point and duration of flowering of Cabernet Sauvignon and Hayward canes during the following season, calculated from fitting the Boltzmann sigmoid function to data. Data are presented for three seasons and are means ± SEM, n = Table 5.10: The effects of treatments applied to vines in February on the size of Cabernet Sauvignon and Hayward inflorescences during the xxii

25 following spring in the 2009 and 2010 seasons. Data are means ± SEM, n = Table 5.11: The effects of treatments applied to vines in February (autumn) on the length of Cabernet Sauvignon and Hayward shoots at anthesis in the subsequent spring. Data are means ± SEM, n = Table 6.1: Dates and DAB that treatments were applied to Shiraz, Hayward and Cabernet Sauvignon shoots in 2009/2010 and 2010/ Table 6.2: Effects of altering carbohydrate supply to shoots prior to anthesis on the dimensions of Shiraz and Hayward Table 6.3: Effects of altering carbohydrate supply to shoots prior to anthesis on the dimensions of Shiraz, Cabernet Sauvignon Table 6.4: Effects of treatments to modify shoot carbohydrate prior to anthesis on inflorescence weight and flower number on Shiraz and Hayward shoots in 2009/2010. Data are means ± SEM, n = 8 inflorescence weight; n = 12 flower number per shoot Table 6.5: Effects of treatments to modify shoot carbohydrate prior to anthesis on inflorescence weight on Shiraz, Cabernet Sauvignon and Hayward shoots and flower number per Cabernet Sauvignon and Hayward shoot in 2010/2011. Data are means ± SEM, n = 4 inflorescence weight; n = 16 flower number per shoot Table 6.6: The effects of treatments to modify carbohydrate supply of shoots prior to anthesis on the percentage of flowers that produced harvestable berries, seed number per berry and mean seed weight in 2009/2010. Data are means ± SEM, n= Table 6.7: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the percentage of Shiraz, Cabernet Sauvignon and Hayward flowers that set harvestable berries, seed number per berry and mean seed weight in 2010/2011. Data are means ± SEM, n = Table 6.8: Effects of altering carbohydrate supply to Shiraz and Hayward kiwifruit shoots prior to anthesis on yield (total fresh berry weight) per shoot and its components mean berry weight and berry number per shoot in the 2009/2010 season. Data for berry number and mean berry weight were log xxiii

26 transformed for analysis and back transformed for presentation. Data are means ± SEM, n = Table 6.9: Effects of treatments on Shiraz, Cabernet Sauvignon and Hayward shoots prior to anthesis on yield (total fresh berry weight) per shoot and its components mean berry weight and berry number per shoot in the 2010/2011 season. Data for berry number and mean berry weight were log transformed for analysis and back transformed for presentation. Data are means ± SEM, n = Table 6.10: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the dry weight of Shiraz and Hayward leaves and stems at anthesis and at fruit harvest in the 2009/2010 season. Data are means ± SEM, n = Table 6.11: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the dry weight of Shiraz, Cabernet Sauvignon and Hayward shoots at anthesis and at fruit harvest in the 2010/2011 season. Data are means ± SEM, n = Table 6.12: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the leaf area of Shiraz and Hayward shoots at anthesis and at fruit harvest in the 2009/2010 season. Data are means ± SEM, n = Table 6.13: Effects of treatments to modify shoot carbohydrate supply on the leaf area of Shiraz, Hayward and Cabernet Sauvignon shoots at anthesis and at harvest in the 2009/2010 seasons. Data are means ± SEM, n = Table 6.14: Effects of treatments to modify carbohydrate supply applied prior to anthesis on the TNC concentration and content of leaves and stems of shoots at anthesis and at fruit harvest in the 2010/2011 season. Data are means ± SEM, n = Table 6.15: Photosynthetic light responses for Hayward, Shiraz and Cabernet Sauvignon leaves from fully exposed shoots at the end of the treatment period in the 2010/2011 season. P max is the light saturated maximum rate of photosynthesis, Rs is the rate of respiration, φ app is the apparent (CO 2 limited) photon yield and PFD sat is the photon flux density at which photosynthesis was light saturated. Data for each vine cultivar are mean values ± SEM, n = xxiv

27 Table 6.16: Coefficients and statistics of fitting a multiplicative model of photon flux density, stomatal conductance and internal CO 2 concentration to photosynthetic assimilation data. The percentage of error mean square (EMS) accounted for by each of the terms and the interaction between terms is presented. The model was highly significant (P < 0.001), n = Table 6.17: Effects of altering carbohydrate supply to Shiraz, Cabernet Sauvignon and Hayward kiwifruit shoots prior to anthesis on the net carbon balance of shoots during the treatment period (from DAB to fruit set DAB) and from fruit set until fruit harvest ( DAB) in the 2010/2011 season. Data are means ± SEM, n = Table 7.1: Overall summary of the timing of treatments to manipulate carbohydrate supply during floral development and the timing and effect of treatments on floral production relative to the untreated control (%) xxv

28 Certificate of Authorship I hereby declare that this submission is my own work and to the best of my knowledge and belief, understand that it contains no material previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at Charles Sturt University or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by colleagues with whom I have worked at Charles Sturt University or elsewhere during my candidature is fully acknowledged. I agree that this thesis be accessible for the purpose of study and research in accordance with normal conditions established by the Executive Director, Library Services, Charles Sturt University or nominee, for the care, loan and reproduction of thesis, subject to confidentiality provisions as approved by the University. Name: Annette Richardson Date: July 2014 Signature: xxvi

29 xxvii

30 Acknowledgements I would like to thank my principal supervisor Dr Dennis Greer for his constant support, his ability to solve problems and help with experiments in another country. He has allowed me to see vines from a different view, tackle photosynthesis and improve my writing skills. Thanks also to his wife Lindsay for making me feel very welcome on my visits to Wagga Wagga. I also thank my supervisors Dr Bruno Holzapfel for his thoughtful comments and wealth of knowledge of grapevines and Dr Ross Atkinson for helpful advice and his ability to view the research from a different perspective. I would particularly like to thank Dr Elspeth MacRae for getting the project started, providing endless encouragement, for many stimulating discussions and helping when life was challenging. Thanks also to Dr Gail Timmerman- Vaughan for her understanding, support and encouragement. I gratefully acknowledge the assistance of Rod and Cindy McIvor, Marsden Estate and Mark and Phillipa Brajkovich, Ivana Wines for providing unlimited access to vines. The companionship and help of Carno, Holly and Gonzo with excavations was always welcome. I would also like to acknowledge the support of Plant & Food Research. I thank my close colleagues Helen Boldingh, Dr Simona Nardozza and Dr Nick Gould who were always supportive and provided good advice. The continued interest and support of many other colleagues especially Dr Ross Ferguson, Dr Robert Schaffer, Dr Mike Currie and Dr Eric Walton as well as co-students Jill Stanley, Peter McAtee and Linda Boyd are gratefully acknowledged. I particularly wish to thank Helen Boldingh, Darienne Voyle and Judith Rees who provided invaluable help with carbohydrate analyses. I thank Duncan Hedderley for his patience and statistical advice. Special thanks to Peggy Kashuba who was a tower of strength and always available to help with field trials whenever she was needed. Thanks also to Mark Astill who provided helpful solutions to tricky problems, particularly constructing xxviii

31 the bird cage. The support from all of the staff of the Kerikeri Research Centre is also acknowledged. Finally I would like to thank my family for their enormous support and to dedicate this thesis to them. I thank James and Tim for providing technical and IT support whenever needed, dealing with possums and for graciously deferring holidays. Without the endless patience, love and support of my husband Mike this thesis would not have happened. I especially thank him for his unconditional support, skill in solving technical problems and general ability to keep our life on track throughout many challenging moments. xxix

32 Abstract The overall aim of this study was to determine whether carbohydrate supply from either photosynthesis or stored reserves limited inflorescence development in grape and kiwifruit vines in a similar way. Both grape and kiwifruit vines have a protracted period of inflorescence development, beginning with the initiation of floral meristems (all the processes needed to commit the meristem to produce an inflorescence) in spring of the first season, followed by a quiescent period before floral morphogenesis (development of four distinct whorls of floral organs - sepals, petals, carpels and stamens) occurs between budswell and anthesis in the second season. Patterns of growth, carbon acquisition and carbohydrate dynamics in Actinidia deliciosa Hayward kiwifruit and Vitis vinifera Shiraz and V. vinifera Cabernet Sauvignon grapevines were generally consistent across vines and seasons. Overall, growth of vines could be divided into two key stages each season: 1. From budbreak to 100 days after budbreak (DAB) when the productive capacity of shoots was established. Mobilisation of carbohydrate reserves supported the initial stages of floral morphogenesis and shoot development (0-30 DAB). Thereafter, shoots became autotrophic but perennial reserves continued to decline until the end of the stage when maximum shoot leaf area and berry numbers were established. 2. From DAB when shoot photosynthetic capacity supported berry development, root growth and carbohydrate accumulation in perennial tissues. Inflorescence initiation occurred during rapid shoot growth in the first season of floral development but was not directly influenced by carbohydrate supply in either Hayward or Cabernet Sauvignon vines. Carbohydrate supply to shoots was modified during inflorescence initiation by phloem girdling (to prevent phloem transport) and either defoliation or shading (to reduce photosynthesis). Only girdling and defoliation of Hayward shoots prevented inflorescence initiation, suggesting floral signals were leaf derived but not specifically carbohydrate. However, the growth of both Hayward and xxx

33 Cabernet Sauvignon shoots was also limited by carbohydrate supply during rapid shoot growth and this limited the capacity of shoots to store reserves in the autumn of that season. Carbohydrate reserves stored in canes and roots in autumn of the first season of floral development were essential to support early floral morphogenesis in both Hayward and Cabernet Sauvignon vines during the spring of the second season of floral development. These reserves were manipulated by modifying photosynthesis and phloem transport during shoot growth in spring or vine carbohydrate accumulation in autumn of the first season of floral development. Reducing photosynthesis 5-fold in spring reduced shoot growth and was associated with a decrease of up to 35% in cane carbohydrate storage. In autumn, reducing photosynthesis decreased carbohydrate storage in canes (by 15%) and roots (by 40-50%), and cane girdling increased cane reserves (by 17%). Modifying cane reserves altered inflorescence production of Hayward shoots by up to 2-fold each season. Girdling Cabernet Sauvignon canes increased inflorescence production of shoots by 30% in two seasons whereas, shading Cabernet Sauvignon vines only reduced inflorescence production by 30% after three seasons, suggesting both root and cane reserves were important for this cultivar. Reducing carbohydrate supply during late floral morphogenesis, particularly through lower photosynthetic rates of Hayward, Shiraz and Cabernet Sauvignon shoots decreased the number of flowers that reached anthesis (by 10-30%) and modified fruit set (by %), so that berry numbers were reduced by 10-30% at harvest. During this period of rapid shoot growth carbohydrate supply also determined the growth and photosynthetic capacity of shoots and hence their ability to support berry growth and reserve accumulation over the remainder of the season. Therefore carbohydrate supply was essential for floral morphogenesis to proceed in both grape and kiwifruit vines, with the early stages dependent on carbohydrate from vine reserves and the later stages on carbohydrate from photosynthesis. xxxi

34 Abbreviations A ABA ANOVA BB BBCH C c i d DAB DAS DM% DW E E-L EMS FS GA GLM g s IAA INF IWB n PAR PFD PFD sat P max REML R s SAM SEM photosynthesis abscisic acid analysis of variance budbreak BASF, Bayer, Ciba-Geigy and Hoechst carbon internal CO 2 concentration day days after 50% budbreak days after 1 September dry matter percent dry weight transpiration Eichorn and Lorenz error mean square floral budbreak gibberellic acid generalised linear model stomatal conductance indole-3-acetic acid inflorescences per floral shoot inflorescences per winter bud replicates photosynthetically active radiation photon flux density photon flux density at which photosynthesis was light saturated light saturated maximum rate of photosynthesis restricted maximum likelihood dark respiration shoot apical meristem standard error of the mean xxxii

35 TNC VPD VSP φapp total non-structural carbohydrate vapour pressure deficit vertical shoot position apparent (CO 2 limited) photon yield xxxiii

36

37 1 Literature review 1.1 Introduction Flowering is a key step in the life cycle of angiosperms, allowing recombination of genes for phenotypic variation that is the basis of natural selection and perpetuation of species Boss et al. (2004). This evolutionary advantage allowed angiosperms to dominate world flora through a diversity of life cycles and growth forms (Zik and Irish 2003). However, characteristics that have led to the evolutionary success of angiosperms (e.g. outcrossing mechanisms like dioecy and self incompatibility) are sometimes at odds with their commercial success (Sedgley 1988). Processes that were optimised to ensure seed dispersal in a specific environment may not be suited to modern day agriculture, where crops are grown as monocultures, in conditions far removed from their natural habitats. Examples of this are wine grapes (Vitis vinifera L.) and Hayward kiwifruit (Actinidia deliciosa (A. Chev.) C.F. Liang et A.R. Ferguson), which have evolved as unsupported vines, prioritising vegetative growth and delaying flower development in a high risk environment, thereby restricting flower production and yield in highly structured modern cultivation systems (Hopping 1990; Carmona et al. 2008). Crop yield is often determined by the success or failure of flowering. The productivity of cultivated crops is much greater than that of their wild ancestors, but there is still considerable variation in yields of different crops and between seasons. The short generation cycle of annuals has allowed rapid development of new cultivars and cultural techniques to increase yields and decrease seasonal variation; for example, annual variation in wheat (Triticum aestivum L.) yield is only 9% (Chloupek et al. 2004). In contrast, the longer breeding cycles of perennials has meant that yield improvement has largely resulted from development of cultural techniques and seasonal variation in yields remains high (for example, 33% in grapes) (Chloupek et al. 2004). Cultural techniques often use large quantities of energy, chemicals and labour to control flowering processes and hence yield (Carmona et al. 2008). In future, climatic change as well as limitations of labour supply and on the use of agricultural chemicals may make such practices unsustainable. 1

38 Therefore, there is considerable urgency to increase our knowledge of flowering in comparatively low yielding perennial crops in order to develop new cultivars and integrated management systems that sustain future yields. Flowering is a complex process, requiring coordination of environmental cues and endogenous factors from floral initiation through to seed set. Significant progress has been made in understanding the development of flowers in model annual crops, particularly the physiological and genetic control of flower induction and morphogenesis in Arabidopsis thaliana (L.) Heynh., Antirrhinum majus (L.) and Sinapsis alba (L.) (Bernier and Perilleux 2005; Corbesier and Coupland 2006; Andres and Coupland 2012). Extrapolating information to perennial crops, where flower development occurs over two seasons and plants must maintain concurrent vegetative and reproductive growth, is progressing (Brunner and Nilsson 2004; Andres and Coupland 2012). Progress has been made in understanding flowering in trees like poplar (Populus sp. L.) (Brunner and Nilsson 2004; Cseke et al. 2005; Jansson and Douglas 2007; Hsu et al. 2011) and birch (Betula pendula Roth.) (Lemmetyinen et al. 2004) and fruiting species like apple (Malus x domestica Borkh.) (Kotoda et al. 2000; Kotoda and Wada 2005; Kotoda et al. 2006; Kotoda et al. 2010), citrus (Citrus sp. L.) (Pillitteri et al. 2004b; Pillitteri et al. 2004a; Munoz-Fambuena et al. 2012) grape (Boss et al. 2003; Carmona et al. 2008) and kiwifruit (Varkonyi-Gasic et al. 2011; Varkonyi- Gasic et al. 2013). New information and reanalysis of existing knowledge should extend our understanding of factors influencing flowering and yields of perennial crops (Srinivasan and Mullins 1981; Dennis 2003; Goldschmidt and Samach 2004; Carmona et al. 2008; Bangerth 2009; Samach and Smith 2013). This will be particularly beneficial in crops like grape and kiwifruit, where yields are constrained by low flower production, but costs of production are high. 2

39 1.2 The flowering process Development of flowers in plants is an integrated, multi-step process that is usually divided into two major phases: floral initiation and floral morphogenesis (Bernier 1988). Floral initiation includes all the processes needed to commit the shoot apical meristem (SAM) to produce an inflorescence (Kinet 1993). Floral initiation begins with sensing of environmental or endogenous cues by leaves, the SAM or other organs, followed by the induction of flowering signals (Bernier and Perilleux 2005). These signals are transmitted to and integrated at the SAM and initiation is completed when the meristem is irreversibly committed to flowering (Bernier et al. 1981). Floral evocation is the last stage of floral initiation, referring only to events that occur at the apex during floral commitment (Evans 1969). Formation of floral meristems and development of four distinct whorls of floral organs (sepals, petals, carpels and stamens) occurs during floral morphogenesis (Bernier 1988; Jack 2004). Environmental and endogenous factors, including carbohydrate metabolism, regulate floral development, determining whether or not flowers reach anthesis (Bernier and Perilleux 2005; Corbesier and Coupland 2005) Floral initiation Which environmental cues induce flowering? Annual plants have a limited opportunity to flower and set seed (Bernier and Perilleux 2005; Andres and Coupland 2012). Therefore continuous, accurate and well integrated monitoring of the environment ensures plants flower during this window of opportunity. The primary environmental signals regulating floral transition in annual plants are photoperiod or vernalisation (a period of low temperatures between 1 and 10 o C), that reliably predict seasonal change (Bernier 1988; Kinet 1993; Andres and Coupland 2012). However, other more variable environmental cues, like ambient temperature, irradiance, water, nutrients and competition from neighbouring plants may also induce flowering time signals (Bernier 1988; Kinet 1993; Samach and Smith 2013). Complex regulatory systems allow integration of many 3

40 environmental signals in plants, to ensure successful flowering outcomes (Blazquez 2000; Corbesier and Coupland 2005; Matsoukas et al. 2012). Many genetic and physiological mechanisms are conserved during flowering of a wide range of species; however, in perennial species juvenility, concurrent vegetative and reproductive growth and an extended period of floral development add complexity to floral initiation compared with that in annuals (Boss et al. 2003; Brunner and Nilsson 2004; Carmona et al. 2008; Song et al. 2013). Seasonal rhythms in trees like poplar and conifers, which are native to northern latitudes, are strongly influenced by photoperiod (Bohlenius et al. 2006; Gyllenstrand et al. 2007; Jansson and Douglas 2007). In other perennials like apple (Dennis 2003; Solomakhin and Blanke 2008) and grape (Boss et al. 2004), a high daily light integral may stimulate flower initiation. Moderate ambient temperatures rather than low temperature vernalisation are also major flowering time cues in some perennials, for example o C for grape, 20 o C in blueberry (Vaccinum corybosum L.) (Wilkie et al. 2008) and o C for subtropicals like citrus, avocado (Persea americana Mill.) and mango (Mangifera indica L.) (Albrigo and Sauco 2004). Environmental flowering time signals are perceived by several plant organs. Photoperiod is generally sensed by leaves, irradiance can be perceived by all photosynthetic organs and light quality is sensed at the SAM (Kinet 1993; Bernier and Perilleux 2005). Vernalisation is generally detected by the SAM and associated leaf primordia, but in some plants the signal may also be sensed by leaves (Bernier 1988; Putterill et al. 2004). Ambient temperature can be sensed by all plant organs (Bernier and Perilleux 2005), while water and nutrient supply are sensed by roots (Bernier et al. 1981) Which endogenous signals are involved in floral initiation? A number of endogenous compounds, including sugars, nutrients and hormones, act as signals during floral initiation (Bernier 1988; Bernier et al. 1993; Matsoukas et al. 2012). Sucrose has multiple roles in the plant s transition to flowering, including signalling, stimulating gene expression and 4

41 controlling carbon and nitrogen supply (Corbesier et al. 1998; Bernier and Perilleux 2005; Matsoukas et al. 2012). Sucrose concentrations in phloem increase significantly during the early stages of floral induction and when this is prevented, flowering is inhibited (Corbesier et al. 1998). Labelling studies have shown that elevated sucrose concentrations arise from breakdown of starch reserves in leaves and stems during flower induction, rather than directly from newly fixed carbon (Lejeune et al. 1991; Lejeune et al. 1993). In suboptimal photosynthetic conditions, application of sucrose to plants can promote flowering (Bagnall and King 2001). Sucrose is also the main factor controlling loading of signalling compounds into phloem in the leaf, and unloading at the SAM (Corbesier et al. 1998). Sucrose, together with gibberellic acid (GA), is needed to increase expression of the key meristem identity gene LEAFY at the SAM (Blazquez et al. 1998; Eriksson et al. 2006). Hydrolysis of sucrose to glucose and fructose (hexoses) by invertase supplies energy to the SAM (Roitsch et al. 2000; Heyer et al. 2004; Koch 2004) and such source sink interactions may involve trehalose-6-phosphate signalling (Lunn et al. 2006). Hexoses, together with cytokinins, also stimulate cell division during floral initiation (Potuschak and Doerner 2001; Jacqmard et al. 2003). A complex relationship between carbon and nitrogen signalling has also been discovered, where sugar signals from leaves stimulate nitrogen export from roots via leaves to the SAM (Corbesier et al. 2001). Studies of endogenous factors known to influence flowering in annuals have been limited to observations of whole plant manipulations in perennial species and results from these studies must be carefully interpreted (Goldschmidt and Samach 2004). In perennials modifying carbohydrate supply by girdling (Cohen 1981; Goldschmidt et al. 1985; Snelgar and Manson 1992; Garcia-Luis et al. 1995; Dennis 2003), root pruning (Khan et al. 1998; McArtney and Ferree 1999a), reduction of crop load (Jonkers 1979; Goldschmidt and Golomb 1982; Burge et al. 1987) and shading or leaf removal (Candolfi-Vasconcelos and Koblet 1990; Snelgar and Manson 1992; Sommer et al. 2000) influences flower initiation, but how this occurs is not yet clear. Some studies suggest that the size of apple, grape and Meterosideros 5

42 excelsa (Sol. Ex Gaertn.) buds affects initiation of flowers, and bud size may be influenced by carbohydrate and GA supply (Antcliff and Webster 1955; Sommer et al. 2000; Bertelsen et al. 2002; Henriod et al. 2003). The ability of buds to take up carbohydrates, via the activity of enzymes like NADdependent sorbitol dehydrogenase and acid invertase that break down imported sugars for utilisation in bud development, has also been shown to affect floral initiation in Japanese pear (Pyrus pyrifolia (Burm.f) Nak.) (Ito et al. 2002; Ito et al. 2004a; Ito et al. 2004b). The hormone GA plays a key role either promoting flowering in plants like Arabidopsis and Lolium temulentum (L.) (Wilson et al. 1992; King and Evans 2003; King 2012), but generally inhibits floral initiation in perennial species like apple (Tromp 2000; Bertelsen et al. 2002), grape (Boss and Thomas 2002), cherry (Prunus avium L.) (Lenahan et al. 2006) and citrus (Goldschmidt et al. 1997). High concentrations of GA, produced by the seeds of a heavy fruit crop, reduce initiation of flowers for the following season, causing biennially bearing cycles in many perennial fruiting tree crops (Bangerth 2006; Samach and Smith 2013). Stimulation of vegetative growth and export of GA from actively growing apices may also reduce floral initiation in vigorous perennials (Boss et al. 2003; Bangerth 2009). Differing responses of annuals and perennials to GA may occur because vegetative meristems are terminated by floral initiation in annuals, whereas vegetative and reproductive meristems may exist concurrently in perennials and development of the current season s crop coincides with initiation of flowers for the following season. Hence the impact of competition for carbohydrates during floral initiation may be greater in perennials, particularly in vines that produce vigorous vegetative growth throughout much of the season. Cytokinins exported from both leaves and roots stimulate enzyme activity (e.g. invertase) and cell division at the SAM during floral initiation in annuals (Corbesier et al. 2003) and cytokinins have been associated with flower induction in litchi and apple (Bangerth 2009). Abscisic acid, salicylic acid, ethylene and brassinsteroids may also act as signals during floral initiation in some annual species (Bernier et al. 1981; Boss et al. 2004; Bernier and Perilleux 2005). 6

43 Considerable progress has been made in understanding how primary flowering time cues are perceived, transmitted and integrated with other processes at the SAM during floral initiation in annual plants. Sugars appear to have multiple roles in floral initiation, from loading of flowering time signals into phloem, to up-regulating expression of genes and stimulating cell division at the apex. Floral initiation is not well understood in perennials, as floral development occurs over a much greater time span with concurrent vegetative growth. Correlative studies suggest that manipulating carbohydrate supply to buds can affect floral initiation, but it is not clear whether this is due to effects of sugars on signalling, stimulation of activity at the SAM or simply to increased growth of perennial buds Floral organ morphogenesis Eudicots, representing about 75% of angiosperms, typically produce flowers with four different classes of floral organs (sepals, petals, stamens and carpels) organised in four or more whorls (Jack 2004). Simple variations in this architecture have resulted in a huge variety of flower sizes, shapes and colours (Zik and Irish 2003). Once the SAM has undergone floral initiation, several processes are required to complete flower development, including formation of the floral meristem, establishment of floral organ identities, differentiation of floral organs and meiosis (Zik and Irish 2003; Irish 2010). These different stages of floral morphogenesis vary in their sensitivity to environmental and endogenous factors (Cawoy et al. 2007). However, the processes that occur during floral morphogenesis are not as well understood as those involved in floral initiation How do environmental factors influence floral morphogenesis? In contrast to floral initiation, floral morphogenesis in annuals is influenced but not regulated by environmental factors (Kinet 1993). In grain crops like wheat and barley (Hordeum vulgare L.), high irradiance stimulates photosynthetic activity in source leaves, and sometimes flowers themselves, increasing assimilate supply and hence the size of the inflorescence (Kinet et al. 1985). Light may also stimulate the enzyme activity in flowers (sinks) in 7

44 maize (Zea mays L.) (Makela et al. 2005). High temperatures can cause flower abortion in pepper (Capsicum annum L.) (Turner and Wien 1994), cowpea (Vigne unguiculata (L.) Walp.) (Mutters and Hall 1992) and tomato (Rudich et al. 1977), while low temperatures increase the duration of ear development and increase the number of florets on each spike in wheat (Warrington et al. 1977). High temperatures may affect sink strength of flowers, for example by reducing the activity of acid invertase in pepper flowers (Aloni et al. 1991). Effects of environmental factors on floral morphogenesis of perennials appear to be similar to those in annuals. Extreme temperatures can cause flower abortion at the beginning of floral morphogenesis in grapes (Petrie and Clingeleffer 2005), apple (Palmer et al. 2003), blackberry (Rubus L. subgenus Rubus Watsch.) (Takeda et al. 2002) and kiwifruit (Richardson et al. 2004). High or fluctuating temperatures during the latter stages of flower development and meiosis, prior to anthesis, also cause abnormalities in apple (Oukabli et al. 2003) and grape (Ebadi et al. 1996) flowers, by affecting their ability to form vascular connections and reduce fruit set in blackberries through reduced pollen viability and stigma receptivity (Stanton et al. 2007) Which endogenous signals affect floral morphogenesis? Carbohydrate supply has a central role in floral morphogenesis of annual plants. Altering carbohydrate supply through shade (Perilleux et al. 1991; Aloni et al. 1996), temperature (Aloni et al. 1991), defoliation (Cawoy et al. 2007), water stress (Andersen et al. 2002; McLaughlin and Boyer 2004; Sun et al. 2004) and competition from other organs (Bertin et al. 2002), may affect all stages of flower development. Early events, such as inflorescence branching, may be modified through the effects of sugar signals on gene expression, for example trehalose-6-phosphate sugar signalling in maize (Satoh-Nagasawa et al. 2006), the regulation of sugar uptake into developing Arabidopsis flowers (Heyer et al. 2004), or the expression of cyclins and cyclin-dependent kinase during cell proliferation in tomato (Baldet et al. 2006). However, the latter stages of anther, carpel, ovule and pollen morphogenesis including meiosis are highly susceptible to changes in the 8

45 amount and composition of carbohydrates supplied to developing organs (Durdan et al. 2000; Andersen et al. 2002; McLaughlin and Boyer 2004; Sun et al. 2004; Cawoy et al. 2007). This can profoundly alter yield through the number of flowers which develop, the size of reproductive organs and their ability to set seed (Bohner and Bangerth 1988; Bertin et al. 2002; Boyer and McLaughlin 2007). Carbohydrate supply also has a key role in the morphogenesis of perennial flowers. This has been clearly demonstrated in perennials like grape where floral morphogenesis is completely separated from inflorescence initiation and there is also intense competition for carbohydrate supply between vegetative and reproductive growth. Low starch reserves in grapevines at budbreak were associated with early flower abortion (Bennett et al. 2005), and low carbon fixation in leaves and the inflorescence itself during the final stages of flower development caused reduced fruit set (Candolfi- Vasconcelos and Koblet 1990; Lebon et al. 2004b; Lebon et al. 2005). Temperature, light, and water stress prior to anthesis also influence flower development via carbon metabolism in grape (Lebon et al. 2008) and blackberry (Stanton et al. 2007). Carbohydrate effects on floral morphogenesis in tree crops like apple (Jonkers 1979; Khan et al. 1998; Dennis 2003), pear (Jonkers 1979) and citrus (Goldschmidt and Golomb 1982; Goldschmidt et al. 1985; Garcia-Luis et al. 1995), are difficult to separate from effects on floral initiation, as the two processes occur sequentially. However, altering the amount of reserves stored in fruit trees and vines over winter can affect the later stages of floral morphogenesis and the ability of flowers to set fruit (Loescher et al. 1990; Zapata et al. 2004b). Low light and nutrient concentrations may also impair floral development particularly during the last stages of floral morphogenesis. Shading apricot trees (Prunus armeniaca L.) during meiosis caused low carbohydrate concentrations in flowers and poor pistil development, resulting in poor fruit set and fruit development (Nuzzo et al. 1999). In grapes, flower quality was reduced and flowers aborted when vines were shaded during meiosis and the final stages of floral morphogenesis (Jackson 1991; Keller and Koblet 1994; Lebon et al. 2004b). 9

46 It has also been suggested that different hormone groups combine to regulate floral organ development during floral morphogenesis. Initially the size of the floral meristem may be influenced by cytokinins, gibberellins and auxins (Chandler 2011). Auxins also have key roles in the development of all floral organs but are particularly important in the development of the gynoecium (Nemhauser et al. 2000). Gibberellins are essential for the development of both petals and stamens of flowers (Koornneef and van der Veen 1980). Cytokinins are necessary for pollen development (Huang et al. 2003) and ethylene may be involved in floral sex determination (Kahana et al. 1999). The link between environmental conditions, endogenous signals and genetic regulation of floral morphogenesis is not as well understood as it is for floral initiation. This may be partly because floral initiation and floral morphogenesis occur consecutively in many plants and it is difficult to separate causal effects. However, it appears that carbohydrates play a central role in coordinating environmental and endogenous signals during floral morphogenesis of both annuals and perennials. Floral morphogenesis is particularly sensitive to changes in carbohydrate supply during the first events of floral organ differentiation and the last stages of sex organ development. Changes in light, temperature and water supply to plants can affect the activity of enzymes involved in sugar unloading and therefore the sink strength of developing flowers. Sucrose and hexoses can also affect the expression of genes controlling floral organ development. Processes controlling floral morphogenesis in perennials may be similar to those in annuals; however, current understanding of the process in perennials is limited. 10

47 1.3 How does flowering behaviour of perennials affect crop yield? Are yields of perennial fruiting crops similar? Fruits or seeds form the saleable yield of many perennial crops and therefore floral development is a major determinant of yield. In fruit crops, fruit number generally has the greatest affect on yield, and is determined by the number of flower or inflorescences produced on a plant, fruit set and fruit retention (Coombe 1976; Mullins et al. 1992). Growth rates of the ovary both pre- and post-anthesis can also affect yield (Coombe 1976; Monselise and Goldschmidt 1982). Perennial fruiting crops produce a wide range of yields, and fruit set data indicate that the range in flower numbers is considerably greater (Table 1.1). Apple, pear and citrus trees are extremely efficient, producing fruit yields of tonnes ha -1. In other tree crops, like stonefruit, where there can be difficulties with pollination (Zucconi 1986), or in trees that produce fruit or nuts containing lipids, like avocado, olive, walnut or macadamia (Macadamia integrifolia Maiden and Betche), yields tend to be lower despite high flower numbers (Wolstenholme 1986). Perennial crops with other plant forms, like grape, kiwifruit and passionfruit (Passiflora sp. L.) vines, brambles or shrubs, generally produce much lower yields as a result of lower flower production. In addition, these non-arboreal crops are often not self-supporting and generally require much higher cultural inputs, for example with training and pruning, to achieve these modest yields (Carmona et al. 2008). Low yields and high management costs have a major impact on the profitability of non-arboreal crops. 11

48 (t ha -1 ) Trees Apple Pear Table 1.1. The range in yield and fruit set (the percentage of flowers that produce mature fruit) for common perennial fruiting crops. Species Yield Fruit- References set (%) 0.2- McKenzie and Rae (1978); Dennis (1986); Palmer et al. 10 (2002); Palmer and Dryden (2006) 5- Stephenson (1981); Jackson and Palmer (1999); 10 Ramos et al. (2008) Peach Stephenson (1981); Zucconi (1986); Looney and Jackson (1999); Jimenez and Diaz (2003) Cherry Stephenson (1981); Looney and Jackson (1999); Granger (2005) Orange Stephenson (1981); Thornton and El-Zeftawi (1983); Spiegel-Roy and Goldschmidt (1996) Grapefruit Thornton and El-Zeftawi (1983); Spiegel-Roy and Goldschmidt (1996) Avocado 5-20 <0.1 Stephenson (1981); Wolstenholme (1986); Morley- Bunker (1999a) Olive Lavee (1986); Morley-Bunker (1999b) Walnut Jackson and McNeil (1999); Rovira et al. (2001); Asadian and Pieber (2005); Ninot et al. (2006) Macadamia Stephenson (1981); Morley-Bunker and McNeil (1999) Vines, Shrubs and Brambles Grape Coombe (1988); Mullins et al. (1992); Bramley and Hamilton (2004); Palliotti et al. (2007) Kiwifruit Hopping (1990); Sale and Lyford (1990); Smith and Buwalda (1994) Passionfruit Morton (1987); Sale (2003) 70 Blueberry Moore et al. (1993); Smagula (1993); Lyrene and Williamson (1997); Thiele (1999) Raspberry Redalen (1976); Thiele (1999); Heidenreich et al. (2008) 100 Brambles Thiele (1999); Heidenreich et al. (2008); Thompson et al. (2008) 12

49 Does flowering behaviour of perennials affect yield? Many perennial fruiting tree crops like apple, pear (Pyrus communis L.), peach, orange (Citrus sinensis L. Osbeck), grapefruit (Citrus x paradise Macfad.), avocado and olive have a strong tendency toward biennial bearing. They produce an excessive number of flowers and fruitlets in the first season and this severely limits resources available for initiation of the following season s flowers. Generally crops that exhibit this flowering pattern produce flowers with comparatively small ovaries and only invest resources in their growth once fruitlet numbers have been selectively reduced (Coombe 1976). Regulation of biennial bearing in fruit trees is well understood and flowering behaviour can be successfully manipulated to maximise yields within these high cropping tree species for commercial gain (Jonkers 1979; Dennis 1981; Goldschmidt and Golomb 1982; Monselise and Goldschmidt 1982; Garcia- Luis et al. 1995; Goldschmidt et al. 1997; Dennis and Neilsen 1999; Goldberg-Moeller et al. 2013; Samach and Smith 2013). By contrast, regulation of flowering and yield is not well understood in other major fruit crops, like grapevines and kiwifruit that have different growth and flowering patterns from those of biennial bearing trees. In both vines, inflorescence meristems are initiated in the season before flowering but flowers develop during spring, prior to anthesis (Pratt 1971; Srinivasan and Mullins 1981; Walton et al. 2001). Grapevines (Coombe 1976; Mullins et al. 1992; May 2000; Shavrukov et al. 2004) and kiwifruit (Davison 1990; Hopping 1990) produce comparatively few inflorescences, but high proportions of flowers set and produce fruit. Grapes and kiwifruit also produce flowers with comparatively large ovaries, for example Corinth grape ovaries increase only 300-fold (Coombe 1976) and Actinidia chinensis Planch. var. chinensis Hort16A 500-fold (Richardson et al. 2011) between anthesis and harvest compared with a 15,000-fold increase in the size of apple ovaries (Tukey and Young 1942) and a 300,000-fold increase in size of avocado ovaries (Coombe 1976) over this period. This contrasting flowering behaviour of grape and kiwifruit vines may be due to the growth pattern and priority for resource allocation in lianas (woody vines) (Mooney and Gartner 1991). 13

50 Does the growth pattern and allocation of resources affect flowering? In their natural environment, vines do not produce a strong aerial perennial structure, but must seek the support of other plants to grow into suitable positions in the canopy (Rowe and Speck 2005; This et al. 2006; Carmona et al. 2007b). This strategy carries much greater risk and uncertainty than growth as a stable free-standing tree and this affects how vines allocate resources between vegetative and reproductive growth (Fisher and Ewers 1991; Mooney and Gartner 1991; Mooney and Winner 1991). Vigorous vegetative growth is a priority for vines throughout much of the growing season (Coombe 1995; Greer et al. 2004). In spring new shoots can grow at a rate of 182 mm.week -1 in grapes (Coombe 1995) and 420 mm.week -1 in kiwifruit (Greer and Jeffares 1998), four - nine times the peak growth rate of apple tree shoots (Abbott 1984), and vine shoots may exceed 3 m in length by the end of the season. Although vines invest few resources in a perennial aerial woody structure, they have a large leaf area and a high root: shoot ratio (1:2) (Castellanos 1991; Mullins et al. 1992; Smith et al. 1992; Greer and Jeffares 1998; McArtney and Ferree 1999a; Bates et al. 2002) compared with trees which typically have a root: shoot ratio of 1:5 or 1:6 (Harris 1992; Palmer 2007). This arrangement enables vines to readily access a favourable light environment above the tree canopy and to support any structural repairs needed (Fisher and Ewers 1991; Carmona et al. 2007b). The patterns and amount of photosynthate produced and stored in deciduous vines and trees are similar (Lakso et al. 2001; Greer et al. 2004; Poni et al. 2007; Holzapfel et al. 2010). In spring, young leaves initially act as carbohydrate sinks, but become net photosynthate exporters when they reach 50% of their final size (Koblet 1969; Kriedemann et al. 1970; Lakso et al. 1999; Petrie et al. 2000; Greer et al. 2003). In grapevines, inflorescences and young berries are able to reassimilate respiratory carbon dioxide (Lebon et al. 2005; Palliotti et al. 2010). By anthesis, the carbon balance of trees and vines becomes positive and photosynthesis reaches equivalent peak rates in mid-summer (Lakso et al. 1989; Corelli-Grappadelli and Magnanini 1993; Petrie et al. 2000; Greer et al. 2004; Wunsche et al. 2005; Poni et al. 2007) before photosynthetic rates decline after harvest (Loescher et al. 1990; 14

51 Greer and Wunsche 2003; Greer et al. 2004). Deciduous fruiting trees and vines generally produce and store excess carbohydrates during the growing season, unless current yields are very high, in order to provide resources for overwintering and early development in spring (Loescher et al. 1990) and as a buffer for periods of stress (Sala et al. 2012). Under normal circumstances replenishment of carbohydrate reserves begins when shoot growth declines in summer and continues until leaf fall (Priestley et al. 1976; Smith et al. 1992; Bennett et al. 2005; Holzapfel et al. 2010). Reserves stored after harvest are particularly important (Greer and Wunsche 2003; Holzapfel et al. 2006; Smith and Holzapfel 2009) as these are among the first used in spring for bud growth (Scholefield et al. 1978). To summarise, perennial fruit crops produce a wide range of yields, from just a few tonnes per hectare in vines to 180 t ha -1 in apple. High cropping fruit trees produce an excess of flowers, have low fruit set and flowers with small ovaries, flowering has been extensively studied in these tree crops and their behaviour can be modified to optimise yields. In contrast, perennial fruit crops with unsupported growth forms, like grapevines and kiwifruit, produce comparatively few flowers, their flowering behaviour is not well understood and fruit yields are low. In vines, vegetative growth is allocated a much greater share of resources than reproductive growth from budbreak until anthesis, to ensure vines survive any mechanical damage and are in a high light environment suited to flowering. As part of this strategy, floral morphogenesis is largely delayed until spring, but limited carbohydrate supply and intense competition for those carbohydrates affect the number of flowers that vines produce. Further understanding of carbohydrate effects on flowering in vines is needed in order to improve the production systems and optimise yield of vines. 15

52 1.4 Is low flower production typical of fruiting vines? Are the flowering behaviours of grape and kiwifruit vines similar? Grape and kiwifruit are the most successful temperate, deciduous, fruiting lianas (woody vines) in commercial cultivation. The only other widely grown fruiting lianas, the Passiflora species, (Passiflora edulis Sims. (passionfruit), P. tripartite var. mollissima (Kunth) Holm-Niel and Jorg (banana passionfruit) and P. ligularis Juss. (sweet granadilla), are also low yielding, but not directly comparable to grape and kiwifruit because they are subtropical evergreens that flower throughout much of the growing season. Both kiwifruit (Sale and Lyford 1990) and grape vines (Carmona et al. 2008) require high cultural inputs to produce relatively low and seasonally variable yields (average t ha -1 for grapes (Mullins et al. 1992; Bramley and Hamilton 2004) and t ha -1 for Hayward kiwifruit), (Sale and Lyford 1990; Smith and Buwalda 1994). Although there is some genetic diversity, both Vitis and Actinidia species are typically low yielding (Thorp et al. 1990; Mullins et al. 1992). Although vine yields vary seasonally they generally do not exhibit classic biennial bearing patterns (McPherson et al. 1994; McPherson et al. 2001; Chloupek et al. 2004; Bennett et al. 2005). Comparatively low yields are due to the low number of inflorescences produced by both vines (Hopping 1990; Clingeleffer 2006) and may also be due to reduced flower numbers per inflorescence in grape (Clingeleffer 2006). Grape buds can produce up to four (Pratt 1971; May 2004), and Hayward kiwifruit up to nine inflorescences per shoot (Brundell 1975c; Hopping 1990) but many floral meristems fail to develop or inflorescences abort part way through development. Typically grapevines produce between one and two inflorescences per shoot (Bennett et al. 2005; Keller 2010), while Hayward kiwifruit generally produce four (Brundell 1975c). Grape inflorescences have three or four orders of branching before ending in a compound dichasium with three to five flowers (Carmona et al. 2007a; Vasconcelos et al. 2009). A Hayward kiwifruit inflorescence is a compound dichasium with a terminal and two lateral flowers (Hopping 1990) although in other species, e.g. Actinidia guilinensis (C.F. Liang) and Actinidia latifolia (Gardn. et Champ.) 16

53 Merr. inflorescences may contain 56 and 198 flowers respectively (Snowball 1997b). Although grape and kiwifruit vines exhibit considerable seasonal variation in flowering and yield, they do not follow extreme biennial bearing cycles of many tree crops with little fruit drop after anthesis (Sale and Lyford 1990; Keller et al. 2004). In kiwifruit 90% of flowers develop into mature fruit (Hopping 1990), and around 50% of grape flowers produce mature fruit, although this may range from 20 to 80% depending on the cultivar and season (Mullins et al. 1992; Keller 2010). Grape and kiwifruit flowers have comparatively large ovaries at anthesis and therefore grow more prior to anthesis and less between anthesis and harvest than other fruit crops. For example, the ovary of Corinth grapes increases 300 times in size between anthesis and harvest (Coombe 1976) while the ovary of A. chinensis Hort16A kiwifruit increases 500 times (Richardson et al. 2011). Low productivity in both grape and kiwifruit vines is not directly related to their genetic and cultural history. Grapes are native to the Transcaucasia region of Eurasia, have a long history of cultivation and are botanically classified as rosids like apple, stonefruit and Arabidopsis (Jansson and Douglas 2007; Goldschmidt 2013). In contrast, kiwifruit are native to Asia, were collected only from the wild until 40 years ago, and are asterids like Antirrhinum, petunia, solanums and blueberries (Ferguson and Huang 2007; Jansson and Douglas 2007). Highly domesticated modern grapevines have many characteristics, including the production hermaphroditic flowers that demonstrate their long cultural history (Goldschmidt 2013). The wild grapes (Hardie 2000) were, however, dioecious like kiwifruit (Hopping 1990). The difference in the evolutionary histories of these vines suggests that relatively low productivity may be a characteristic of vine behaviour particularly their patterns of growth and floral development Are floral initiation and morphogenesis in grape and kiwifruit similar? Inflorescence initiation in both kiwifruit (Snelgar and Manson 1992) and grape vines (Lavee et al. 1967; Pratt 1971; Srinivasan and Mullins 1981; 17

54 Watt et al. 2008) begins months before flowers reach anthesis, during the peak of vegetative growth. This has been confirmed by expression of orthologues of the Arabidopsis meristem identity genes LEAFY and AP1 (Walton et al. 2001; Boss et al. 2003; Carmona et al. 2008). The first bud that develops in the leaf axil of a developing grape shoot in spring generally grows to form a summer shoot (Srinivasan and Mullins 1981). Then the primary latent bud forms in the axil of the first bract on the summer shoot and produces on average two bracts, six to ten leaf primordia and two to four uncommitted primordia (anlagen) (Pratt 1971; Srinivasan and Mullins 1981). Secondary and tertiary latent buds form in the axils of the bracts on the primary latent bud, and together, these three latent buds form a compound bud (Pratt 1971; Srinivasan and Mullins 1981). Uncommitted primordia on the primary latent bud divide into two arms and undergo up to three orders of branching, or develop into a tendril during the two months after initiation (Pratt 1979; Srinivasan and Mullins 1981; Boss et al. 2003) (Figure 1.1). Kiwifruit buds develop in a similar way to those of grape but produce only primary and secondary latent buds in the leaf axils of current season s shoots. The primary latent bud produces four bracts and leaf primordia (Brundell 1975c; Walton et al. 1997). Secondary latent buds form in the axils of the four basal bracts and simple dome-shaped meristems form in the axils of the remaining leaf primordia of the primary latent bud (Brundell 1975c; Walton et al. 1997). As kiwifruit (Watson and Gould 1994; Snowball 1996) and grape (May 2000; Sommer et al. 2000) shoots continue to grow during summer, buds at more distal nodes progressively undergo inflorescence initiation and branching, but later developing buds have fewer structures present in the meristem (Snowball 1996). 18

55 Figure 1.1: Key stages of floral development in grape buds. Reproduced from Dunn and Martin (2007). 19

56 Table 1.2: Key stages of floral morphogenesis in grape buds using the Eichorn-Lorenz (E-L) and BASF, Bayer, Ciba-Geigy and Hoechst (BBCH) models. Reproduced from Coombe (1995) and May (2004). Time from bud E-L growth BBCH growth Flower bud burst (days) stage stage development Dormant 1 00 inflorescence primordia floral meristem sepals corolla stamens emergence of carpel placenta and ovule primordia development of style and stigma meiosis formation of embryo sac capfall and pollination fertilisation 20

57 Season 2 Season 1 Initiation (buds) Dormant (buds) Inflorescence number and size Inflorescence and flower size Morphogenesis Figure 1.2. Key stages in the floral development of kiwifruit buds as described by Brundell (1975c) and Walton et al. (1997). After initiation and early inflorescence branching, grape (Pratt 1971; Srinivasan and Mullins 1981) and kiwifruit (Walton et al. 2001) buds become quiescent for several months during autumn and winter. It is not known why this occurs but it may be because of vigorous growth of new shoots, sequential initiation of new buds along shoots, high demand for carbohydrates by other sinks or delayed investment in reproductive growth in case vines are damaged or no longer in a suitable environment for flowering 21

58 (Carmona et al. 2007b). Quiescent buds can be forced to grow soon after they develop, by removal or abortion of the shoot apex and by providing suitable growing conditions. In grapevines, this stimulates growth of a summer shoot from the prompt bud in the last intact leaf axil and a small inflorescence or a transitional structure between a tendril and an inflorescence sometimes forms (Mullins et al. 1992; May 2000; Boss et al. 2003). In contrast, removing the apex of a kiwifruit shoot rarely produces a floral shoot from buds of the main commercial cultivar A. deliciosa Hayward, but may produce flowers in A. eriantha (Benth.) vines (Walton 1995). Floral organ morphogenesis begins during the earliest stages of budswell in spring and continues for a protracted period until anthesis in grape (Srinivasan and Mullins 1981; Boss et al. 2003) and kiwifruit buds (Brundell 1975c; Polito and Grant 1984) (Figures 1.1 and 1.2, Tables 1.2 and 1.3). During budswell in early spring, further branching of the grape inflorescence primordia may occur and then, at budbreak, each inflorescence primordium divides to form three to five floral primordia (May 2000; May 2004; Carmona et al. 2007b). Each floral primordium produces sepals, petals, stamens and a gynoecium with two ovules in each ovary (Srinivasan and Mullins 1981; Boss et al. 2003; May 2004) (Table 1.2). Floral morphogenesis occurs in a similar manner in both staminate and pistillate kiwifruit vines, with only minor differences in the timing of events in terminal flowers (Brundell 1975c; Hopping 1990) (Table 1.3). At budswell, 10 days before budbreak, kiwifruit inflorescence meristems form bracts, lateral flower meristems and sepal primordia. Then after budbreak, petals, stamens and a gynoecium form sequentially, with each pistillate flower containing around 35 carpels, each with 20 ovules (Brundell 1975c; Polito and Grant 1984; Hopping 1990). 22

59 Table 1.3: Key stages of floral morphogenesis in kiwifruit buds (Brundell 1975c; Polito and Grant 1984; Hopping 1990). Time from Axillary bud Flower bud development budburst (days) development Pistillate ( Hayward ) Staminate ( Alpha ) Dormant uninitiated flower primordium -10 budswell swollen flower primordium -6 advanced budswell initiation of bracts initiation of sepals 0 budburst initiation of petals 5 advanced budburst initiation of stamens 9 initiation of stigma 10 open cluster 11 initiation of stigma 15 advanced open cluster 16 gynoecial development stops 18 formation of gynoecial plateau 27 anther and filament initiation 34 anther and filament initiation 38 pollen grain formation 40 calyx split 45 ovule initiation 50 calyx split 53 open flower 60 open flower 23

60 When is flowering potential in grape and kiwifruit buds lost? Flowering potential can be lost at three stages: 1. during inflorescence initiation and bud development in the preceding season, 2. during early floral morphogenesis around budbreak, and 3. later in floral morphogenesis during the development of sex organs in grape (Srinivasan and Mullins 1981; May 2000; Dunn and Martin 2007; Lebon et al. 2008) and kiwifruit vines (Brundell 1975c; Grant and Ryugo 1982; Piller et al. 1998). In grapes, the number of inflorescences that develop is determined during initiation in the preceding season (Candolfi-Vasconcelos and Koblet 1990; Bennett et al. 2005; Watt et al. 2008). This may also be true of kiwifruit, as all buds are potentially floral but the number of floral nodes within a bud may vary (Brundell 1975a; Grant and Ryugo 1982; Watson and Gould 1994; Snowball 1997a). The number of flowers that develop on a grape inflorescence is largely determined by the degree of branching of the inflorescence that occurs either immediately after initiation or during the early stages of floral development in spring (Dunn and Martin 2007; Lebon et al. 2008). In kiwifruit buds, primordia at node five often fail during the early stages of floral morphogenesis, while those at higher nodes may abort at later stages (Brundell 1975c; Walton and Fowke 1993). Disrupting pollen and ovule formation in both grape (Lebon et al. 2004b; Petrie and Clingeleffer 2005) and kiwifruit (Piller et al. 1998) flowers during the final stages of floral morphogenesis affects both fruit set and the early growth of fruit When do carbohydrates limit flower development of grape and kiwifruit? Does carbohydrate supply affect inflorescence initiation and early bud development? Manipulating grape (Candolfi-Vasconcelos and Koblet 1990; Bennett et al. 2005) and kiwifruit (Snelgar and Manson 1992) vines by leaf removal, shading and girdling during summer and autumn of the season preceding anthesis affects their subsequent flower production. Inflorescence initiation in these vines appears to be sensitive to leaf signals but it is not known how leaf removal affects sugar or other flowering time signals (Lavee et al. 1967; Candolfi-Vasconcelos and Koblet 1990; Snelgar and Manson 1992; Snowball 24

61 1996). The size of developing grape (Sommer et al. 2000) and kiwifruit buds (Watson and Gould 1994; Snowball 1996) affects their subsequent capacity to produce flowers and bud development may be influenced by carbohydrate supply, but this has not been investigated. These studies suggest that carbohydrates may have various roles during inflorescence initiation and early development of grape and kiwifruit buds, but how they affect these processes has not been described Do the amounts of stored reserves affect floral morphogenesis? Vines accumulate large amounts of carbohydrates, in their roots, trunk, perennial canopy structure and shoots/canes from soon after anthesis until leaf fall (Smith et al. 1992; Bates et al. 2002; Holzapfel et al. 2010). These reserves supply 30-60% of the carbohydrates needed for early shoot development in spring (McArtney and Ferree 1999b) and are the major source of carbohydrates for developing grape and kiwifruit shoots until three to four weeks after budburst but continue to support shoot development until anthesis (Lebon et al. 2008). Defoliating or shading vines during the period of reserve accumulation in autumn can reduce flower numbers by 50% in the following season (Candolfi-Vasconcelos and Koblet 1990; Snelgar et al. 1991; Snelgar et al. 1992; Manson et al. 1994; Petrie et al. 2003; Bennett et al. 2005; Holzapfel et al. 2006). Conversely early harvest of fruit from grape vines has been shown to increase flower numbers by 50% (Holzapfel et al. 2006). However few studies have directly linked carbohydrate storage and subsequent floral production. Reserves in several organs are likely to support floral morphogenesis in grape and kiwifruit buds during spring. During dormancy, carbohydrates in the root system vary from 18 to 75% of total grapevine reserves (Lebon et al. 2008; Holzapfel et al. 2010) and 80% of starch reserves in Hayward vines (Smith et al. 1992). Decreasing root reserves through leaf removal or shading of grapevines during the previous season significantly reduces subsequent flowering, although these treatments may also influence reserves in other organs (Bennett et al. 2005). Specifically reducing carbohydrate concentrations in grape canes (Sommer et al. 2000; Goffinet 25

62 2004) or the dry weight of kiwifruit canes (Grant and Ryugo 1984) during their development increased bud mortality and decreased flower production in spring compared with that in untreated canes. High carbohydrate concentrations in over-wintering buds are thought to protect meristems from low temperatures (Wample and Bary 1992; Jones et al. 1999; Richardson et al. 2007); however, these reserves may also be needed to support very early growth of the meristems. The relative importance of storage organs has not been described in either grape or kiwifruit vines Does carbohydrate supply affect flower quality and fruit set in vines? Remobilisation of vine reserves continues throughout spring until anthesis, although leaves begin to export carbohydrates three or four weeks after budbreak (Piller and Meekings 1997; McArtney and Ferree 1999b; Greer et al. 2003; Poni et al. 2006; Greer et al. 2011). The transition from storage reserves being the main source of carbohydrates to photosynthetic supply as the main source occurs during meiosis and is a critical stage of grape flower development (Lebon et al. 2004b). Interrupting carbohydrate supply to flowers at this stage causes abnormal flower development and results in coulure or abscission of up to 80% of flowers after anthesis (Mullins et al. 1992; Lebon et al. 2004b). This is also a critical stage in the development of kiwifruit flowers, when pollen or ovules develop and starch reserves accumulate in the styles of pistillate flowers to support pollination (Brundell 1975c; Hopping and Jerram 1979). Carbohydrate supply is also essential to ensure rapid development of the ovary prior to anthesis (Lai et al. 1990; Petrie and Clingeleffer 2005). Kiwifruit and grape flowers accumulate significant carbohydrate reserves during this period to support fruit set and rapid growth of ovaries after fruit set (Kliewer 1965; Richardson et al. 2011). Shading, defoliation, girdling, or pruning vines during this period of development can affect fruit set and early fruit growth (Coombe 1962; Brown et al. 1988; Piller et al. 1998; Zapata et al. 2004a) but how these effects occur has not been determined. Grape and kiwifruit vines both bear comparatively low yields of fruit, because of the low number of inflorescences and flowers that they produce. Low 26

63 flower numbers appear to be a function of vines prioritising vegetative growth over reproduction for much of the season. Vigorous vegetative growth competes with developing flowers for carbohydrate supply, particularly during floral morphogenesis, which is delayed until early spring. Flowering processes in both grape and kiwifruit vines appear to be limited by the availability of carbohydrates at three points during floral development: 1. during inflorescence initiation and bud development in the season preceding anthesis, 2. during the early stages of floral morphogenesis around budbreak in spring and 3. during pollen and ovule development, immediately before anthesis and during fruit set. Apparently no previous studies have examined effects of carbohydrate limitation on all three stages of flower development, despite the apparent sensitivity of flower development to carbohydrate supply in both grape and kiwifruit vines. 27

64 1.5 Thesis aims and hypotheses The goal of this study is to determine whether carbohydrate supply limits reproductive development prior to anthesis in a comparable manner and through similar processes in grape and kiwifruit vines. Key questions that form the basis of this research proposal are: 1. Does variation in carbohydrate supply from stored or current assimilates affect floral development of grape and kiwifruit vines in a similar way? 2. Which stages of floral development in grape and kiwifruit are most sensitive to variations in carbohydrate supply? 3. How does carbohydrate supply limit flower development in grape and kiwifruit vines during each of these stages? 4. Are carbohydrate limitations of flowering in grape and kiwifruit typical of the behaviour of vines? These questions form the basis of four hypotheses and related objectives that will direct the proposed research. 1. Hypothesis: Seasonal patterns of vine development and associated carbohydrate dynamics of grape and kiwifruit vines are similar during the key stages of development. The following objectives will be used to assess this hypothesis in Chapter 3: 1. To describe the phenology and seasonal patterns of growth and carbon acquisition of grape and kiwifruit vines. 2. To quantify seasonal changes in carbohydrate concentrations in key organs of grape and kiwifruit vines. 2. Hypothesis: The number of inflorescences produced is determined by carbohydrate supply during inflorescence initiation and early bud development in grape and kiwifruit. The following objective will be used to assess this hypothesis in Chapter 4: 28

65 1. To determine whether the size, carbohydrate concentration and content of developing buds can be changed during inflorescence initiation. 2. To determine whether manipulating carbohydrate supply at different stages of inflorescence initiation can influence the number of inflorescences produced. 3. Hypothesis: That carbohydrate supply from reserve pools affects the early stages of floral morphogenesis and hence inflorescence numbers through the amount of carbohydrates available and the location of those carbohydrates. The following objectives will be used to assess this hypothesis in Chapter 5: 1. To quantify or determine how the amount of reserves stored in the previous season affects floral morphogenesis. 2. To determine which reserve pools are most important during early floral morphogenesis. 4. Hypothesis: That the amount of carbohydrates supplied from leaves to flowers during the later stages of flower development determines flower number and the ability of the flowers to set and develop into fruit. The following objectives will be used to assess this hypothesis in Chapter 6: 1. To determine how critical carbohydrate supply is to formation of grape and kiwifruit flowers immediately prior to anthesis. 2. To measure how changes during late floral development affect fruit set and berry development. 29

66 30

67 2 General materials and methods The methods described in this chapter were used in one or more of the experiments in following chapters. Some of the techniques described in detail here are mentioned briefly in following chapters. Methods specifically used in individual experiments are described in each chapter. 2.1 Plant material Experiments were carried out on Vitis vinifera L. cv. Shiraz, Vitis vinifera L. cv. Cabernet Sauvignon grapevines and Actinidia deliciosa (A. Chev.) C.F. Liang et A.R. Ferguson) cv. Hayward kiwifruit. Hereafter these are referred to as Shiraz, Cabernet Sauvignon and Hayward vines. The study area in the Bay of Islands, Northland, New Zealand has a significant area of commercial kiwifruit vines, and a smaller area of grape vineyards. Readily accessible commercially grown Hayward and Shiraz vines were available for long-term seasonal monitoring (Chapters 3) or short-term spring studies (Chapter 6). However, only Hayward and Cabernet Sauvignon vines were available for multi-season vine manipulation experiments (Chapter 4 and 5). Therefore for consistency between experiments, a more distant site with commercially grown Cabernet Sauvignon vines was included in both the seasonal monitoring (Chapter 3) and short term spring studies (Chapter 6) in the final 2010/2011 season. Although the vines were grown at different sites and under different management this study examined basic physiological properties of vines Commercial vines Briefly, experimental vines were grown in standard commercial orchard/vineyards at three sites: Kerikeri Research Centre ( Hayward ), Marsden Estate ( Shiraz ) and Ivana Wines ( Cabernet Sauvignon ) (Table 2.1; Figures ). All three orchards were located in the Bay of Islands region, New Zealand, where the climate is described as warm temperate. The soils at each site were derived from basalt, but formed during different volcanic events and elevation ranged from 80 to 200 m above sea level. 31

68 Each scion cultivar was grafted onto a rootstock, with Hayward on seedling Bruno rootstocks, Shiraz on clonal m470 rootstocks and Cabernet Sauvignon on clonal SO4 rootstocks. At the beginning of these studies (July 2008) Hayward vines were 24-years-old, Shiraz vines five-years-old and Cabernet Sauvignon vines 20-years-old. Hayward vines were trained on a horizontal pergola trellis (Sale and Lyford 1990), while Shiraz and Cabernet Sauvignon vines were trained vertically on a Smart-Dyson split canopy or a vertical shoot position (VSP) trellis respectively (Gladstone and Dokoozlian 2003) (Figures ). All vines were cane pruned in winter, and then fruiting shoots were trimmed at either four leaves past the fruit on Hayward vines or at 2-3 m length (Shiraz and Cabernet Sauvignon vines) during summer. Approximately 3-4 leaves were plucked from around Shiraz and Cabernet Sauvignon bunches in December. Dioecious Hayward vines were interplanted with M52 male vines and flowers were pollinated by bees (6 hives/ha) while hermaphrodite Shiraz and Cabernet Sauvignon flowers were wind pollinated. In addition, Hayward vines were regularly irrigated and had significant quantities of fertiliser applied. Shiraz and Cabernet Sauvignon vines were not regularly irrigated and had limited fertiliser applications. 32

69 Table 2.1: Details of the location, site characteristics, vine material and cultural techniques used on the three commercial orchards/vineyards used. Details Kerikeri Research Centre Marsden Estate Ivana Wines Vineyard Location S E S E S E Elevation 80 m 130 m 200 m Soil type Okaihau gravelly, friable clay. Kerikeri friable clay. Ruatangata friable clay. Vine scion Actinidia deliciosa cv. Hayward Kramer clone Vitis vinifera cv. Shiraz unknown clone Vitis vinifera cv. Cabernet Sauvignon unknown clone Rootstock Actinidia deliciosa Bruno seedling Vitis riparia Michaux x Vitis rupestris Scheele cv m470 Vitis berlandieri Planchon x Vitis riparia Michaux cv. SO4 Year of planting Within row 2.5 m 2.0 m 1.8 m spacing Between row 4.5 m 3.0 m 2.7 m spacing Trellis Pergola Smart-Dyson split canopy Vertical shoot position Row North - south North - south East - west orientation Winter pruning 16 canes: 1.8 m and 20 buds 2 canes: 0.9 m and 10 buds 4 canes: 0.9 m and 10 buds Summer pruning Fruiting shoots pruned at 4 leaves past the fruit Fruiting shoots topped at 2-3 m in December and January. Fruiting shoots topped at 2-3 m in December and January. Pollination Bees 4 hives/ha Wind Wind Fertiliser 1.5 t/ha Kiwifruit mix Bio Ag foliar None (NPKSMg 0:3.5:18.7:8.1:5.9) 300 kg/ha CAN (N: 27) spray with trace elements Irrigation Drip irrigated Rain fed Rain fed Sward 1m herbicide strip, with grass interrow management Pest and disease See Appendix 1 for Pest and Disease programme, Birdnetting after veraison for Shiraz vines 33

70 Figure 2.1: Hayward vines on Bruno rootstocks were 24-years-old, canepruned and grown on a pergola at the Kerikeri Research Centre. Figure 2.2: Shiraz vines on m470 rootstocks were five-years-old, cane-pruned and grown on a Smart-Dyson split canopy trellis at Marsden Estate. 34

71 Figure 2.3: Cabernet Sauvignon vines on SO4 rootstocks were 20-yearsold, cane pruned and trained on a VSP system at Ivana Wines Vineyard Potted vines Hayward and Cabernet Sauvignon vines were also grown in pots at the Kerikeri Research Centre (Figure 2.4). No Shiraz vines were available at this site. Details of these vines and their management are presented in Table 2.2. Hayward vines were one-year-old plants derived from cuttings at the beginning of this study (July 2008), while Cabernet Sauvignon vines grafted on clonal SO4 rootstocks were nine-years-old. Hayward vines were grown on a T-Bar trellis, while Cabernet Sauvignon vines were grown on VSP trellis. Vines were cane pruned in winter and fruiting shoots pruned in summer as described above for commercially grown vines. All vines were grown in 30 litre pots containing Daltons Premium potting mix (Daltons, Matamata, New Zealand). Regular dressings of Sierra Blend slow release fertiliser (Scotts-Sierra Horticultural Products, Marysville, Ohio, USA) were applied to vines. Water was applied to vines via a drip irrigation system throughout the growing season, with up to three applications per day during peak summer demand. 35

72 Figure 2.4: Hayward and Cabernet Sauvignon vines were grown in pots at the Kerikeri Research Centre for use in multi-season studies. 36

73 Table 2.2: Details of the location, site characteristics, vine material and cultural techniques used on the Hayward and Cabernet Sauvignon vines grown in pots for multi-season studies. Details Kerikeri Research Centre Location S E Elevation Vines scion 80 m Actinidia deliciosa cv. Hayward Rootstock own roots SO4 Year of planting Pot size Within row spacing Between row spacing Vitis vinifera cv. Cabernet Sauvignon 30 L with Daltons Premium potting mix 1.0 m 2.5 m Trellis T-Bar Vertical shoot position Row orientation Winter pruning Summer pruning Fertiliser Irrigation Pest and disease North - south Cane pruned 4-6 canes, 18 buds per cane Fruiting shoots pruned at 4 leaves past the fruit Cane pruned 2-3 canes, 13 buds per cane Fruiting shoots topped at trellis height 2-3 m in December / January. 80 g Sierra Blend (NPK 21:1.8:9 + trace elements) per year Microirrigation 2x4 L/hour drippers per vine. Water applied according to daily evapotranspiration see Appendix 1 for Pest and Disease programme 37

74 2.2 Meteorological data All experimental vines were growing within 2 km and 10m altitude of a standard meteorological station. Data for the Kerikeri Research Centre ( Hayward commercial and all potted vines) were collected from the meteorological station at the Centre, data for the Marsden Estate ( Shiraz commercial) came from the adjacent Kerikeri Airport meteorological station and data for the Ivana Wines Vineyard ( Cabernet Sauvignon commercial) came from the nearby Kaikohe DSIR EDR meteorological station. Daily maximum and minimum temperatures and rainfall data were from the two Kerikeri sites from 1 May 2008 to 31 December 2011 and from the Kaikohe site from 1 May 2010 to 31 December In addition daily solar radiation data was available from the Kerikeri Research Centre and the Kaikohe DSIR EDR sites, but was not available from the Kerikeri Airport site. Mean monthly values were calculated for temperature, rainfall and solar radiation data from each site and data are presented in Appendix Vine phenology Budbreak and flowering The time course of budbreak was determined at regular intervals of between two and four days from the start of bud movement until no new buds had developed. A bud was said to have broken when the developing shoot began to elongate and showed a small amount of green tissue. This is defined as green tip (E-L stage 4) for Shiraz and Cabernet Sauvignon grapevines (Eichorn and Lorenz 1977; Coombe 1995) and as advanced bud burst for Hayward kiwifruit (Brundell 1975b). The time course of flowering (anthesis) was determined in a similar way except observations were made at two - three day intervals. Anthesis for grape flowers was defined as having occurred when flowers that had lost their caps (calyptras), and the percentage of flowers in this state was visually estimated for each inflorescence (E-L stage 19-26) (Eichorn and Lorenz 1977; Coombe 1995). To determine total flower number per inflorescence, inflorescences were covered with a net bag (0.5 mm mesh) during anthesis to collect flower caps (calyptras) which were subsequently counted to determine total flower 38

75 number per inflorescence. A kiwifruit flower was determined to be at anthesis when the petals had opened sufficiently to allow easy access to bees (McPherson et al. 2001). Total budbreak was calculated from the number of buds that had broken as a percentage of the total number of buds per cane. Floral budbreak was calculated from the number of buds that produced shoots with inflorescences as a percentage of the total number of buds per cane. Flowering was calculated from the number of flowers that had reached anthesis as a percentage of the total number of flowers. The number of inflorescences produced per dormant winter buds (IWB) was determined from the following measurements according to the method of (Snelgar et al. 1997): IWB = BB x FS x INF Where: BB = the proportion of all dormant buds that produce shoots in spring. FS = the proportion of shoots that produced inflorescences. INF = the number of inflorescences produced on each floral shoot. For Hayward an inflorescence is defined as a terminal flower and up to two lateral flowers (Hopping 1990) Phenological stages of development Phenological stages of vine development were described using the modified E-L system for Shiraz and Cabernet Sauvignon grapevines (Lorenz et al. 1994; Coombe 1995) and an E-L type system for Hayward kiwifruit (Table 2.3). Data on budbreak, early shoot development, flowering, and fruit development and leaf fall were combined to describe timing of vine development for each vine cultivar (Jackson 1994). Shoot development was determined from shoot node number, leaf opening, shoot length, inflorescence and pedicel growth. Fruit development was assessed from berry growth, seed development and the total soluble solids concentration (Brix) in the juice of fruit. Fruit maturation began at veraison in grapes (the beginning of berry colouration and softening) and seed maturation in 39

76 Hayward (the beginning of seed colouring). Fruit were harvested from vines when grape berries had reached a Brix of about 20 and Hayward had reached at least 6.2 Brix. Leaf fall was assessed from the number of leaves that had abscised from canes. 40

77 Table 2.3: Key phenological stages described using the modified E-L (Eichorn-Lorenz) and BBCH (BASF, Bayer, Ciba-Geigy and Hoechst) growth stage models for Shiraz and Cabernet Sauvignon grapevines reproduced from Lorenz et al. (1994); Coombe (1995); May (2004). Key BBCH stages for Hayward kiwifruit (Salinero et al. 2009) and a newly developed E-L type system for Hayward. Cultivar Shiraz and Hayward Phenological stage E-L grape Cabernet Sauvignon BBCH grape E-L type kiwifruit BBCH kiwifruit Dormant Beginning of budswell Beginning of budbreak First leaf separated Five leaves, shoots 10 cm inflorescence visible Inflorescence well developed pedicels elongated Flowering begins % flowering Full bloom Fruit set Berries pea size (grape)/ green rapid growth (kiwifruit) Veraison or seeds maturing Harvest Beginning of leaf fall End of leaf fall

78 2.4 Leaf area estimation Leaf dimensions were converted into leaf area measurements using the coefficients established by obtaining the line of best fit between measured leaf area (LI3100 Li-Cor, Lincoln, Nebraska, USA) and the product of maximum leaf width and length along the main vein for 200 leaves of varying size from each vine cultivar (see Appendix 3). The coefficient of slope for leaf width (cm) * leaf length (cm) vs. measured leaf area (cm 2 ) ± SEM for each vine are: Hayward ± Shiraz ± Cabernet Sauvignon ± Gas exchange The rate of photosynthesis was measured using the Li-Cor 6400 open path photosynthesis system (Li-Cor Inc., Lincoln, Nebraska, USA) fitted with a carbon dioxide (CO 2 ) mixer. The CO 2 concentration was set at 400 µmol mol -1 at the leaf surface. The stomatal ratio was set at 0 as both grape (Mullins et al. 1992) and kiwifruit (Ferguson 1990) leaves have stomata only on the underside (hypostomatous). Photosynthetic light response curves were carried out on Hayward kiwifruit, Shiraz and Cabernet Sauvignon grapevines at anthesis (December) during the 2010/2011 growing season. Irradiance was provided by the LED light source ( B, Li-Cor Inc., Lincoln, Nebraska, USA). Photosynthetic responses were measured on the youngest mature leaf on a shoot from four separate vines. This was repeated for each vine cultivar. Initially leaves were allowed to equilibrate at a photon flux density (PFD) of 1200 µmol m -2 s -1, then the PFD was increased to 2000 µmol m -2 s -1 and photosynthetic rates stabilised before measurements were recorded. Photon flux densities were then reduced in 14 steps, and photosynthetic rates measured until dark when respiration was recorded. For each leaf a non-linear curve was fitted using a hyperbolic tangent function of the form: 42

79 A = [P max x Tanh (PFD x φ app )/P max ] Rs to photosynthetic light response data according to Greer and Halligan (2001). A is the measured rate of photosynthesis, P max is the light saturated maximum rate of photosynthesis, Rs is the rate of respiration, φ app is the apparent photon yield, PFD is the photon flux density and Tanh is the hyperbolic tan function. All gas exchange measurements on experimental vines were made between 10 am and 3 pm, when photosynthesis rates are generally consistent (Petrie et al. 2009; Greer et al. 2011) under natural light, temperature and vapour pressure deficit conditions. Leaves greater than 35 mm in width were enclosed in the chamber in situ at their normal orientation, readings were allowed to stabilise and then three consecutive measurements were made for each leaf. On each measured shoot, photosynthetic rates were determined for every second leaf along the shoot, until leaves were too small to fit in the chamber. 2.6 Carbohydrate analysis Tissue extraction All tissues collected for carbohydrate analysis were removed from vines (as described for each experiment) and immediately frozen in liquid nitrogen and then stored at -80 o C. Vegetative samples were freeze dried and ground, while fruit samples were ground fresh in liquid nitrogen. A subsample of tissue from each sample was extracted for 1h at 60 o C using 80% ethanol with adonitol added as an internal standard. Extracted samples were centrifuged and the supernatant decanted off. The residue was suspended in 80% ethanol, centrifuged again and both supernatants combined Starch determination For all tissues except buds, the insoluble residue was analysed for starch using the method of Smith et al. (1992). The residue was autoclaved for 1 hour and then incubated at 55 o C with amyloglucosidase in acetate buffer 43

80 (0.25 M, ph 4.5) for 1 hour. The glucose concentration in solution was then measured colourimetrically at 510 nm (UV-Vis Spectrophotometer, Shimadzu, UV-160A) using a manual glucose oxidase method (Trinder 1969). For buds, which were too small to analyse using the method used for other tissues, starch was extracted from the insoluble residue using a method modified from MacRae et al. (1992). The pellet was extracted in 0.5 ml dimethylsulfoxide and ml 8N HCl at 60 o C for 75 minutes. A subsample was taken and diluted with 0.3 ml of 0.5 M citrate buffer (ph 5.25). Starch concentrations were measured by incubating 45 µl of the dilute starch extracts 4 µl of 10 ml of 100 mg/ml1.4 U/mg amyloglucosidase. The mixture was incubated for 1 hour at 60 o C and then glucose measured using an assay mix containing 0.3 M triethanolamine buffer ph 7.6, 4 mm MgSO 4, 3 mm ATP neutralised with 22 mm sodium bicarbonate, 0.5 mm NADP, 0.68U hexokinase and 0.3U glucose 6P dehydrogenase. A starch standard was run with all analyses and data corrected for starch recovery. Glucose concentrations were subsequently converted to starch equivalents by multiplying glucose values by 0.9 and expressed as mg starch per g of dry or fresh tissue as appropriate Soluble carbohydrate determination A sub-sample of the supernatant from the original tissue extraction was taken, the ethanol evaporated using a stream of nitrogen gas and the sample redissolved in ultra pure water. The soluble sugars were analysed using DIONEX ICS-3000 Reagent-Free IC (RFIC ) system (DIONEX, USA) with a CarboPac PA20 column. Chromatograms were analysed using Chromeleon software. 2.7 Data handling and statistical analysis Data handling Datasets were checked for recording and transcription errors by sorting data by treatments and scanning data to identify any unusual values. Data was also sorted to identify any unusually large or small data. Residual plots of all data were also examined for any outliers. In each case, any unusual data identified was checked. Where data required transformation to conform to 44

81 the assumptions underlying analysis of variance (ANOVA), the method outlined in Fernandez (1992) was used and data was back-transformed for presentation. Most experiments used a fully randomised split plot design for each vine cultivar with individual or adjacent groups of vines as blocks to remove random variation due to position within a block or previous history of the vine. Canes or shoots were generally used as experimental units and data were averaged per experimental unit. Although the field grown vines were grown at different locations, the amount of replication for each cultivar was more than adequate to statistically compare these vines within each season. No further accounts of site difference were considered as these could not be tested statistically Data analysis Statistical analysis was carried out using GenStat 14 th edition for Windows (VSN International Ltd., Hemel Hempstead, Hertfordshire, United Kingdom). Most experimental analyses were carried out using Restricted Maximum Likelihood (REML) linear mixed models to determine the effects of treatments on measured variables. Treatment comparisons were made by using appropriate least significant differences for the predicted means at the appropriate level of comparison. Relationships between factors affecting carbon (C) assimilation or bud performance were examined using the generalised linear model (GLM) procedure Data plots and curve fitting All data was plotted and lines fitted using Origin version 8.5 (Originlab Corporation, Northhampton, Massachusetts, USA). Curves were fitted to photosynthetic light response data with non-linear fits using a hyperbolic tangent function as described above. Modelling of budbreak, flowering and shoot growth data as a function of time was carried out using a Boltzmann sigmoid function (Greer 2001; Seleznyova and Greer 2001): 45

82 Where x 0 = point of inflexion dx = duration of the window of rapid expansion A 1 = lower asymptote A 2 = upper asymptote And A 2 >> A 1 46

83 3 Seasonal patterns of kiwifruit and grapevine development 3.1 Introduction Understanding and quantifying seasonal growth patterns are an essential part of studying plants. Plants follow characteristic patterns, for example, temperate, woody grape and kiwifruit vines are dormant during winter, buds formed during the previous summer begin to swell and grow in spring (Brundell 1975b; Srinivasan and Mullins 1981; Mullins et al. 1992; Boss et al. 2003), new leaves are produced and expand, inflorescences and flowers develop and the main shoot axis extends (Williams et al. 1985b; Coombe 1988; Mullins et al. 1992; Snowball 1995). Shoots continue to grow rapidly over an extended period until anthesis, days after budbreak (DAB). New leaves, initiated distal to the current season s inflorescences, subtend axillary buds that initiate leaf and inflorescence primordia for the following season (Pratt 1971; Srinivasan and Mullins 1981; Walton et al. 1997). Once inflorescence primordia are initiated there is no further morphogenesis in Hayward inflorescences (Walton et al. 2001) and only limited branching in grape inflorescences in that growing season (Pratt 1971; Srinivasan and Mullins 1981; Watt et al. 2008). In the current season, shoot growth generally slows after fruit set as more vine resources are allocated for fruit development (Williams et al. 1985b; Buwalda and Smith 1990; Mullins et al. 1992). The growth of grape and kiwifruit berries both follow a double sigmoid pattern over about 100 days for non-climacteric grape berries (Mullins et al. 1992) and 160 days for climacteric Hayward berries (Hopping 1976). In addition, vines lay down reserves in perennial organs during summer and autumn for the following season (Smith et al. 1992; Holzapfel et al. 2010). In grape, flushes of root growth have been observed both when vine growth resumes in spring and after fruit harvest (Freeman and Smart 1976; Williams and Matthews 1990; Bates et al. 2002); however, Eissenstat et al. (2006) found that root growth generally occurs between bloom and veraison. There are few studies of kiwifruit vine roots but several growth patterns have been observed including a single period of growth in late summer (Buwalda and Hutton 1988), a flush 47

84 in summer followed by a second peak of growth in autumn (Black 2012) or no distinct seasonal trend (Reid et al. 1993). Quantifying the amount of total non-structural carbohydrates (TNC) in plant organs can provide an understanding of the balance between the supply of photosynthate and the demand for carbohydrate, including the use of reserves, for growth. Previous studies of carbohydrate dynamics in grape and kiwifruit have often focussed on single stages of vine growth, particularly fruit development (Kliewer 1965; Scholefield et al. 1978; Hamman et al. 1996; Richardson et al. 1997; Boldingh et al. 2000; Bates et al. 2002; Weyand and Schultz 2006), on specific organs such as roots or buds (Wample and Bary 1992; Jones et al. 1999; Bennett et al. 2005; Richardson et al. 2007) or a limited number of carbohydrates (Winkler and Williams 1945; Smith et al. 1992; Bates et al. 2002). In grapevines sucrose, glucose and fructose are the main soluble sugars present in tissues (Kliewer 1966) and significant amounts of sucrose, glucose, fructose, myo-inositol and planteose are found in kiwifruit tissues (Boldingh et al. 2000; Klages et al. 2004). The fruit of both species contain high levels of glucose and fructose, with lower concentrations of sucrose (Amerine and Root 1960; Kliewer 1965; Okuse and Ryugo 1981; Walton and De Jong 1990; Boldingh et al. 2000). However, kiwifruit berries also store large quantities of starch during their development (Richardson et al. 1997; Boldingh et al. 2000). In both vines, starch is a key component of reserve storage in perennial woody tissues (Smith et al. 1992; Holzapfel et al. 2010). In this study, to examine the hypothesis that seasonal patterns of grape and kiwifruit vine development are influenced in a similar manner by carbohydrate supply; vine phenology, growth, carbon acquisition, carbohydrate pools and carbon budgets of vines were compared during key stages. Although the vines were grown at different sites and under different management this study examined basic physiological properties of vines. Vine phenology and shoot growth were described in detail over three seasons for Hayward and Shiraz vines and for one season for Cabernet Sauvignon vines (for consistency with other experiments). Seasonal changes in TNC 48

85 concentrations in key organs of vines were also compared over three seasons. In the final season, carbon acquisition (photosynthesis), shoot biomass and carbohydrate content were then combined to develop carbon budgets for shoots and whole vines. These data were integrated to provide an overview of key events during vine development. 49

86 3.2 Materials and methods Plant material Research was carried out on commercial blocks of 24-year-old Hayward vines on Bruno rootstocks growing at Kerikeri Research Centre, Kerikeri, New Zealand, five-year-old Shiraz grapevines on m470 rootstocks growing at Marsden Estate, Kerikeri, New Zealand; and 22-year-old Cabernet Sauvignon grapevines on SO4 rootstocks growing at Ivana Wines, Kaikohe, New Zealand. Measurements were made on Hayward and Shiraz vines over three seasons (2008/2009, 2009/2010 and 2010/2011) and on Cabernet Sauvignon vines for one season (2010/2011) so that the seasonal growth patterns of all vines used in later experiments were documented. Full details of vines, growing conditions and cultural management are given in Chapter 2 (Table 2.1). Briefly, Hayward vines were trained on a pergola system, Shiraz vines on a Smart-Dyson split canopy system and Cabernet Sauvignon vines on a Vertical shoot position system. All vines were cane pruned in winter to leave 16 one-year-old canes on each Hayward vine, two one-year-old canes on Shiraz vines, and four one-year-old canes on Cabernet Sauvignon vines. Vines were summer pruned to limit extension of fruiting shoots, remove lateral growth and maintain replacement canes for the following season. Three leaves were plucked from shoots on Shiraz and Cabernet Sauvignon vines, in early summer, to keep bunches open. Both Shiraz and Cabernet Sauvignon vines were grown without irrigation, while Hayward vines were drip irrigated (see Chapter 2) Vine phenology The phenological development of the vines was monitored at regular intervals over three seasons (see Chapter 2). The times of budbreak and flowering were determined from regular visual assessments of vines during the 2008/2009 season. In the 2009/2010 and 2010/2011 seasons, 16 typical one-year-old canes (two canes on each of eight vines) were tagged at each site during winter dormancy. The dimensions and bud number were recorded for each cane. Time of budbreak and flowering were assessed at 50

87 two - four day intervals according to the protocol described in Chapter 2. Thereafter, all observations of vine development were expressed as days after the mid-point (50%) of budbreak (DAB). On each monitored cane total budbreak (BB) was calculated from the number of buds that had broken as a percentage of the total number of buds per cane. Floral budbreak (FS) was calculated from the number of buds that produced shoots with inflorescences as a percentage of the total number of buds per cane. The average number of inflorescences produced on each floral shoot (INF) was calculated from the total number of flowers per cane divided by the total number of floral shoots. The number of inflorescences produced per dormant winter buds (IWB) was determined from the following measurements IWB = BB x FS x INF according to the method of Snelgar et al. (1997), (see Chapter 2). Time of shoot growth was determined from regular visual observations of shoots during the 2008/2009 and 2009/2010 seasons. In the 2010/2011 season, one typical shoot on each tagged cane was selected for shoot growth measurements. The length, basal diameter of the shoot, and width and length of all leaves on each shoot were measured at 7-10-day intervals during active leaf and shoot growth and then at approximately 20-day intervals thereafter, until the end of the season. Leaf dimensions were converted into leaf area measurements using the calibrations established by obtaining the line of best fit between measured leaf area and the product of leaf width and length for data from 200 leaves of varying size from each vine cultivar (see Appendix 3). Once the outer photosynthetic tissues in a shoot degenerated and the shoot changed colour from green to brown it was defined as a cane (Jackson 1994). On each date that tissue samples were collected for carbohydrate analysis over three seasons, six shoots equivalent to those sampled for monthly carbohydrate analysis, were collected from each site. The length, basal diameter and leaf number were measured on each shoot. All leaves were then removed from each shoot and the leaf area determined using a LI

88 leaf area meter (Li-Cor Inc, Lincoln, Nebraska, USA). The weight of the leaves, fruit and stem were recorded before all material was dried at 65 o C for 48 hours or until tissue reached a steady weight and then reweighed to determine tissue dry weights. Root activity was determined at each date by visually assessing whether the samples excavated for carbohydrate analysis had actively growing white tips, no growth or dead roots present. Fruit maturation began at veraison in grapes (the beginning of berry colouration and softening) and seed maturation in Hayward. Fruit were harvested from vines when grape berries had reached a Brix of about 20 and Hayward had reached 6.2 Brix. Leaf fall was assessed by the percentage of leaves that had abscised from canes Gas exchange measurements Photosynthetic light response curves were measured on Hayward, Shiraz and Cabernet Sauvignon leaves at anthesis (December) during the 2010/2011 growing season using the Li-Cor 6400 open path photosynthesis system (Li-Cor Inc, Lincoln, Nebraska, USA) fitted with a carbon dioxide (CO 2 ) mixer as described in Chapter 2. Photosynthetic responses were measured on the youngest mature leaf on a shoot from four separate vines. For each leaf a non-linear curve was fitted to photosynthetic light response data using a hyperbolic tangent function (Greer and Halligan 2001) (see Chapter 2). In the 2010/2011 season, the rates of photosynthesis were measured on eight of the measured shoots during shoot development. All measurements were made between 10 am and 3 pm under natural light, temperature and vapour pressure deficit conditions. Leaves greater than 35 mm in width were enclosed in the Li-Cor 6400 chamber at their in situ orientation, readings were allowed to stabilise and then three consecutive measurements were made. On a shoot, every second leaf was measured, until leaves were too small to measure. Measurements were made at fortnightly intervals during rapid shoot development and at monthly intervals thereafter. 52

89 All gas exchange measurements collected over the growing season were collated and photosynthesis (A) was modelled as a function of stomatal conductance (g s ), internal CO 2 concentration (c i ) and PFD for each cultivar. A general linear model (GLM) was used to model photosynthesis (Greer 2012) using the equation: A = f(g s, c i, PFD) Net carbon balance The total carbon acquired per shoot per day was calculated during shoot growth (Greer et al. 2011). Photosynthetic rates were assumed to be constant throughout the day (Cartechini and Palliotti 1995) and photosynthetic measurements for leaves were integrated using actual day length values. Photosynthetic rates were combined with the calculated area of each leaf and then summed over all leaves to calculate daily CO 2 fixation rates per shoot. Rates of dark respiration were calculated from light response data for each vine cultivar. Respiration rates were integrated over the night period and combined with daily leaf area per shoot to calculate a daily respiration per shoot in a similar manner to CO 2 fixation. Photosynthetic acquisition and respiration of shoots were converted to g of carbon using the molecular fraction of carbon in CO 2 (Greer et al. 2003). Net carbon acquisition was calculated as the mathematical sum of the daily carbon acquisition and respiration per shoot. Carbon sequestered in the shoot as biomass was determined from shoot measurements and dry weight data. Shoot dry weight was converted to g of carbon by assuming that 45% of the biomass was elemental carbon (Walton and Fowke 1995). The net daily carbon balance for each shoot was determined as the difference between daily carbon acquisition and the daily carbon accumulated as biomass in a shoot Sampling and carbohydrate analysis Samples for carbohydrate analysis were collected from vines at approximately monthly intervals over the three seasons. Collections began 53

90 in July 2008 for Hayward vines and from November 2008 for Shiraz vines and continued through to June 2011 when vines were dormant. Samples were also collected from Cabernet Sauvignon vines between June 2010 and June Sampling was rotated in blocks so that no vine was sampled more than once a year. Samples were collected from vines during late morning to mid-afternoon (10 am - 3 pm). Four replicates each of developing axillary buds, leaves, internodes, one-year-old canes and fine roots were sampled. Material from three separate vines was bulked per replicate. From each shoot, samples of leaves, axillary buds and internodes were taken from each of three nodes located two nodes above the most distal inflorescence on a shoot. In addition, the most distal 50 mm of a one-year-old cane subtending current season s shoots was sampled. A sample of fine roots (< 2 mm) was collected from an excavation 150 mm diameter by 150 mm deep within the herbicide strip of the row, approximately m from the trunk of a vine. In the final season (2010/2011), one inflorescence or bunch ( Shiraz and Cabernet Sauvignon ) or fruit ( Hayward ) were also collected from each sampled shoot. Berry subsamples were taken by randomly selecting nine berries from Cabernet Sauvignon and Shiraz bunches or taking a 2 mmthick equatorial slices from Hayward fruit. All bud, leaf, internode, inflorescence, fruit and one-year-old cane tissue was immediately frozen in liquid nitrogen and stored at -80 o C. Root material was kept on ice for a short period, washed and then frozen in liquid nitrogen and stored at -80 o C. Samples were freeze-dried, extracted and analysed for starch colourimetrically and soluble carbohydrates by ion chromatography as described in Chapter 2. Total sugar concentrations were calculated from the sum of individual sugars and TNC concentrations were calculated by summing the starch and total sugar concentrations. TNC content of shoot tissues was calculated by multiplying the TNC concentration of an organ by its total dry weight and then summing values per shoot at each sampling date. As the total vine biomass of vines could not be measured on commercial vineyards/orchards and as shoot components from this study were similar to those of Smith et al. (1992) and Field (2013) it was assumed that the total 54

91 vine biomass of the components of vines was similar to that measured in previous studies of Shiraz and Hayward. The carbon content in total vine dry mass was calculated by multiplying organ dry weights by 0.45 (Walton and Fowke 1995). Vine TNC budgets were calculated using the TNC concentrations of roots, one year old cane, leaves, internode and fruit measured at each site prior to harvest (23 March 2011 Shiraz and Cabernet Sauvignon and 27 April 2011 Hayward ) and the appropriate biomass of each organ from earlier studies of grape Field (2013) and kiwifruit Smith et al. (1992). The carbon content of the carbohydrate (non-structural) pool was calculated by multiplying the TNC content of vine organs by 0.42 (average carbon content of the carbohydrates in tissues). The total carbon content of the structural component of tissues was calculated by subtracting the nonstructural carbon content (carbohydrate) from the total carbon content of tissues Statistical analysis All data were analysed using GenStat 14 th edition for Windows (VSN International Ltd., Hemel Hempstead, Hertfordshire, United Kingdom) using REML linear mixed models to analyse effects of vine cultivar and season on vine phenology, shoot growth, gas exchange, shoot net carbon balance, and the carbohydrate composition of shoot components. Differences between treatment means were determined at the appropriate level of significance, and estimated means and standard errors were presented. Relationships between carbon assimilation, light, gas exchange parameters and leaf characteristics were examined using the GLM procedure in Genstat. All data was plotted using Origin version 8.5 (Originlab Corporation, Northhampton, Massachusetts, USA). Curves were fitted to budbreak, flowering and shoot growth data across time using the Boltzmann function (Seleznyova and Greer 2001) and a non linear curve was fitted to light response data using a hyperbolic tangent function (Greer and Halligan 2001). See Chapter 2 for further details. 55

92 3.3 Results Vine phenology Vine growth patterns The pattern of vine development was very similar for all vines over the three seasons (Figure 3.1). Budbreak and flowering were clearly separated in spring, with the greatest separation for Shiraz vines (75 days), followed by Cabernet Sauvignon (64 days) with the least separation for Hayward vines (53 days). The growth of shoots occurred over a very similar period, beginning at budbreak and generally finishing in mid-summer ( DAB), except in the 2010/2011 season when shoot growth continued until DAB. The period of fruit growth was consistent between seasons, but was days longer on Hayward than Shiraz and Cabernet Sauvignon vines. The greatest difference between vines was the longer periods of new root growth in Hayward compared with those in Shiraz and Cabernet Sauvignon vines. Hayward roots grew at the end of winter prior to budbreak ( days before budbreak) and from mid-summer until fruit harvest ( DAB). By contrast, Shiraz and Cabernet Sauvignon vines produced new roots for a short period in mid-summer ( DAB) and then again in late autumn/early winter, after fruit harvest ( DAB). A summary of the dates of key developmental stages of vines over three seasons is shown in Table 3.1. These data, particularly when 50% of the buds that developed in spring had broken (DAB), are used throughout the chapter as the timeline for vine development. Shiraz vines were the first to begin developing in spring, closely followed by Hayward and then Cabernet Sauvignon. Budbreak was delayed in the 2010/2011 season compared with the two previous seasons. There was little difference in the mid-point (50%) of flowering between the three seasons (Table 3.1). However, Hayward vines reached mid-flowering approximately 17 days earlier than Shiraz or Cabernet Sauvignon vines. The beginning of fruit maturation varied between vines, with veraison and seed maturation beginning around DAB for Shiraz and Cabernet Sauvignon berries and slightly later in Hayward fruit around DAB. Shiraz and Cabernet Sauvignon 56

93 berries reached commercial fruit maturity around 200 DAB and Hayward fruit, again slightly later, at 230 DAB. Leaf fall was also accelerated by approximately 15 days on Shiraz and Cabernet Sauvignon vines compared with Hayward vines. 57

94 58 Table 3.1: The date and DAB for key growth stages of Hayward and Shiraz vines (2008/ /2011) and Cabernet Sauvignon vines (2010/2011). Vine and 50% budbreak 50% flowering Veraison/seed Harvest date 80% leaf fall season date (DAB) Hayward 2008/09 28 September (0) 2009/10 30 September (0) 2010/11 5 October (0) Shiraz 2008/09 21 September (0) 2009/10 18 September (0) 2010/11 20 September (0) Cabernet Sauvignon 2010/11 2 October (0) date (DAB) 23 November (55) 25 November (56) 27 November (53) 3 December (73) 2 December (75) 27 November (68) 5 December (64) maturation (DAB) 4 March (157) 3 March (154) 28 February (146) 5 February (137) 5 February (140) 29 January (128) 5 February (126) (DAB) 27 May (241) 25 May (237) 17 May (224) 21 April (212) 31 March (194) 2 April (194) 5 April (185) date (DAB) 26 June (271) 30 June (273) 28 June (266) 17 June (269) 28 May (252) 2 June (255) 2 June (243) 58

95 59 Figure 3.1: Timing of key growth stages for Hayward (A, B and C), Shiraz (D, E and F) and Cabernet Sauvignon (G) vines. Data are for the 2008/2009 season (A and D), 2009/2010 season (B and E) and the 2010/2011 season (C, F and G). 59

96 Budbreak and flowering The processes of budbreak and flowering followed a sigmoidal pattern of increase during spring (Figures 3.2 and 3.3). In both the 2009/2010 and 2010/2011 seasons, Shiraz buds were the first to break and reached 50% budbreak 17.6 ± 0.2 and 19.7 ± 2.1 days after 1 September (DAS) respectively, significantly (P < 0.001) earlier than Hayward vines where 50% budbreak occurred 29.9 ± 0.5 and 36.4 ± 1.9 DAS in the 2009/2010 and 2010/2011 seasons respectively. In 2010/2011, Cabernet Sauvignon vines began to break bud significantly (P < 0.001) later than Shiraz buds but earlier than Hayward buds, reaching 50% budbreak 33.2 ± 0.6 DAS. The duration of budbreak (days between 20 and 80%) did not differ significantly between the vines (Shiraz 7.0 ± 1.9 days; Hayward 10.0 ± 2.0 days; Cabernet Sauvignon 13.7 ± 2.8 days) or season (7.4 ± 2.7 days in 2009/2010 cf ± 1.6 days in 2010/2011). There was also a significant interaction between vine cultivar and season (P < 0.001) on the mid-point of flowering. During the 2009/2010 season, Hayward vines reached the mid-point of flowering at 72.9 ± 0.4 DAS, compared with 83.5 ± 0.5 DAS for Shiraz vines. In the 2010/2011 season, Hayward vines reached the mid-point of flowering at 77.2 ± 0.3 DAS and Shiraz at 80.3 ± 0.5 DAS but the mid-point of flowering on Cabernet Sauvignon vines was slightly later at 84.3 ± 0.5 DAS. In 2009/2010, the duration of flowering was significantly (P < 0.001) shorter on Hayward canes (1.6 ± 0.5 days) than Shiraz canes (6.1 ± 0.5 days) but differences between the vines in 2010/2011 were not significant. For Shiraz vines, the duration of flowering was significantly (P < 0.001) shorter in 2010/2011 (3.6 ± 0.5 days) than in 2009/2010 (6.1 ± 0.5 days). By contrast, the duration of flowering on Hayward vines was significantly (P < 0.01) longer in 2010/2011 (3.7 ± 0.5 days) than in 2009/2010 (1.6 ± 0.5 days). 60

97 Figure 3.2: Dynamics of budbreak in A, 2009/2010 and B, 2010/2011 for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ; 2010/2011 only) vines. Data are the means ± SEM, n = 16. The lines are the fit of data to the Boltzmann sigmoid function. Figure 3.3: Dynamics of flowering in A, 2009/2010 and B, 2010/2011 seasons for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( 2010/2011 only) vines. Data points are the means ± SEM, n = 16. The lines are the fit of data to the Boltzmann sigmoid function. 61

98 Root growth The three vines had markedly different patterns of both root growth and root loss (die back from root tips) (Figure 3.4). Hayward vines produced new roots at the end of winter prior to budbreak ( days before budbreak) and again from early summer until fruit harvest ( DAB). In contrast, Shiraz and Cabernet Sauvignon vines produced new roots for a short period ( DAB) in mid-summer and again in late autumn/early winter ( DAB), after fruit harvest. Patterns of root growth were relatively consistent between seasons; however, root loss was more seasonally variable. All vines had a period of root loss in spring around DAB, and this coincided with the period of rapid shoot growth. In Hayward vines, there was a second period of root loss in late winter (2008/2009 and 2010/ days before budbreak) or in autumn (2009/ DAB). Shiraz vines also showed root loss in winter (2010/ DAB) and autumn (2008/2009 and 2010/ DAB). In the 2010/2011 season, root loss in Cabernet Sauvignon vines occurred at the same times as that in Shiraz vines. 62

99 63 Figure 3.4: Periods of root loss (die back of root tips observed at sampling) and root growth (white root tips present at sampling). Data are for Hayward vines (A, 2008/2009, B, 2009/2010 and C, 2010/2011) Shiraz (D, 2008/2009, E, 2009/2010 and F, 2010/2011) and Cabernet Sauvignon (G, 2010/2011). 63

100 Amount of budbreak and inflorescence production There was a significant (P < 0.001) vine cultivar effect on the total number of dormant buds that broke and developed into shoots (Table 3.2). At the end of spring, budbreak on both Shiraz (2009/2010 and 2010/2011) and Cabernet Sauvignon (2010/2011) canes was more than two-fold higher than that on Hayward canes. There was also a significant (P < 0.05) effect of season on the amount of budbreak, with approximately 20% more buds breaking in the 2010/2011 season compared to the 2009/2010 season. Floral budbreak was also 2-fold higher (P < 0.001) on Shiraz and Cabernet Sauvignon canes than on Hayward canes and floral budbreak was also 25% higher (P < 0.05) in the 2010/2011 season than the 2009/2010 season. Hayward buds produced 1.7-fold more inflorescences per shoot (P < 0.001) than either Shiraz or Cabernet Sauvignon buds. The combined effects of vine cultivar on budbreak and inflorescence number per floral shoot resulted in differences in inflorescence numbers per bud between vines not being significant. However, there was a significant (P < 0.05) seasonal effect on the number of inflorescences produced per winter bud. Both Hayward and Shiraz vines produced 50% more inflorescences per bud in the 2010/2011 season than the 2009/10 season. Table 3.2: Budbreak and floral production of vines. Data are mean values ± SEM, n = 16 for Hayward and Shiraz vines (2009/2010 and 2010/2011 seasons) and for Cabernet Sauvignon vines (2010/2011 season). Vine and Budbreak Floral Inflorescence Inflorescence season Hayward (%) budbreak (%) /floral shoot /winter bud 2009/ ± ± ± ± / ± ± ± ± 0.18 Shiraz 2009/ ± ± ± ± / ± ± ± ± 0.08 Cabernet Sauvignon 2010/ ± ± ± ±

101 Patterns of shoot growth The pattern of shoot growth was similar across all vines in 2010/2011 (Figure 3.5). In spring, shoot length increased rapidly (10-17 mm day -1 ), as did both shoot diameter ( ) and leaf area (16-41 cm 2 day -1 ) from around 25 DAB until just after fruit set (60-70 DAB). However, the growth of Shiraz and Cabernet Sauvignon shoots was limited to increases in shoot diameter after this, largely due to pruning practices such as topping and lateral removal. In contrast, unpruned Hayward shoots continued to grow, with increases in both shoot length and leaf area occurring until well into the period of fruit development (120 DAB). Total leaf area on each Hayward shoot (4060 ± 445 cm 2 ) was significantly (P < 0.001) greater and over twice that of both Shiraz (1714 ± 118 cm 2 ) and Cabernet Sauvignon (1048 ± 120 cm 2 ) shoots. The mean leaf area of Hayward and Shiraz leaves was similar at 175 ± 15 cm 2 and 165 ± 14 cm 2, respectively, but Cabernet Sauvignon leaves were significantly (P < 0.001) smaller (73 ± 8 cm 2 ). The maximum length of unpruned Hayward shoots (1747 ± 329 mm) was also similar to that of pruned Shiraz shoots (1310 ± 100 mm). However, both Hayward (P < 0.01) and Shiraz (P < 0.05) shoots were significantly longer than pruned Cabernet Sauvignon shoots (760 ± 117 mm). Hayward shoots also had a significantly (P < 0.001) greater basal diameter (14.9 ± 0.9 mm) than the other two vines. The basal diameter of Shiraz shoots (10.7 ± 0.6 mm) was significantly (P < 0.05) greater than that of Cabernet Sauvignon shoots (9.3 ± 0.4 mm). These differences in shoot dimensions were reflected in fresh weight of fruit carried on the shoots, with both Hayward (367 ± 69 g) and Shiraz shoots (452 ± 49 g) carrying significantly (P < 0.01) more fruit than Cabernet Sauvignon shoots (101 ± 10 g). 65

102 Figure 3.5: The development of Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots across the 2010/2011 season. A, total leaf area per shoot, B, shoot length and C, shoot basal diameter. Data are means ± SEM, n = 16. The patterns of shoot growth and development (Figure 3.5) were generally reflected in seasonal changes in the dry weight (DW) of the different shoot components in 2010/2011 (Figure 3.6). In keeping with this, differences between the maximum total DW of Hayward (174.5 ± 13.2 g) and Shiraz shoots (113.8 ± 4.9 g) were significantly different (P < 0.05) while shoots of both vines were significantly (P < 0.001) larger than Cabernet Sauvignon shoots (40.5 ± 1.6 g). Allocation of the total shoot weight was divided equally between leaves, stem and fruit on Hayward and Cabernet Sauvignon shoots. However, after the first 100 DAB fruit became the dominant component on Shiraz shoots (40-60% of shoot total DW) until harvest (Figure 3.6), 73% greater than the fruit DW of Hayward shoots and almost four-fold higher than that of Cabernet Sauvignon shoots. During the latter stages of shoot development, particularly after fruit harvest ( DAB), the contribution of the leaves to shoot dry weight declined by 38-58%. 66

103 Furthermore, there was an increase in the DW of the stems (61-68%) but an overall decline in shoot dry weight of all three vines (Figure 3.6). Figure 3.6: Changes in the total dry weight per shoot during the 2010/2011 season. Data for A, Hayward ( ), leaves ( ), stems ( ), and fruit ( ), B, Shiraz ( ), leaves ( ), stems ( ), and fruit ( ) and C, Cabernet Sauvignon ( ), leaves ( ), stems ( ), and fruit ( ) shoots. Note the different scales for each vine cultivar. Data are means ± SEM, n = Gas exchange Photosynthetic light response curves Photosynthetic light responses for mature leaves of each vine cultivar are shown in Figure 3.7 with a hyperbolic tangent fitted to the data from each leaf. Mean values of the parameters describing the curves for each cultivar are shown in Table 3.3. Generally, the photosynthetic light responses for each cultivar were similar, with no significant differences between P max (the light saturated rate of photosynthesis) (P = range ± µmol (CO 2 ) m -2 s -1 ), the photon flux density at which the maximum photosynthetic rate occurred (PFD sat ) (P = 0.84 range ± µmol (photons) m -2 s -1 ) and the apparent photon yield of leaves (Φ app ) (P = range ± mol (CO 2 ) mol (photons) -1 ). The only 67

104 parameter that differed between vines was the dark respiration rate (Rs), where Shiraz leaves had a significantly lower rate (0.31 ± 0.11 µmol (CO 2 ) m -2 s -1 ) than either Cabernet Sauvignon (P < 0.05; 0.72 ± 0.11 µmol (CO 2 ) m -2 s -1 ) or Hayward leaves (P < 0.01; 1.03 ± 0.07 µmol (CO 2 ) m -2 s -1 ). There was no significant difference between respiration rates of Hayward and Cabernet Sauvignon leaves. Figure 3.7: Photosynthetic light responses of youngest mature leaves from A, Hayward ( ), B, Shiraz ( ) and C, Cabernet Sauvignon ( ) shoots at anthesis (December 2010). Data are mean values ± SEM, n = 4. 68

105 Table 3.3: Parameters describing photosynthetic light responses for Hayward, Shiraz and Cabernet Sauvignon leaves measured at anthesis, where P max is the light saturated maximum rate of photosynthesis, Rs is the rate of respiration, φ app is the apparent (CO 2 limited) photon yield and PFDsat is the photon flux density at which photosynthesis was light saturated according to Greer and Halligan (2001). Data for each vine cultivar are means ± SEM, n = 4. Vine cultivar P max Rs Φ app PFD sat µmol (CO 2 ) m -2 s -1 µmol (CO 2 ) m -2 s -1 mol (CO 2 ) mol (photons) -1 µmol (photons) m -2 Hayward 15.3 ± ± ± ± 81 Shiraz 13.6± ± ± ± 81 Cabernet Sauvignon 15.6 ± ± ± ± 38 P value s Seasonal patterns of gas exchange Gas exchange was measured on every second leaf along developing shoots of each vine cultivar throughout the season. There were significant effects (P < 0.001) of vine cultivar and time of season on rates of photosynthesis, stomatal conductance, internal CO 2 concentration and transpiration rates (Table 3.4; Figure 3.8). When averaged over the growing season, Hayward leaves had overall significantly (P < 0.001) higher rates of photosynthesis with mean values of 11.5 ± 0.2 µmol m -2 s -1 compared with the lower rates of 7.0 ± 0.2 µmol m -2 s -1 for Shiraz and 6.7 ± 0.2 µmol m -2 s -1 for Cabernet Sauvignon leaves (Table 3.4). There was a similar trend in seasonal transpiration rates, with Hayward having a significantly (P < 0.001) higher mean rate of 4.13 ± 0.06 mmol m -2 s -1 compared with 2.86 ± 0.05 mmol m -2 s -1 for Shiraz and 2.73 ± 0.07 mmol m -2 s -1 for Cabernet Sauvignon leaves. Hayward leaves also had the highest (P < 0.001) stomatal conductance, with a mean rate of ± mol m -2 s -1 over the growing season, compared with ± mol m -2 s -1 for Shiraz and ± mol m -2 s -1 for Cabernet Sauvignon leaves. However, Shiraz leaves had the highest mean seasonal internal CO 2 concentration at 320 ± 1.7 µmol mol -1 compared with 296 ± 1.4 µmol mol -1 for Hayward leaves and 299 ± 2.2 µmol mol -1 for Cabernet Sauvignon leaves. 69

106 Across the growing season, gas exchange attributes varied on shoots of all vines (Figure 3.8). Rates of photosynthesis were initially low soon after budbreak, but the rates increased rapidly to a peak at around 50 DAB, just prior to anthesis. Thereafter, rates of photosynthesis of Hayward shoots declined slightly then stayed relatively constant for the remainder of the season. Photosynthetic rates of Shiraz and Cabernet Sauvignon shoots also remained constant after anthesis but declined during the later stages of fruit development. Seasonal trends in stomatal conductance of leaves from each cultivar were similar to the photosynthetic rates. Mean shoot transpiration rates of all vines were low early in the season, increased by 50 DAB, remained high until 100 DAB, and then declined again. Transpiration rates of Shiraz and Cabernet Sauvignon leaves remained low for the rest of the growing season; however, transpiration rates of Hayward leaves increased again later in the season. There was no clear seasonal trend for leaf internal CO 2 levels over the growing season (data not shown). Table 3.4: Gas exchange parameters for Hayward, Shiraz and Cabernet Sauvignon leaves averaged over the growing season. Environmental conditions ranged between: temperature o C, PFD µmol (photons) m -2 s -1, CO µmol CO 2 mol -1, VPD kpa). Data are mean values ± SEM. Hayward n = 1120, Shiraz n = 922 and Cabernet Sauvignon n = 545. Vine Photosynthesis Transpiration Stomatal Internal CO 2 cultivar (µmol (CO 2 ) m -2 s -1 ) rate (mmol m -2 s -1 ) conductance (mol m -2 s -1 ) concentration (µmol mol -1 ) Hayward 11.5 ± ± ± ± 1 Shiraz 7.0 ± ± ± ± 2 Cabernet Sauvignon 6.7 ± ± ± ± 2 P value < < < <

107 Figure 3.8: Rates of A, photosynthesis, B, stomatal conductance, and C, transpiration measured concurrently under ambient conditions (temperature o C, PFD µmol (photons) m -2 s -1, CO µmol CO 2 mol -1, VPD kpa) over the growing season. Data for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots are mean values for every second leaf on a shoot ± SEM, n = Gas exchange along the shoots At about the mid stage of shoot development (80 DAB) photosynthesis increased significantly (P < 0.001) from low rates in the oldest leaves (5.0 ± 1.2 µmol m -2 s -1 ) at the base of the shoot to maximum rates at leaf positions 7-9, where rates varied from 12.3 to 13.1 ± 1.2 µmol m -2 s -1 (Figure 3.9). Thereafter, photosynthetic rates in younger leaves near the tip of the shoot declined to about 8.6 ± 2.0 µmol m -2 s -1 though these decreases were not significant. Rates of transpiration followed a similar pattern along a shoot, with maximum rates of 5.6 ± 0.3 mmol m -2 s -1 in the middle of the shoot (leaves 5 to 11) and significantly (P < 0.001) higher than rates of leaves at the base of the shoot (4.2 ± 0.3 mmol m -2 s -1 ). Stomatal conductance of leaves at the base of the shoot (0.24 ± 0.02 mol m -2 s -1 ) was also significantly (P < 0.001) lower than maximum rates of ± 0.02 mol m -2 s -1 at leaf positions 5-11, but again the decline in rates of leaves at the tip of the shoot (0.26 ± 0.04 mol m -2 s -1 ) was not statistically significant. 71

108 Figure 3.9: Changes in gas exchange with leaf position along a shoot from the base to the tip at around 80 DAB. Rates of A, photosynthesis, B, stomatal conductance and C, transpiration. Measurements were made concurrently under ambient conditions (temperature o C, PFD µmol (photons) m -2 s -1, CO µmol CO 2 mol -1, VPD kpa) at approximately 80 DAB, during the very early stages of fruit development. Data are for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots. Data are means ± SEM, n = Gas exchange model From the general linear model (GLM) applied to the seasonal photosynthetic data, the factors that consistently explained the largest amount of variation in photosynthetic assimilation were stomatal conductance, leaf internal CO 2 concentration and photon flux density. Effects of leaf position, time of season and leaf temperature were also generally statistically significant, but accounted for less than 10% of the variance. As these shoot factors had little effect on the variance when added to either additive of multiplicative models of photosynthesis, they were excluded from further analysis. However, the effect of cultivar was consistently significant (P < 0.001) and, therefore, these data were analysed separately for each cultivar. When the key factors of stomatal conductance and internal CO 2 concentration were fitted to an additive model for each of the three vines, r 2 ranged between 0.58 and

109 Including PFD into the model did not greatly improve the fit (from r to 0.88). Using a multiplicative model with stomatal conductance, internal CO 2 concentration and the interaction between these factors considerably improved the amount of variance accounted for (r ) (Table 3.5). Again, adding PFD to the model only improved the r 2 by 0.01 or less, so the multiplicative model using stomatal conductance and internal CO 2 concentration was chosen as the best option. The coefficients for each cultivar are shown in (Table 3.5). The intercept, c i and g s.c i parameters were significantly (P < 0.001) larger, and g s parameter significantly (P < 0.001) lower for Hayward leaves than Shiraz and Cabernet Sauvignon leaves. The only difference between Shiraz and Cabernet Sauvignon leaves was a significantly (P < 0.001) lower g s coefficient for Shiraz leaves. Table 3.5: Coefficients and statistics from fitting a multiplicative model of stomatal conductance and internal CO 2 concentration to photosynthetic data. The percentage of error mean square (EMS) accounted for by each of the terms and the interaction between terms, are presented. The number of measurements used to develop the model for each cultivar (n) is also shown. Vine cultivar Intercept g s c i g s. c i r 2 Hayward n = 1120 Shiraz n = 922 Cabernet Sauvignon n = 545 (µmol m -2 s -1 coefficient (m 2 s mol -1 ) (% EMS) 9.9 ± ± 1.6 (47) 3.7 ± ± 1.2 (19) 3.9 ± ± 1.9 (52) coefficient (mol µmol -1 ) (% EMS) ± (42) ± (41) ± (36) coefficient (m 2 s µmol - 1 ) (% EMS) ± 0.01 (11) ± 0.01 (40) ± 0.01 (12) Net carbon balance of shoots over the growing season In early spring, rates of shoot carbon acquisition were low (Figures 3.10 A - C) as few leaves had developed (Figure 3.5) and photosynthetic rates were also low (Figure 3.8). All vines showed a rapid increase in the rate of carbon 73

110 acquisition during the first DAB. Peak values of carbon acquisition varied significantly (P < ) between vines, from 2.7 ± 0.4 g C shoot -1 day -1 for Hayward vines, 1.2 ± 0.1 g C shoot -1 day -1 for Shiraz and 0.4 ± 0.1 g C shoot -1 day -1 for Cabernet Sauvignon vines. This reflected differences in both the leaf area and the photosynthetic rates of the leaves from the different vines (Figures 3.5 and 3.8). Rates of carbon acquisition remained high for Hayward shoots over the rest of the measurement period until just prior to fruit harvest (220 DAB). Acquisition rates also remained high for Shiraz shoots, but declined 30 days prior to fruit harvest (193 DAB). In contrast, the rates of carbon acquisition by Cabernet Sauvignon shoots were much lower, reaching a peak at berry set (60 DAB) and then declining consistently over the remainder of the growing season, reaching very low levels at fruit harvest (184 DAB). Rates of carbon accumulation in shoot biomass were also significantly different (P < 0.001) between the three vines (Figures 3.10 D - F). Carbon accumulated rapidly in Hayward shoots soon after budbreak, reaching maximum daily rates of 0.76 ± 0.03 g C shoot -1 day -1 around 100 DAB, and thereafter, the rates remained high until close to fruit harvest. By contrast, accumulation of carbon in Shiraz biomass was slow during the early stages of shoot development up to 50 DAB, increased rapidly from 50 to 100 DAB, but then more slowly, reaching maximum daily rates of 0.34 ± 0.02 g C shoot -1 day -1 by 150 DAB. Carbon accumulation as biomass was slowest in Cabernet Sauvignon shoots but rates increased linearly from budbreak, reaching a maximum of 0.12 ± 0.01 g C shoot -1 day -1 at 150 DAB. Rates of carbon accumulation declined in both Shiraz and Cabernet Sauvignon shoots as fruit growth slowed. During the first 30 days of shoot development, the daily net carbon balance of shoots from all three vines was very low or negative (Hayward ± 0.01, Shiraz ± 0.001, Cabernet Sauvignon ± g C shoot -1 day -1 ), as the carbon required to support rapid shoot growth was greater than that produced by the young leaves (Figures 3.10 G - I). Shoots of all three vines then showed a rapid increase in their net carbon balance, reaching 74

111 maximum surpluses of 1.92 ± 0.15, 0.93 ± 0.07 and 0.35 ± 0.07 g C shoot -1 day -1 for Hayward, Shiraz and Cabernet Sauvignon shoots respectively between 75 and 100 DAB. Again, the net carbon gain per day was significantly (P < 0.001) greater in Hayward shoots compared to other vines, and higher (P < 0.001) in Shiraz compared with Cabernet Sauvignon shoots. The net daily carbon balance followed a similar trend to net carbon acquisition of shoots over the growing season (Figures 3.10 A - C). Hayward and Shiraz shoots reached maximum rates around 100 DAB and remained high throughout most of fruit development. The daily carbon balance of Shiraz shoots declined markedly during fruit maturation prior to harvest but remained positive. In contrast, the daily carbon balance of Cabernet Sauvignon shoots peaked at 75 DAB, and then declined consistently throughout the remainder of the season, becoming negative once again at harvest. 75

112 76 Figure 3.10: Changes in rates of carbon acquisition (A - C), carbon accumulated as biomass (D - F) and the net carbon balance of shoots (G - I) of Hayward ( A, D, G), Shiraz ( B, E, H) and Cabernet Sauvignon ( C, F, I) shoots. Note the different scales between Hayward and other vines for each parameter. Data are means ± SEM, n = 6. 76

113 3.3.3 Total carbohydrate concentrations Seasonal changes in total carbohydrate concentrations Seasonal changes in TNC concentrations in leaves, the internode and developing axillary buds of shoots are shown for Hayward and Shiraz over three seasons and for Cabernet Sauvignon in 2010/2011 in Figure In the leaves, TNC concentrations rose rapidly (Figures 3.11 A - C), to reach maximum levels at anthesis (Hayward ± 5-9 mg gdw -1, Shiraz ± mg gdw -1 and Cabernet Sauvignon 140 ± 10 mg gdw -1 ) when the leaf area of shoots was rapidly expanding and photosynthetic rates were high (Figures 3.5 and 3.8). TNC concentrations in leaves then declined, coincident with shoot growth slowing at 100 DAB (Figure 3.5) and concentrations remained constant at around 60 ± 5 mg gdw -1 for Hayward, 120 ± 6 mg gdw -1 for Shiraz leaves and 215 ± 20 mg gdw -1 for Cabernet Sauvignon leaves for the rest of the growing season. However, the Shiraz leaves generally had significantly (P < 0.001) higher concentrations (mean across the growing season 143 ± 9 mg gdw -1 ) than Hayward (mean 84 ± 6 mg gdw -1 ). In the 2010/2011 season, when Cabernet Sauvignon vines were added to the study, leaves from this cultivar had significantly (P < 0.05) lower TNC concentrations (108 ± 11 mg gdw -1 ) from early in the season until fruit set than occurred in Shiraz leaves (157 ± 8 mg gdw -1 ), but then had significantly higher (P < 0.01) concentrations (172 ± 14 mg gdw -1 ) than either Hayward (78 ± 5 mg gdw -1 or Shiraz (110 ± 10 mg gdw -1 ) leaves during the remainder of the growing season. TNC concentrations in the axillary buds of all vines were markedly lower than that occurring in the leaves (Figures 3.11 D - F); however, over much of the season the TNC concentrations of axillary buds of Shiraz and Cabernet Sauvignon shoots remained relatively constant, though an increase in concentration occurred at the end of the season before leaf fall (Figures 3.11 D - F). In Shiraz buds, TNC concentrations remained at around ± 2-12 mg gdw -1 in 2008/2009 (Figure 3.11 D) and ± 4-14 mg gdw - 1 in 2009/2010 (Figure 3.11 E). In 2010/2011 TNC concentrations in axillary buds of both Shiraz and Cabernet Sauvignon shoots were lower, at around ± 1-10 mg gdw -1 (Figure 3.11 F). TNC concentrations in Hayward 77

114 axillary buds (50-60 ± 3-5 mg gdw -1 ) were not significantly different from those in Shiraz (50-70 ± 3-5 mg gdw -1 ) and Cabernet Sauvignon buds 45 ± 4 mg gdw -1 ) at the beginning of their development. However, concentrations in Hayward buds declined to ± 1-5 mg gdw -1 during shoot elongation, significantly (P < 0.05) lower than the concentrations in Shiraz ( ± 2-10 mg gdw -1 ) and Cabernet Sauvignon (50-66 ± 2-10 mg gdw -1 ) buds. Concentrations remained low in Hayward buds throughout the rest of the growing season, (Figures 3.11 D - F). TNC concentrations in the internodes of Hayward and Shiraz shoots were also markedly lower than the concentrations in the leaves. Generally TNC concentrations in the internodes of these vines were similar and followed similar patterns over development. Concentrations peaked in spring when shoots began to grow rapidly (Hayward ± mg gdw -1, Shiraz shoots ± 7-14 mg gdw -1 ) (Figures 3.11 G - I). During the rapid shoot extension and early fruit development phase, up to about 100 DAB, TNC concentrations declined to ± 3-10 mg gdw -1 in Hayward internodes and to ± 3-13 mg gdw -1 in Shiraz internodes. Thereafter, the concentration of TNC in the internodes of shoots generally increased steadily until the end of the growing season, when TNC concentrations in Shiraz internodes ranged from 100 to 150 ± 3-6 mg gdw -1 and in Hayward internodes from 80 to 100 ± 4-6 mg gdw -1. However in the 2009/10 season, there was no clear pattern in the changes in TNC concentrations in Shiraz internodes (Figure 3.11 H). Although internodes of Cabernet Sauvignon shoots initially had significantly lower TNC concentrations during early shoot development (60-80 ± 5-7 mg gdw -1 ) in the 2010/2011 season, thereafter, concentrations were not significantly different to those concentrations in the internodes of the other vines (Figure 3.11 I). 78

115 79 Figure 3.11: Changes in TNC in the leaves (A - C), developing axillary buds (D - F) and internodes G - I) of shoots during the 2008/09 (A, D, G), 2009/10 (B, E, H) and 2010/11 (C, F, I) growing seasons. Data are shown for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) shoots. Note different scales for leaves, axillary buds and internode TNC concentrations. Data points are the means ± SEM, n = 4. 79

116 Prior to anthesis in the 2010/2011 season, the TNC concentration of inflorescences from all three vines were similarly low, ranging between 60 and 78 ± 4-5 mg gdw -1 at DAB (Figure 3.12). Thereafter, the TNC concentrations in Hayward berries increased steadily and were significantly greater (P < 0.001) than those of the other vines during early fruit development. Concentrations of TNC in Shiraz and Cabernet Sauvignon berries remained low during the first part of fruit development and then, after veraison, increased steeply during the last two months of fruit development. At harvest, final TNC concentrations were significantly (P < 0.05) higher in Shiraz (614 ± 32 mg gdw -1 ) than Hayward (486 ± 50 mg gdw -1 ) and Cabernet Sauvignon (477 ± 35 mg gdw -1 ) berries. Figure 3.12: Changes in TNC concentration in inflorescences and berries of three vines across the 2010/2011 season. Data are shown for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) berries. Data are means ± SEM, n = 4. The seasonal pattern of TNC concentrations in older vine organs, such as the wood subtending the current season s shoots and the fine roots were generally similar to each other (Figure 3.13). In the three vines, maximum TNC concentrations ( ± 3-7 mg gdw -1 ) measured in the wood just 80

117 prior to budbreak, were not significantly different (Figures 3.13 A - C). Thereafter, there was a decline in TNC concentrations in the wood until the end of shoot growth when minimum concentrations ranged between 27 and 41 ± 2-8 mg gdw -1 and were not significantly different between the vines. There was a steady increase in TNC concentrations in wood throughout autumn until maximum concentrations were reached in winter. Seasonal changes in TNC concentrations in fine roots of vines were of a similar magnitude to those in the wood (Figures 3.13 D - F). Concentrations of TNC were highest in fine roots during winter. Maximum concentrations in Hayward and Cabernet Sauvignon roots were ± 9-16 mg gdw -1 and not significantly different. However, maximum concentrations in the roots of Shiraz vines ( ± mg gdw -1 ) were significantly (P < 0.05) higher than those in the roots of the other two vines. TNC concentrations declined rapidly to minimum concentrations in spring. In the 2009/2010 season, both Hayward and Shiraz roots reached minimum TNC concentrations of 19 ± 5 mg gdw -1 and 13 ± 1 mg gdw -1 respectively at DAB. However, in the following season the minimum concentrations of ± 1-15 mg gdw -1 were not reached until DAB. Concentrations then remained low for much of summer (until 150 DAB), before increasing steadily during autumn, until maximum concentrations were reached in the winter each season (Figures 3.13 D - F). There was a very marked increase in the TNC concentrations of Shiraz roots, of about 100 mg gdw -1, during the autumn of both seasons (Figures 3.13 D - F). 81

118 Figure 3.13: Changes in TNC concentration in wood during A, 2008/2009, B, 2009/10 and C, 2010/11 seasons and fine roots over D, 2008/2009, E, 2009/10 and F, 2010/11 seasons. Data are shown for Hayward ( ), Shiraz ( ) and Cabernet Sauvignon ( ) vines. Data are means ± SEM, n = Seasonal changes in total sugars and starch concentrations The starch and total sugar concentrations are shown for both annual growth (Figure 3.14) and perennial growth in 2010/2011 (Figure 3.15). Seasonal changes in the concentration of starch and soluble sugars in the organs of all vines tended to follow one of three different patterns. Fruit of all the vines tended to have low starch (4-28 mg g DW -1 ) and total sugar concentrations (47-77 mg g DW -1 ) early in their development with a rapid increase in concentrations from around 100 DAB until harvest (Figures 3.14 J - L). Before anthesis, the inflorescences of each cultivar contained both starch and sugars; however, sugars made up the highest proportion of total carbohydrate during these early stages of berry development (0-100 DAB). Thereafter, the concentrations of carbohydrates increased rapidly from veraison in Shiraz and Cabernet Sauvignon berries and from slightly before seed maturation began in Hayward berries. However, the form of carbohydrate accumulated in berries differed between the vines, with 82

119 significantly (P < 0.001) more soluble sugars accumulating in Shiraz and Cabernet Sauvignon compared to Hayward berries and more (P < 0.001) starch accumulating in Hayward berries than the other two vines. The second seasonal pattern of change in carbohydrate concentrations occurred in the leaves, internodes, and to a lesser extent, axillary buds of all vines (Figure 3.14). These organs generally had the highest concentrations of carbohydrate during the early stages of development (until DAB) and lower, either steady or slightly increasing concentrations thereafter. However, starch was the predominant carbohydrate (57%) that accumulated during early leaf development (Figures 3.14 A - C) whereas soluble sugars made up 80-90% of carbohydrate in rapidly developing internodes (Figures 3.14 G - I) and axillary buds (Figures 3.14 D - E). From around 100 DAB, after the initial period of rapid shoot growth, Hayward and Shiraz leaves had higher concentrations of sugars (80% of total carbohydrate concentrations) and lower concentrations of starch, consistent with export from the leaves. Although the total sugar the leaves of Cabernet Sauvignon shoots were similar to the leaves of other vines during the later stages of shoot development, starch concentrations were significantly (P < 0.001) higher than those in the leaves of other vines and equivalent to total sugars in Cabernet Sauvignon leaves during this period. During early shoot development, the concentration of sugars in the internodes of both Hayward and Shiraz were significantly (P < 0.01) higher than those in the internodes of Cabernet Sauvignon shoots. By contrast, there was a gradual accumulation of starch reserves to mg g DW -1 in the internodes of all vines from 100 DAB when the concentration of total sugars remained relatively constant at about mg g DW -1. Starch concentrations also increased in the axillary buds of Shiraz and Cabernet Sauvignon vines during the later stages of shoot development. The third pattern of seasonal change in carbohydrates occurred in the perennial wood (Figures 3.15 A - C) and roots (Figures 3.15 D - F)) of all the vines. During rapid shoot growth over the first 100 DAB, there was generally a marked decrease in the starch concentrations to less than 20 mg g DW -1 in 83

120 these tissues, and this is consistent with these organs supplying carbohydrate to support new shoot growth. Thereafter, the starch concentrations increased throughout the remainder of the growing season as reserves were replenished peaking at mg g DW -1 in perennial wood and mg g DW -1 in roots depending on vine cultivar. Concentrations of sugars in these organs showed no consistent pattern over the growing season. Changes in starch concentrations in both the perennial wood and roots showed marked seasonal changes in Shiraz vines. However, in both Hayward and Cabernet Sauvignon vines seasonal changes in starch concentrations in the perennial wood were greater than occurred in the roots. The peak concentrations in the wood of both Shiraz and Cabernet Sauvignon vines and the roots of Shiraz vines were significantly (P < 0.01) higher than those in wood and roots of Hayward vines. 84

121 85 Figure 3.14: Changes in starch ( ) and total sugar concentrations ( ) in the leaves (A - C), developing axillary buds (D - F) internodes G - I) and fruit (J - L) of shoots during the 2010/11 growing seasons. Data are shown for Hayward (, ), Shiraz (, ) and Cabernet Sauvignon (, ) shoots. Note different scales for fruit concentrations. Data points are the means ± SEM, n = 4. 85

122 Figure 3.15: Changes in starch and total sugar concentrations in the wood (A - C), and root (D - F) of vines during the 2010/11 growing seasons. Data are shown for Hayward (starch, total sugars ), Shiraz (starch, total sugars ) and Cabernet Sauvignon (starch,total sugars ). Data points are the means ± SEM, n = Total carbohydrate content of shoots Hayward shoots had a significantly higher (P < 0.01) total carbohydrate content in leaves (5.5 ± 0.8 g shoot -1 ) than Shiraz (1.7 ± 0.1 g shoot -1 ) and Cabernet Sauvignon (2.0 ± 0.1 g shoot -1 ) shoots (Figure 3.16), because of their greater leaf area and leaf mass (Figures 3.5 and 3.6). The maximum carbohydrate content of Hayward (8.0 ± 1.4 g shoot -1 ) and Shiraz stems (5.9 ± 0.2 g shoot -1 ) were not significantly different, but Hayward shoots had a (P < 0.01) higher stem carbohydrate content than Cabernet Sauvignon shoots (2.6 ± 0.3 g shoot -1 ) (Figure 3.16). The carbohydrate content of Hayward leaves was significantly (P < 0.001) 2-fold higher than stems during the first half of the season, but the carbohydrate content of leaves and stems of Shiraz and Cabernet Sauvignon shoots were similar during this period. From around 100 DAB, the carbohydrate content of stems increased steadily. Later, after harvest, the carbohydrate content of leaves declined 86

123 while that in stems continued to increase, reaching maximum values at the end of the season in shoots of all vines. Figure 3.16: Changes in total shoot carbohydrate content of A, Hayward leaves ( ) and stems ( ), B, Shiraz leaves ( ) and stems ( ) and C, Cabernet Sauvignon leaves ( ) and stems ( ) across the 2010/2011 season. Data are mean values ± SEM, n = 4. The fruit component had a dramatic effect on carbohydrate content of shoots during the growing season (Figure 3.17). Between 70 and 170 DAB the combined increase in fruit dry weight and fruit carbohydrate concentration produced a significant (P < 0.001) and 10-fold increase in shoot carbohydrate content for Hayward, a 15-fold increase for Cabernet Sauvignon shoots and a 22-fold increase for Shiraz shoots, (Figures 3.6, 3.12 and 3.17). At harvest, the carbohydrate content of fruit on Shiraz shoots (47.0 ± 2.4 g) was significantly higher (P < 0.001) than that on Hayward shoots (21.0 ± 3.3 g), and both were significantly higher than that on Cabernet Sauvignon shoots (9.3 ± 0.8 g). Both Hayward (30.5 ± 3.1 g) and Shiraz shoots (51.1 ± 2.8 g) had significantly (P < 0.001) higher total carbohydrate content than 87

124 Cabernet Sauvignon shoots (12.5 ± 1.0 g) (Figure 3.17). The difference in shoot carbohydrate content between Shiraz and Cabernet Sauvignon shoots was partially due to the large bunches carried by Shiraz shoots in the 2010/2011 season (430 ± 17 g) compared with Cabernet Sauvignon shoots (194 ± 15 g). Figure 3.17: Changes in carbohydrate content of shoots (solid symbols) and fruit (open symbols) throughout the growing season for Hayward shoots ( ), Shiraz shoots ( ) and Cabernet Sauvignon shoots ( ). Data are means ± SEM, n = Whole vine total carbon budget A whole vine carbon budget was calculated for vines prior to harvest using dry weight data from whole Shiraz vines (Field 2013) and Hayward vines (Smith et al. 1992) as whole vine biomass could not be obtained from the commercial vines used in this study and annual shoot biomass from the current study was comparable to that measured in these previous studies. Whole vine biomass data and the TNC concentrations from this study were converted to total carbon and non-structural carbon respectively, and the difference between the two was structural carbon (Table 3.6). In the earlier studies, Hayward vines were almost eight times larger than Shiraz vines on a total dry weight basis and, therefore, also on a total and structural carbon 88

125 basis. Hayward vines also had a higher percentage of total carbon in the internode (12.3%) and leaves (14.1%) than Shiraz vines (internode 7.7% and leaves 9.0%), a similar proportion in fruit (Hayward 24.5% Shiraz 20.6%), but a lower proportion in perennial cordon (Hayward 12.1% Shiraz 16.7%), trunk (Hayward 7.7% Shiraz 16.9%) and roots (Hayward 24.3% Shiraz 29.1%) (Table 3.6). However, Hayward vines also have an additional canopy component of one-year-old canes, which contained 5% of the total vine carbon. The distribution of structural carbon in the three vines was similar to the distribution of total carbon (Table 3.6). On a whole vine basis, Hayward vines had more than four times as much total non-structural carbon per vine as Shiraz and Cabernet Sauvignon vines (Table 3.6). The percentage of non-structural carbon allocated to different organs varied considerably between the vines (Table 3.6). Shiraz vines had the highest proportion of non-structural carbon in roots (33.0%) and fruit (32.7%), but the lowest proportion of TNC in the cordon (6.4%) and leaves (10.7%). By contrast, Cabernet Sauvignon had the highest proportion of non-structural carbon in the trunk (13.2%), cordon (13.0%) and leaves (22.0%) and the lowest proportion in the internode (8.0%) and fruit (21.7%). Hayward vines had the highest proportion of non-structural carbon in internodes (18.5%) and the lowest proportion in the vine trunk (4.4%). 89

126 90 Table 3.6: Estimated whole vine carbon allocations determined from vine dry weight data from Field (2013) for Shiraz and Cabernet Sauvignon vines and data from Smith et al. (1992) for Hayward vines. Total non-structural carbon content per vine component was calculated using dry weight data and TNC concentrations from immediately prior to fruit harvest (23 March 2011 Shiraz and Cabernet Sauvignon and 27 April 2011 Hayward ). Total structural carbon is the difference between total carbon and total non-structural carbon. Percentages each component contributed to the total vine are shown in brackets. Vine component Shiraz Roots 1042 (29.1) Trunk 605 (16.9) Cordon 596 (16.7) One-yearold cane Internode 275 (7.7) Leaves 323 (9.0) Fruit 735 (20.6) Total carbon g vine -1 (%) Cabernet Sauvignon 1042 (29.1) 605 (16.9) 596 (16.7) Total structural carbon g vine -1 (%) Hayward Shiraz Cabernet Sauvignon 6579 (24.3) 2070 (7.7) 3263 (12.1) (5.0) 275 (7.7) 323 (9.0) 735 (20.6) 3330 (12.3) 3825 (14.1) 6615 (24.5) 930 (28.7) 583 (18.0) 574 (17.8) 996 (30.6) 563 (17.3) 555 (17.0) Total non-structural carbon g vine -1 (%) Hayward Shiraz Cabernet Sauvignon 6363 (24.9) 2002 (7.8) 3155 (12.4) (4.9) 247 (7.6) 286 (8.8) 623 (19.3) 249 (7.6) 253 (7.8) 666 (20.4) 3047 (11.9) 3531 (13.8) 6148 (24.1) (33.0) 22.3 (6.5) 22.0 (6.4) 46.1 (14.5) 41.8 (13.2) 41.1 (13.0) Hayward (14.1) 68.0 (4.4) (7.0) (6.3) 27.6 (18.1) 36.6 (10.7) (32.7) 25.4 (8.0) 69.4 (22.0) 68.8 (21.7) (18.5) (19.2) (30.5) Total

127 3.4 Discussion This chapter examines the hypothesis that seasonal patterns of development, in both grape and kiwifruit are similar and that the development of both vines are strongly influenced by carbohydrate supply. In this chapter the phenology, carbon acquisition, carbohydrate accumulation and the distribution of biomass and in key organs of Shiraz, Cabernet Sauvignon and Hayward vines was compared during vine development. Although the vines were grown at different sites and under different management this chapter examined basic physiological properties of vines. Generally, the seasonal patterns of growth and changes in carbohydrate concentrations were similar between the vines. However, there were differences in the extent of growth with higher budbreak on Shiraz and Cabernet Sauvignon vines, but greater growth of shoots, particularly leaves, and longer periods of root growth in Hayward vines. These differences influenced carbohydrate acquisition, which was greater in Hayward shoots because of their larger leaf area. However, the concentration of TNC in key organs was higher in Shiraz and Cabernet Sauvignon vine organs than those in Hayward vines Seasonal patterns of vine growth Phenological development of Shiraz and Cabernet Sauvignon grapevines and Hayward kiwifruit vines followed similar seasonal patterns. The pattern of budbreak was generally similar for the three vines, with Shiraz buds breaking first, followed by Cabernet Sauvignon and then Hayward buds. This variation in timing was typical for different cultivars of both grape (Coombe 1988) and kiwifruit vines (Thorp et al. 1990). In addition, the seasonal response of the Shiraz and Hayward buds to environmental conditions was similar, with an earlier budbreak in the 2009/2010 season and later development in 2010/2011. This difference was likely due to the 1.2 o C lower mean temperature in the 2009 winter (see Appendix 1) advancing the date of budbreak in the 2009/2010 season. Previous studies of both Hayward (McPherson et al. 1994; McPherson et al. 2001) and several grape cultivars (Williams et al. 1985a; Moncur et al. 1989; Lavee and May 1997; Dokoozlian 1999) have shown that the timing of budbreak is advanced by 91

128 cool winter temperatures. The flowering of each vine cultivar followed a sigmoidal pattern over time that was comparable to budbreak. However, although Hayward canes were the last to break bud, their inflorescences reached anthesis first, followed by those on Shiraz and then Cabernet Sauvignon canes. Therefore, although the timing of events was similar, the period of floral morphogenesis was shorter in Hayward (average 54 days) than Shiraz (average 72 days) and Cabernet Sauvignon (63 days) vines. Another major difference in the timing of development between the vines was the earlier fruit maturation and harvest on Shiraz (average 32 days earlier) and Cabernet Sauvignon (33 days earlier) vines than that on Hayward shoots. For Hayward, both the amount of budbreak and the number of inflorescences produced per winter bud fell within the range previously reported for vines in this region (McPherson et al. 1994; McPherson et al. 2001). There are no comparable data for Shiraz and Cabernet Sauvignon in this region; however, budbreak fell within the reported range for Sultana, Semillon and Cabernet Sauvignon grapevines in Australia (Antcliff and Webster 1955; Dunn and Martin 2000; Holzapfel et al. 2006). Hayward vines produced twice as many inflorescences per floral shoot as Shiraz and Cabernet Sauvignon. However, as Shiraz and Cabernet Sauvignon canes produced twice as many floral shoots as Hayward, the number of inflorescences produced per winter bud was, in fact, similar between the three vines. Higher budbreak and floral production on both Hayward and Shiraz vines in the 2010/2011 appeared to occur after a warm, rather than cool winter as previously reported for Hayward (McPherson et al. 1995; McPherson et al. 2001; Snelgar et al. 2008) and grape cultivars (Lavee and May 1997). This suggests other factors, such as carbohydrate supply, may also have an influence on inflorescence production. After budbreak, rapid changes in shoot growth occurred and followed a similar pattern in all three vines. Thereafter, the rate of shoot growth declined in the early stages of fruit development (100 DAB) in the first two seasons, but continued until around 145 DAB in the 2010/2011 season. In 92

129 the 2010/2011 season, minimum temperatures in December and January were o C higher, on average, than in the previous two seasons (see Appendix 1) and 2 o C higher than the 30 year average for this region. A previous study of Pinot Noir and Riesling vines showed that similar increases in minimum temperatures resulted in a 34-63% increase in shoot growth rates (Hendrickson et al. 2004) and high temperatures in controlled environment studies of Hayward vines also resulted in increased shoot growth rates (Greer and Jeffares 1998; Richardson et al. 2004). Both the current and these previous studies are consistent with the conclusion that shoot growth of grape and kiwifruit vines was temperature-dependent and also supported the conclusion that the observed differences in shoot growth during the different seasons in the present study was influenced by the seasonal temperatures. The seasonal pattern of fruit dry weight accumulation per shoot was similar between the vines. During the early stages of fruit development, the increase in fruit biomass per shoot was slow and less than that of vegetative growth. However, around 100 DAB (about 50 days after anthesis in Hayward and 30 days after anthesis in Shiraz and Cabernet Sauvignon ) the rate of accumulation of fruit biomass increased for all vines, as has been reported in previous studies of grape (Coombe 1976) and kiwifruit (Hopping 1976). In Shiraz, the rate of increase in fruit biomass was much greater than the increase in shoot vegetative biomass; however, the allocation between fruit and vegetative biomass was proportionally similar on both Cabernet Sauvignon and Hayward shoots. Although the timing of shoot development for the three vines was generally similar, the actual amount of growth varied markedly. Hayward and Shiraz shoots grew to a similar length but the total leaf area and dry weight of Hayward shoots (excluding berries) was twice that of Shiraz shoots. However, due to a higher budbreak, there was approximately twice the number of shoots per Shiraz parent cane as per each Hayward parent cane. Therefore, the overall shoot mass per parent cane was similar between the two vines. Furthermore, both Hayward and Shiraz shoots 93

130 were at least twice the size and carried twice the biomass of berries of the Cabernet Sauvignon shoots. This conforms with Cabernet Sauvignon vines being recognised as having a lower vigour than Shiraz vines (Dry and Gregory 1988). However, it is also possible that vine management (two canes on Shiraz vines and four canes on Cabernet Sauvignon ) may also have contributed to this difference, resulting in shoot mass per vine being similar between the two vines. Patterns of root growth were consistent between seasons for each vine cultivar, but the extent of seasonal root growth differed between the vines. Both Shiraz and Cabernet Sauvignon vines had two short periods of root growth, the first occurred near the end of rapid shoot growth in spring and the second occurred after harvest at the end of the growing season. This pattern of root growth was similar to the bimodal pattern of root growth previously observed in grapevines as reviewed by Williams and Matthews (1990). In Hayward vines, root growth began before anthesis, continued throughout the later stages of rapid shoot development and also through much of the fruit development period. A second shorter period of root growth occurred in Hayward vines after fruit harvest in winter. This pattern of root growth was more extensive than previous reports of a single major flush of root growth in Hayward vines in late summer (Buwalda and Hutton 1988) or a flush in summer followed by a second peak in autumn (Black 2012). Therefore overall, the timing of root growth of Hayward vines was broadly similar but extended over a much greater period than occurred with either Shiraz or Cabernet Sauvignon vines. The pattern of root loss (death of root tips) also varied between the three vines and across seasons. There was a consistent period of root loss in all three vines around 50 DAB in spring, coinciding with the occurrence of rapid shoot development, floral morphogenesis, minimum TNC concentrations in roots and the beginning of a positive net daily carbon balance in shoots. Similarly Black (2012) observed an extended period of root death in Hayward vines at about the same stage of development, although Reid et al. (1993) found that Hayward vines had a high turnover of roots, and hence 94

131 root loss occurred throughout the growing season. By contrast, Comas et al. (2005) found that root mortality peaked in Concord grapevines after the flush of root growth in spring rather than before it as observed in this study. Thus, the loss of roots in late spring in several studies is consistent with the depletion of root reserves Carbohydrate composition Total non-structural carbohydrate (TNC) concentrations in leaves, axillary buds and internodes were higher in Shiraz and Cabernet Sauvignon shoots compared to those in Hayward shoots. Leaves were sampled from vines during late morning - mid-afternoon when photosynthesis was high, carbohydrate concentrations were increasing and some carbohydrate export was occurring (Chaumont et al. 1994; Greer 1999; Black et al. 2012). During early leaf growth, TNC concentrations in leaves immediately distal to the last inflorescence generally declined and then remained constant throughout the remainder of the season in all vines. This is similar to reports from mature Carignane grape leaves (Winkler and Williams 1945) and similar to the change in starch concentrations in leaves on Hayward vines (Smith et al. 1992). The percentage of starch was initially high in newly developed leaves of all vines, consistent with little export occurring, but the starch concentration decreased as photosynthesis reached peak rates and excess carbon was available to be exported from leaves. At this point, sugars became the major compound in the Hayward and Shiraz leaves, again, consistent with carbon export occurring. In several other grape cultivars, sucrose levels also increased in the later stages of leaf development, although starch concentrations were not measured (Kliewer 1966; Wu et al. 2011). However, in the present study, starch remained the major compound in Cabernet Sauvignon leaves throughout the season, although high sugar concentrations were also present. This strongly reflects the different pattern of shoot development and reinforces the conclusion that little carbon export occurred from the Cabernet Sauvignon leaves compared with the leaves of the Hayward and Shiraz shoots. 95

132 The TNC concentration in axillary buds on Hayward shoots was high early in the season and then declined, however, TNC concentrations in Shiraz and Cabernet Sauvignon axillary buds generally increased over the growing season with some starch accumulation. The concentration of TNC in grape buds in this study was similar to that reported in Merlot (Koussa et al. 1998), but in the present study, the seasonal change in carbohydrates were less marked. There are no previous studies on the TNC changes in whole Hayward buds throughout the season; however, the increase in sugars in buds in autumn was similar to that previously occurring in Hayward axillary bud meristems (Richardson et al. 2010). There was an increase in internode sugars early in the development of actively growing shoots of all the vines and such an increase has been previously associated with active cell division in A. deliciosa fruit (Nardozza et al. 2013). Higher concentrations of sucrose in shoots have been previously reported in grapevines during active shoot development (Kliewer 1966; Hamman et al. 1996). Sugar concentrations within internodes of all vines declined after the initial period of rapid shoot extension up to 100 DAB. Thereafter starch concentrations in the internodes of maturing canes increased. This is comparable to the accumulation starch in Carignane (Winkler and Williams 1945) and Hayward (Smith et al. 1992) shoots. Furthermore, the accumulation of carbohydrates in internodes from around 100 DAB occurred at a similar time to the accumulation of starch in perennial wood and roots in this study. Despite the fundamental differences in fruit physiology between Hayward (starch accumulation and climacteric) and Shiraz and Cabernet Sauvignon (soluble sugar accumulation and non-climacteric) berries, the early patterns of carbohydrate accumulation in inflorescences and berries were similar. Flowers and young berries of each cultivar had similar concentrations of TNC and all the inflorescences contained starch. In each case, the percentage of starch in inflorescences declined to low levels at or soon after anthesis. The patterns of change in starch and sugar concentrations were broadly similar to that previously reported in Pinot Noir and Gewürztraminer inflorescences 96

133 (Lebon et al. 2004b) but there have been no previous reports of the carbohydrate composition and seasonal change in inflorescences of Hayward or other Actinidia species. Sugar concentrations increased in the berries of all vines during the early stages of fruit development when cell division, a metabolically expensive process, was occurring. Increasing sugar concentrations have been shown in early berry development of grape and kiwifruit (Kliewer 1966; Klages et al. 1998; Boldingh et al. 2000; Wu et al. 2011). Later in development, grape berries began to accumulate sugars and Hayward berries began to accumulate starch. These patterns are similar to that occurring during the development of a number of grape varieties (Kliewer 1965; Wu et al. 2011) and Actinidia species (Klages et al. 1998; Boldingh et al. 2000). Sugar and starch accumulation appear to have a consistent role across different vines in supporting early cell division and then the transition from cell division to increased cell expansion and sugar/starch accumulation in fruit (Nardozza et al. 2013). Concentrations of TNC in the perennial organs (wood and roots) generally followed similar patterns across the three vines, despite differences in the timing of root growth. In perennial wood, peak carbohydrate concentrations occurred in winter, then levels declined rapidly during the peak period of annual growth (up to 100 DAB), before concentrations in wood increased as reserves were replenished again in late summer/autumn. Peak TNC concentrations in the perennial wood of Shiraz and Cabernet Sauvignon reached 15% of DW but were slightly lower in Hayward at 10-12% of DW. This is lower than the peak concentrations of around 20% recorded in previous studies of grape cultivars, for example Carignane (Winkler and Williams 1945), Thompson Seedless (Williams 1996) and Chardonnay (Bennett et al. 2005), but similar to that in Shiraz wood (Holzapfel and Smith 2012). The major compound found in the wood of all three vines were starch (40-80%) and seasonal trends in starch were similar to those in previous studies (Winkler and Williams 1945; Smith et al. 1992; Williams 1996; Weyand and Schultz 2006; Holzapfel and Smith 2012). This decline in starch concentrations reflects the importance of the wood to store reserves 97

134 and the mobilisation of reserves to support vines, especially before the shoot becomes autotrophic. The seasonal pattern of change in TNC concentration in the roots was similar to that in wood but the TNC concentration in roots generally declined earlier, around budbreak and also began to accumulate earlier from 50 to 100 DAB, although this was seasonally dependent. The pattern also reflects the seasonal trends found in previous grapevine and kiwifruit studies (Winkler and Williams 1945; Smith et al. 1992; Williams 1996; Boldingh et al. 2000; Bates et al. 2002; Bennett et al. 2005; Holzapfel and Smith 2012) and for other deciduous crops (Oliveira and Priestley 1988; Forshey and Elfving 1989; Kozlowski 1992). The concentration of TNC in the roots of Shiraz vines at 15-20% DW was the highest of vines in this study (10% in Cabernet Sauvignon and Hayward ) but still lower than those recorded in previous grapevine studies (20-30%) (Winkler and Williams 1945; Williams 1996; Bennett et al. 2005; Holzapfel and Smith 2012). However, the TNC concentration in Hayward roots were higher than those recorded in a previous study of A. deliciosa vines (5%) (Boldingh et al. 2000). Starch was the major carbohydrate found in roots of all three vines, comprising 40-80% of TNC concentrations, emphasising the importance of roots as primary storage organs. The percentage of starch in Shiraz roots in the current study increased rapidly as shoot development slowed around 150 DAB but was slower to increase in Cabernet Sauvignon roots, where net export from the shoots was low and Hayward where the root biomass was larger and the fruit remained on vines for longer. Again, this seasonal pattern of starch mobilisation and accumulation was consistent with root carbohydrate reserves forming the primary carbohydrate sink to supply the developing shoot in spring Gas exchange Photosynthetic light responses of the youngest mature leaves of the three vines were almost identical. The light saturated rates of photosynthesis, the PFD at which photosynthesis was light saturated, and the apparent photon yield of leaves were all similar and well within the range reported in earlier 98

135 studies of Cabernet Sauvignon (Poni and Intrieri 2001), Sangiovese (Cartechini and Palliotti 1995), Sultana (Kriedemann 1968), Semillon (Greer et al. 2011) Hayward (Greer and Halligan 2001) and other Actinidia deliciosa, A. chinensis, A. eriantha and A. arguta genotypes (Jo et al. 2008). The only difference in gas exchange between the vines was the low dark respiration rates recorded for Shiraz compared with Hayward and Cabernet Sauvignon leaves. Low rates from well exposed Shiraz leaves on non-fruiting shoots in this study (0.31 ± 0.01 µmol (CO 2 ) m -2 s -1 contrasted with higher rates of well exposed leaves from fruiting Shiraz shoots (1.16 ± 0.29 µmol (CO 2 ) m -2 s -1 ) (see Chapter 6), though they were comparable with those of shaded fruiting shoots (0.32 ± 0.07 µmol (CO 2 ) m -2 s -1 ) in that Chapter. Dark respiration rates recorded on well exposed Shiraz leaves from fruiting shoots (0.4 ± 0.2 µmol (CO 2 ) m -2 s -1 ) (Rogiers and Clarke 2013) were similar to those measured in the current study. The dark respiration rates for both Hayward (1.03 ± 0.07 µmol (CO 2 ) m -2 s -1 ) and Cabernet Sauvignon leaves (0.72 ± 0.11 µmol (CO 2 ) m -2 s -1 ) were comparable to those in earlier studies (Piller and Meekings 1997; Greer and Jeffares 1998; Escalona et al. 2012). Thus, from this and other studies, it would suggest the photosynthetic light responses appeared to be an intrinsic property of the vine habit. Although the photosynthetic light response patterns for leaves of the three vines were similar, gas exchange patterns varied markedly between Hayward and the other vines across the growing season. Photosynthetic rates for Hayward were generally higher, rising more rapidly early in the season and then maintaining higher rates through to the end of the season. In both Shiraz and Cabernet Sauvignon shoots, average photosynthetic rates declined during fruit maturation. These patterns of seasonal variation in photosynthesis were tightly coupled with the stomatal conductance which differed in a similar manner between the vines. Consistent with this, from the general linear modelling (Table 4.5), stomatal conductance alone accounted for 20 to 50% of the variance in seasonal photosynthesis and the model coefficient was large, indicative of a close correlation. Furthermore, over 60% of the variance was accounted for when stomatal conductance 99

136 interacted with the internal CO 2 concentration. It was not clear what caused the stomatal conductance to vary across the growing season, although variation in VPD ( kpa) might have contributed. However, in a comparison of 10 common grapevines, Rogiers et al. (2009) demonstrated a 2-fold difference in stomatal conductance between cultivars, including Cabernet Sauvignon and Shiraz vines. Thus, differences in stomatal conductance between the present vines were, most likely, genotypic. Some differences in the seasonal photosynthesis and conductance patterns between Hayward and the two grape cultivars may also have been explained by the difference in canopy structure, duration of leaf development and the pattern of fruiting. Hayward vines were grown on a horizontal pergola system and replacement shoots were on the exterior of the canopy and, therefore, continually exposed to sunlight throughout the day. By contrast, Shiraz and Cabernet Sauvignon vines were grown on a vertical trellis system where shoots received some degree of shading during the day (Smart 1985). At the end of the season, higher photosynthetic rates and stomatal conductance in Hayward leaves than that in leaves of the two grape cultivars may have been due to the stage of fruit development. For example, fruiting sinks are known to increase leaf photosynthetic rates (Greer et al. 2003) and fruit development continued for 30 days longer on Hayward shoots than on grape shoots. Therefore, Hayward shoots had an advantage of maintaining higher photosynthetic rates later in the season Carbon acquisition and balance The maximum rate of carbon acquisition of shoots in this study was highest for the Hayward shoots at 1.92 g C shoot -1 day -1 compared with 0.93 C g shoot -1 day -1 for Shiraz shoots and 0.35 g C shoot -1 day -1 for Cabernet Sauvignon. These rates are well within the range reported previously (Greer and Jeffares 1998; Greer et al. 2003; Greer and Sicard 2009; Greer et al. 2011). The maximum rate of carbon gain of shoots occurred between DAB in Hayward, DAB in Shiraz and DAB in Cabernet Sauvignon. These differences in carbon acquisition reflected differences in both total shoot leaf area (Hayward 0.37 ± 0.04 m 2, Shiraz 100

137 0.24 ± 0.1 m 2 and Cabernet Sauvignon 0.09 ± m 2 ) and the mean photosynthetic rates (Table 3.4). Daily carbon acquisition rates were also reflected in the daily rates of carbon accumulation as biomass in shoots of the respective vines (Figure 3.6), with the highest rates for Hayward 0.76 ± 0.03 g C shoot -1 day -1 compared with Shiraz 0.34 ± 0.02 g C shoot -1 day -1 and Cabernet Sauvignon 0.12 ± 0.01 g C shoot -1 day -1. The net daily carbon balance in shoots of all vines was negative until around 40 DAB and this was similar to reports from earlier studies of Hayward (35-60 DAB) (Greer and Jeffares 1998; Greer et al. 2003) and Semillon grapevines (28-40 DAB) (Greer and Sicard 2009; Greer et al. 2011). The inability of shoots to produce sufficient carbon by photosynthesis during this early shoot development is indicative of reserves being consumed to support the biomass accumulation and shoot growth. Notably, this period was also associated with reduced TNC reserves (shoots are heterotrophic) in oneyear-old wood and roots, and also root death. This dependence on reserves to support early shoot development has been extremely well established in grapevines by Winkler and Williams (1945); Williams (1996); Bennett et al. (2005); Holzapfel and Smith (2012) but not kiwifruit vines. As the leaf area of shoots continued to expand they became progressively autotrophic, as carbon was acquired at a faster rate than it was being sequestered into biomass, resulting in a net carbon balance of the shoot coming into surplus and export occurring (Hale and Weaver 1962; Koblet 1969). For Hayward and Shiraz shoots, the net daily carbon balance remained high throughout fruit development, suggesting all the carbon required for fruit growth was being met and the surpluses were being exported from the shoot. Consistent with this, carbon acquisition rates declined just prior to harvest. However, for Cabernet Sauvignon shoots, both daily carbon acquisition and the net carbon balance peaked during the early stages of berry development, and across the whole season, were at markedly lower rates than the other two vines. These differences were, however, not a result of an intrinsically lower photosynthetic capacity as rates along the shoot and during the growing season were comparable with the Shiraz vines. By contrast, 101

138 Cabernet Sauvignon vines were distinguished by fewer and markedly smaller leaves and an overall lower shoot leaf area and this would have accounted for the reduced carbon acquisition. Similarly, the Cabernet Sauvignon vines had shorter stems, a lower crop load and a markedly reduced allocation of dry matter to the shoots compared with the Shiraz vines and this again reflected the much reduced sequestration of carbon into biomass. Thus, in contrast to the other vines, the Cabernet Sauvignon vines had a marked decline in the net daily shoot carbon balance during the period of reproductive growth. It would appear that the crop load of the Cabernet Sauvignon vines could not be sustained by the supply of carbon but also suggested little to no carbon was exported from shoots on these vines. It is possible that these vines had been previously over cropped and had developed an unbalanced growth pattern (Holzapfel et al. 2010) Whole vine carbon budgets Generally, Hayward kiwifruit vines have a much larger biomass and total carbon content (Smith et al. 1992; Greer et al. 2004) than grapevines (Mullins et al. 1992; Holzapfel et al. 2010). In Hayward vines, total carbon and structural carbon were evenly distributed between the perennial (roots, trunk and perennial wood) and annual biomass (shoots, leaves and fruit) (Smith et al. 1992). In the grapevines, there was a greater percentage (64%) of total and structural carbon in the perennial vine components (Field 2013). However, both grape and kiwifruit vines generally had a larger portion of carbon allocated to roots than trunks or wood. By contrast, in terms of the annual component of vines, almost half of the total and structural carbon was allocated to the fruit. Allocation of non-structural carbon within the vines was more variable than the structural allocations. The portion of non-structural carbon allocated to roots was greater in Shiraz (33.0%) than Hayward (14.1%) and Cabernet Sauvignon (14.5%) vines. This strongly reflects the differences between the vines in the surplus amounts of carbon available for export from the shoots (see Figure 3.10) and root biomass. However, in Cabernet Sauvignon vines a greater portion of non-structural carbon was allocated to the perennial vine 102

139 structure (26.2%) than that in either Shiraz (12.9%) or Hayward (17.7%) vines. This provides further evidence that Cabernet Sauvignon vines had been previously over cropped and were having to utilise root reserves primarily to support fruit growth (Candolfi-Vasconcelos et al. 1994; Holzapfel et al. 2010). Consistent with this conclusion, in Shiraz vines (32.7%) and Hayward vines (30.5%) a greater portion of non-structural carbon was allocated to fruit than in the Cabernet Sauvignon vines (21.7%). However, the total non-structural carbon in perennial organs at harvest in the Shiraz (157 g per vine) and Cabernet Sauvignon (129 g per vine) vines was less than previously reported by Holzapfel et al. (2010). Furthermore, these estimates of total non-structural carbon are markedly lower than the 391 g per vine for Hayward on an equivalent basis (roots, trunk and cordon) or 488 g per vine if one-year-old canes are included. However, on a percentage basis, total non-structural carbon comprised 7% of total carbon in perennial organs of Shiraz and 5.8% in Cabernet Sauvignon vines compared with only 3.4% in Hayward vines. These differences clearly emphasises the higher carbohydrate concentrations of grapevines relative to their biomass compared with the lower concentrations for the larger, more vigorous kiwifruit vines. This difference may reflect the longer period of domestication of grapevines, when vines have been selected for fruiting performance rather than vegetative vigour Summary To examine the hypothesis that seasonal patterns of development of grape and kiwifruit vines was influenced in a similar manner by carbohydrate supply, vine phenology, shoot growth, carbon acquisition, carbohydrate pools and carbon budgets of vines were compared during key stages of development. The development of vines was remarkably similar. During the first 30 DAB of shoot development, growth of new shoots was rapid and carbohydrate concentrations in annual organs were high; however, the net carbon balance of shoots was negative as photosynthetic rates were insufficient to support new growth and floral development. At this time starch concentrations in perennial wood and roots decreased rapidly and root death occurred, indicating that early shoot growth and floral development were 103

140 largely supported by import from perennial reserves. From around anthesis (50-70 DAB) through to the early stages of fruit growth (100 DAB), leaf area reached a maximum, photosynthetic rates peaked, shoot vegetative growth and biomass accumulation was rapid and carbohydrate concentrations in shoots generally decreased. From around 100 DAB the balance of growth shifted, from rapid shoot growth to increased fruit and root growth, with accumulation of carbohydrate in fruit and as reserves in internodes, wood and roots, as photosynthetic rates generally remained high. After fruit harvest although photosynthetic rates declined, carbohydrate continued to be stored in shoot internodes and perennial organs through until leaf fall. Therefore the growth of the vines could be divided into two stages, DAB when rapid shoot growth occurred and vine reserves declined, and DAB when fruit growth, root growth and carbohydrate accumulation occurred. Although growth patterns were similar between vines there were quantitative differences with, for example, the amount of budbreak being greater in Shiraz and Cabernet Sauvignon vines while the growth of shoots and the periods of root growth were greater in Hayward vines. These differences influenced carbohydrate acquisition, which was greater in Hayward shoots because of the larger leaf area. Partitioning also differed, with a greater portion of carbon in the annual growth of Hayward vines than in Shiraz and Cabernet Sauvignon vines. The TNC concentration in key organs was higher in Shiraz and Cabernet Sauvignon vines than those in Hayward vines. The general similarity in the pattern of growth and changes in carbohydrate acquisition and accumulation of the vines suggests that variation in carbohydrate supply at key stages of vine development may have similar effects on their performance. However given the greater growth and lower concentrations of TNC in Hayward vines, it is likely that altering carbohydrate supply may have a greater effect on the development of these vines and this is explored in later chapters. 104

141 4 Carbohydrate supply during inflorescence initiation and bud development 4.1 Introduction Inflorescence initiation, the first stage of inflorescence development, begins in developing grape and kiwifruit axillary buds around the time of anthesis of the current season s crop, 12 months before the newly initiated floral meristems become fully developed inflorescences. The timing of inflorescence initiation in both grape and kiwifruit buds has been determined from the appearance of floral primordia (Lavee et al. 1967; Walton et al. 1997) and the expression of homologues of meristem identity genes (Walton et al. 2001; Boss et al. 2003; Carmona et al. 2008). As shoots grow during spring and early summer, vegetative and floral development is concurrent, with developing buds undergoing inflorescence initiation sequentially along the shoot (Lavee et al. 1967; Snowball 1995; May 2000), and therefore the process of inflorescence initiation occurs over several weeks (Walton et al. 1997; Bennett et al. 2005). Once floral primordia are initiated in grape buds, they divide into two arms and may undergo a limited amount of branching before becoming quiescent until the following spring (Srinivasan and Mullins 1981; Boss et al. 2003). In contrast, kiwifruit floral primordia undergo no further development before becoming quiescent until the following season (Walton et al. 1997; Walton et al. 2001). Sugars have several crucial roles in floral initiation including signalling, phloem loading, cell division and supplying energy for growth (Bernier and Perilleux 2005). Previous studies of inflorescence initiation in both grape and kiwifruit have involved leaf removal (Swanepoel and Archer 1988; Snelgar and Manson 1992; Snowball 1996). However, it is not clear whether the effects of leaf removal are due to the modification of flowering time signals from leaves (Lavee et al. 1967; Candolfi-Vasconcelos and Koblet 1990; Snelgar and Manson 1992; Snowball 1996) or altering the supply of sugars to support bud development (Hopping 1977; Candolfi-Vasconcelos and Koblet 1990; Watson and Gould 1994; Snowball 1996; Sommer et al. 2000; Lebon et al. 2004b; Bennett et al. 2005). Modifying carbohydrate supply to 105

142 developing buds during inflorescence initiation, by shading, girdling or root pruning of vines, influences subsequent inflorescence production by grape (Candolfi-Vasconcelos and Koblet 1990; McArtney and Ferree 1999a; Bennett et al. 2005) and kiwifruit buds (Fabbri et al. 1991; Snelgar and Manson 1992; Snowball 1996). However, in these studies it is difficult to separate the effects of treatments on inflorescence initiation, and shoot and bud development in the current season, from effects from the development of flowers occurring immediately prior to anthesis in the following season. Therefore, in this chapter carbohydrate supply was manipulated during early bud development to determine how carbohydrate supply influenced inflorescence initiation during the current season s shoot growth and hence, bud floral performance in the subsequent spring. The applied treatments involved a combination of girdling (to restrict shoot import/export via phloem) and reducing photosynthate supply through either defoliation (reduced leaf area) or shading (reduced photosynthesis rate). As these treatments may also affect floral signals, therefore, for all treatments it will be assumed that the changes in bud performance may be related to both the measured carbohydrate changes and also to putative floral signals, the measurement of which is beyond the scope of the thesis. 106

143 4.2 Materials and methods Plant material This experiment was carried out during the 2009/2010 and 2010/2011 growing seasons. At the beginning of the experiment Hayward (unknown clone) vines were two-years-old and growing on their own roots and nineyears-old Cabernet Sauvignon (unknown clone) vines were grown on SO4 rootstocks. No Shiraz vines were available for this experiment. Both vines were grown in 30 litre pots (see Chapter 2) (Figure 4.1), rather than in a commercial vineyard/orchard, so that treated canes could be easily maintained over consecutive experimental seasons. Briefly, vines were spaced 1 m apart within rows with 3 m between rows. Hayward vines were grown on a mini T-Bar system with a short (0.5 m) leader and four one-yearold canes, each approximately 1.1 m long with 18 buds. Cabernet Sauvignon vines were grown on a VSP system with two canes, each 1.5 m long with 15 buds. Vines were managed according to standard practices including drip irrigation and regular additions of a compound, slow-release fertiliser (see Chapter 2). Figure 4.1 Pot grown vines were used in this experiment with Hayward vines in the foreground (row one) and Cabernet Sauvignon vines in rows three - five) Experimental design To determine effects of carbohydrate supply on inflorescence initiation and early bud development, a blocked split plot experimental design was used. 107

144 Thirty-three vines of each vine cultivar were used, split into 11 replicate blocks each of three vines. The stage of development treatment was randomly allocated to one vine (plot) in each block. Shoot manipulation treatments were then randomly allocated to four equivalent shoots (subplots), suitable as replacement canes, within each vine (plot). Each treatment was replicated 11 times. The experiment began in October when inflorescence initiation is thought to begin when 7-8 leaves (E L 14-15) had opened on Cabernet Sauvignon shoots (Pratt 1971; Swanepoel and Archer 1988; Watt et al. 2008) and leaves (E L 16) had opened on Hayward shoots, (Walton et al. 2001) and treatments were completed in February, when the production of new buds had ceased (Coombe 1995; Snowball 1995) (Table 4.1). The trial was repeated over two growing seasons and new material was used in each season. There were three stages of development when shoot manipulation treatments were applied: 1. Early: from early shoot development until anthesis (E L 14-19) when shoots were growing rapidly, producing new buds and autotrophic. 2. Mid: from anthesis until rapid berry growth (E L 19-31) when shoots were growing rapidly, producing new buds, maturing older buds and exporting carbohydrate. 3. Late: from the beginning of rapid berry growth until to the beginning of fruit seed maturation (veraison for Cabernet Sauvignon and seed colouring in Hayward ) (E L 31-35) when shoot growth and bud development had slowed, and shoots were exporting and storing carbohydrate. There were four treatments, including three treatments to manipulate shoot carbohydrate supply and an untreated control. Treatments included: 1. Untreated control: shoots were intact and exposed to full sun. 2. Girdled: shoots were phloem girdled to prevent carbohydrate import/export and exposed to full sun (Figure 4.2A). 108

145 3. Girdled and defoliated: shoots were phloem girdled to prevent carbohydrate import/export, all leaves except the most basal four leaves of a shoot were removed and the shoot was exposed to full sun. This treatment limited production of floral signal and photosynthate by reducing leaf area (Figure 4.2C). 4. Girdled and shaded: shoots were phloem girdled to prevent carbohydrate import/export and shaded to reduce carbohydrate production through reduced photosynthesis rates (Figure 4.2B). Each treatment period lasted for approximately 38 days and at the end of each period, the girdles had healed and the shade cloth was removed from shoots. Shoots were girdled mm from their base using a sharp pair of sidecutter pliers to carefully sever the phloem, leaving a 1 mm wide girdle. The girdled section was supported by taping a thin splint of bamboo to the shoot (Figure 4.2A). At the end of each treatment period, the girdle had healed and the splints were removed. All leaves, except the four most basal leaves, were removed from defoliated shoots and any developing leaves/shoots were removed during the treatment period (Figure 4.2C). Very few leaves developed on defoliated shoots after the treatment period. The four basal leaves were retained to keep shoots alive (Greer and Sicard 2009). Shoots were shaded by enclosing them in bags made from shade cloth (Solarshade TM, Permathene, New Zealand) (Figure 4.2B). Calibration measurements showed that the shade cloth reduced average photosynthetically active radiation (PAR) levels by 85% (and PFD to 83 cf. 991 µmol m -2 ) and gave a slight (0.7%, 0.3 o C) increase in mean temperature (see Appendix 4). 109

146 110 Table 4.1: Calendar date and days after budbreak (DAB) of each stage of development (E-L stage) that shoot treatments were applied to Cabernet Sauvignon and Hayward shoots, during the 2009/2010 and 2010/2011 growing seasons. Stage of Cabernet Sauvignon Hayward development (E-L stage) Date DAB Untreated shoot node number Date DAB Untreated shoot node number 2009/2010 Early (14-19) 23/10/09-30/11/ /10/09-27/11/ Mid (19-31) 30/11/09-5/1/ /11/09-4/1/ Late (31-35) 5/1/10-12/2/ /1/10-12/2/ /2011 Early (14-19) 28/10/10-30/11/ /10/10-22/11/ Mid (19-31) 30/11/10-11/1/ /11/10-3/1/ Late (31-35) 11/1/11-19/2/ /1/11-11/2/

147 A B C C Figure 4.2: Girdling treatments applied to shoots. Examples A, a girdled Cabernet Sauvignon shoot exposed to full sun, B, a girdled and shaded Cabernet Sauvignon shoot and C, a girdled and defoliated Hayward shoot Shoot measurements In this study, shoots were defined as the vegetative growth from buds that occurred in the current season and had a green and photosynthetically active stem. Once the photosynthetic tissues degenerated and the stem changed colour from green to brown it was defined as a cane. After winter dormancy, these shoots became one-year-old canes. All shoots were measured at the beginning of the experiment and at the end of each stage of development. Measurements included shoot basal diameter, shoot length and node number. In the winter of 2009, three replicate canes of each treatment were removed at winter pruning and sampled for carbohydrate analysis. At the end of each growing season, seven replicate Cabernet Sauvignon canes 111

148 (from each treatment combination) were retained for measurements in spring. All Hayward replicate canes were retained, except the three replicate canes of each treatment required for winter carbohydrate samples in 2009, giving a maximum of eight replicate Hayward canes from each 2009/2010 treatment combination and 11 replicate Hayward canes (treated shoots) from each 2010/2011 treatment combination Winter cane composition During winter at the end of the 2009/2010 season, three replicate canes per treatment combination were used to provide bud and cane samples for carbohydrate analysis. Six buds were excised from each cane, weighed, and immediately frozen in liquid nitrogen and stored at -80 o C for carbohydrate analysis (see Chapter 2). In addition, the length, diameter and fresh weight of the internode immediately below each sampled bud was measured before a subsample of internode was frozen in liquid nitrogen and stored at -80 o C for carbohydrate analysis. The remainder of the internode was measured, weighed and the dry weight determined after drying at 65 o C for 24 hours or until samples reached a stable weight. Sampling was not repeated in 2010/2011 season in order to retain the maximum number of replicates for evaluation in spring Phenology During spring in the season following shoot treatment, the time of budbreak was monitored three times per week on treated canes as described in Chapter 2. Prior to anthesis, individual buds on treated canes were visually assessed to determine if they had produced a shoot and if so, how many inflorescences were present on that shoot. Hayward buds produced very few lateral flowers, so these were not counted. From these data, budbreak (%), floral budbreak (%), inflorescence number per floral shoot and inflorescence number per winter bud were calculated (see Chapter 2). Inflorescence numbers per bud (IWB), budbreak (BB), floral budbreak (FS), and inflorescences per shoot (INF) in spring 2010 were each modelled as a function of bud and internode dry weight and carbohydrate content in the 112

149 previous winter, for the two vines. A general linear model (GLM) was used to model the data using the equation. IWB = f (bud carbohydrate + internode carbohydrate) Statistical analysis All data were analysed using GenStat 14 th edition for Windows (VSN International Ltd., Hemel Hempstead, Hertfordshire, United Kingdom) using REML linear mixed models assuming a fully randomised split plot design to analyse treatment differences in shoot growth, cane composition, budbreak and flowering data. Differences between treatment means were determined at the appropriate level of significance, and estimated means and standard errors were presented. Relationships between carbohydrate content, dry weight and bud performance were examined using the GLM procedure. All data was plotted using Origin version 8.5 (Originlab Corporation, Northhampton, Massachusetts, USA) and curves were fitted to budbreak and shoot growth data across time using the Boltzmann function (Seleznyova and Greer 2001). See Chapter 2 for further details. 113

150 4.4 Results Shoot growth Shoot growth of both vines was much greater during the 2010/2011 season than in the 2009/2010 season, resulting in an 80% increase in the final length of shoots, a 40% increase in node number and a 25% increase in shoot basal diameter. In addition, untreated Cabernet Sauvignon shoots grew almost twice as long (P < 0.001) (2.33 ± 0.12 m in 2009/2010; 4.02 ± 0.23 m in 2010/2011) as untreated Hayward shoots (1.34 ± 0.11 m in 2009/2010; 2.46 ± 0.21 m in 2010/2011) in both experimental seasons (Figure 4.3). The number of nodes that developed on shoots followed the same trends, so that Cabernet Sauvignon shoots produced significantly (P < 0.001) more nodes per shoot (31.7 ± 1.5 nodes in 2009/2010; 42.6 ± 2.6 nodes in 2010/2011), than Hayward shoots (22.3 ± 1.4 nodes in 2009/2010; 32.9 ± 2.6 nodes in 2010/2011). However, untreated Hayward shoots had a significantly (P < 0.001) greater shoot basal diameter (11.06 ± 0.32 mm in 2009/2010; ± 0.33 mm in 2010/2011) than Cabernet Sauvignon shoots (9.62 ± 0.28 mm in 2009/2010; ± 0.31 mm in 2010/2011). Shoots of the two vines followed different patterns of extension growth (Figure 4.3). Rapid growth of Cabernet Sauvignon shoots occurred during both the early and mid stages of the shoot development treatment, with 40-50% of final shoot length produced during each stage, and this pattern was consistent between seasons. By contrast, a large but variable portion of Hayward shoot growth (87% in 2009/2010 and 60% in 2010/2011) occurred in the early treatment stage with slower growth during the remaining two stages. This difference between the two vines was also reflected in the dynamics of shoot growth with a significant (P < 0.001) difference of about 20 days between the middle point of growth (50% of development completed) of Cabernet Sauvignon shoots (54.3 ± 0.9 DAB in 2009/2010; 67.6 ± 2.2 DAB in 2010/2011) and Hayward shoots (36.3 ± 1.0 in 2009/2010; 43.0 ± 2.1 DAB in 2010/2011) (Tables 4.2 and 4.3). In 2009/2010, the duration of shoot growth (days between 20 and 80% of growth) was also significantly (P < 0.001) longer for Cabernet Sauvignon shoots (47.4 ± 1.3 days) compared to Hayward shoots (39.1 ± 1.3 days). 114

151 There were significant vine cultivar*shoot and treatment*stage of development interactions on shoot growth (Figure 4.3) and on the dynamics of growth (Tables 4.2 and 4.3). The growth of Cabernet Sauvignon shoots that were girdled and defoliated either at the early or mid-stages of development when rapid growth was occurring, was significantly (P < ) slower and shoot length at the end of the stage was only 70% of untreated shoots (Figures 4.3A - F) in both experimental seasons. In addition, the growth of Cabernet Sauvignon shoots that were girdled and shaded during early development was also significantly (P < 0.05) less and the length of shoots at the end of the stage was only 72% (2009/2010) or 33% (2010/2011) of that of untreated shoots. The effects of these shoot treatments persisted throughout the remainder of shoot extension growth, and were consistent over the two experimental seasons. In 2009/2010, the mid-point of shoot growth was significantly (P < 0.001) 15.6 days earlier for Cabernet Sauvignon shoots that were girdled and shaded at the early and mid development stages as well as 11.4 days (P < 0.05) earlier for the mid development girdle and defoliation treatment than the untreated shoots (Table 4.2). In addition, the duration of shoot growth was significantly (P < 0.05) 15 days shorter on shoots girdled and shaded during the mid development treatment stage than on untreated shoots. In 2010/2011, only the vine cultivar*shoot treatment interaction was significant (P < 0.01) for shoot growth dynamics (Table 4.3). The mid-point of shoot growth was significantly (P < 0.05) earlier (57.8 ± 4.4 DAB) and the duration of growth shorter (47.3 ± 4.9 days) on Cabernet Sauvignon shoots that had been girdled and defoliated (averaged over stage of development) compared to untreated shoots (72.3 ± 4.4 DAB; 62.7 ± 5.0 days). By contrast, the pattern of growth of the girdled Cabernet Sauvignon shoots was similar to that of the untreated shoots in both seasons. Shoot treatments did not have consistent effects on the growth of Hayward shoots over the two experimental seasons (Figures 4.3G - L). In 2009/2010, Hayward shoots that were girdled and defoliated during early development grew only 74% (P < 0.05) as much as untreated shoots during that period 115

152 (Figure 4.3G). By contrast, the growth of shoots that were girdled and shaded, during either the early or mid development treatment, increased significantly (P < 0.05), to 144% and 121% of that of untreated shoots, during the treatment period (Figures 4.3G and H). However, effects of shoot treatments were not maintained and there were no significant differences in shoot length at the end of the 2009/2010 season. By comparison, in 2010/2011 the growth of shoots that were either girdled or girdled and defoliated during rapid early development was significantly (P < 0.05) less (50%) than that of untreated or girdled and shaded shoots, and these effects were maintained until the end of shoot development (Figures 4.3J - L). With respect to the dynamics of Hayward shoot growth, in 2009/2010 the mid-point of shoot growth on shoots that were girdled during early or late development was significantly (P < 0.05) 9.6 days later than that on untreated shoots (Table 4.2). The early girdle and shade treatment significantly (P < 0.05) reduced the duration of shoot growth compared to shoots that were either just girdled (15.3 days) or girdled and defoliated (13 days) during early development and also to shoots that were girdled and shaded (16.1 days) during the mid development treatment. Overall, the duration of growth on shoots that were just girdled was 17.2 days less when compared with that of untreated shoots (Table 4.3). In contrast to the 2009/2010 season, the growth of girdled and shaded shoots and untreated shoots was similar in 2010/2011 except that the mid-point of shoot growth was significantly (P < 0.05) 14.9 days later on girdled and shaded shoots than on untreated shoots. 116

153 117 Figure 4.3: Growth of shoots at each stage of development for Cabernet Sauvignon A - C, 2009/2010 and D - F, 2010/2011; Hayward G - I, 2009/2010 and J - L, 2010/2011. Shoot treatments include untreated, girdled, girdled and defoliated and girdled and shaded. The lines are the fit of data to the Boltzmann sigmoid function. Data are means ± SEM, n = 11. Note the differences in scale between Cabernet Sauvignon and Hayward shoots. For Cabernet Sauvignon 50% anthesis occurred DAB and veraison DAB, and for Hayward 50% anthesis occurred DAB and seed maturation DAB. 117

154 118 Table 4.2: Effects of treatments applied during different stages of inflorescence initiation in the 2009/2010 season on the mid-point (50%) and duration of growth of Cabernet Sauvignon and Hayward shoots calculated from fitting the Boltzmann sigmoid function. Data are means ± SEM, n = 11. Shoot treatment Stage of development Mid-point of shoot growth (DAB) Duration of shoot growth (days from 20-80% growth) Cabernet Hayward Cabernet Hayward Sauvignon Sauvignon Untreated 58.3 ± ± ± ± 4.8 Girdled early 53.7 ± ± ± ± 4.6 Girdled/defoliated early 50.7 ± ± ± ± 4.6 Girdled/shaded early 42.7 ± ± ± ± 4.4 Girdled mid 64.2 ± ± ± ± 4.6 Girdled/defoliated mid 46.9 ± ± ± ± 4.4 Girdled/shaded mid 42.7 ± ± ± ± 4.4 Girdled late 58.3 ± ± ± ± 4.6 Girdled/defoliated late 60.5 ± ± ± ± 4.6 Girdled/shaded late 57.0 ± ± ± ± 4.4 P value Vine <0.001 <0.001 Treatment < Stage Vine*treatment Vine*stage Treatment*stage Vine* treatment*stage

155 119 Table 4.3: The effects of shoot treatments applied during different stages of inflorescence initiation development in the 2010/2011 season on the mid-point and duration of growth of Cabernet Sauvignon and Hayward shoots calculated from fitting the Boltzmann sigmoid function. Data are means ± SEM, n = 11. Shoot treatment Stage of development Mid-point of shoot growth (DAB) Duration of shoot growth (days from 20-80% growth) Cabernet Hayward Cabernet Hayward Sauvignon Sauvignon Untreated 72.3 ± ± ± ± 8.2 Girdled early 69.3 ± ± ± ± 8.4 Girdled/defoliated early 57.6 ± ± ± ± 8.4 Girdled/shaded early 69.5 ± ± ± ± 9.4 Girdled mid 85.8 ± ± ± ± 8.4 Girdled/defoliated mid 49.8 ± ± ± ± 8.1 Girdled/shaded mid 54.7 ± ± ± ± 8.1 Girdled late 75.9 ± ± ± ± 8.1 Girdled/defoliated late 66.1 ± ± ± ± 8.4 Girdled/shaded late 57.0 ± ± ± ± 8.1 P value Vine < Treatment Stage Vine*treatment < Vine*stage Treatment*stage Vine* treatment*stage

156 4.4.2 Axillary bud and internode dry weight in winter In the winter of 2010, after the shoot treatments had been imposed in the 2009/2010 season, the dry weight of dormant axillary buds and their associated internodes differed significantly between the two vines (Table 4.4). Untreated Cabernet Sauvignon buds were significantly (P < 0.05) 139% heavier and internodes (P < 0.001) 185% heavier than those from untreated Hayward shoots. For bud dry weight, there was a highly significant effect of interaction between vine and stage of shoot treatment (Table 4.4) and this is shown in Figure 4.4A. The mean weight of Cabernet Sauvignon buds from the early stage of treatment (across all shoot treatments) was significantly (P < 0.001) and markedly reduced by over 40%, but there was no significant difference in the average weight of treated and untreated buds at the two later stages of development. By contrast for Hayward canes, the mean bud weight (across all shoot treatments) from shoots treated in late development were significantly (P < 0.05) 26% heavier than that of untreated buds, but there was no effect at the other stages of development. In addition, the shoot treatment*stage interaction was also highly significant (P < 0.001). Pooled across the two vines, there was an overall trend for the dry weight of dormant axillary buds to increase by 30% between the early and late shoot treatment applications, although effects varied with shoot treatment (Figure 4.4B). For instance, the dry weights of buds from shoots that had either been girdled and defoliated (20.0 ± 3.7 mg) or girdled and shaded (22.1 ± 3.3 mg) during early shoot development, were significantly (P < 0.05) reduced to only 70% of the weight of untreated buds (30.5 ± 3.2 mg) (Figure 4.4B). Furthermore, the buds from shoots girdled and defoliated in mid development were significantly (P < 0.05) heavier (38.7 ± 3.2 mg) than untreated buds. It is notable that the dry weight of buds from shoots that were girdled or girdled plus shaded both increased consistently from early to late development, whereas the weight of buds from girdled and defoliated shoots appeared to increase rapidly from the early to mid development, but declined thereafter. 120

157 Figure 4.4: A, Effects of vine cultivar (averaged over shoot girdling treatments) and stage of shoot development treatment on the dry weight of dormant buds compared with dry weight of untreated buds in winter Vines are Cabernet Sauvignon and Hayward. B, Effects of shoot girdling treatments (averaged over vine cultivar) and stage of development treatment on the dry weight of dormant buds compared with the dry weight of untreated buds in winter Shoot treatments include untreated, girdled, girdled and defoliated and girdled and shaded. Data are means ± SEM, n = 6-9. For internode dry weights, there were no significant interactions between the vine cultivar, shoot treatment or stage of development (Table 4.4). The internodes from girdled and defoliated canes (averaged across vines and stage of development) was markedly lighter (1.50 ± 0.26 g), by about 40%, than those from shoots that were just girdled (2.77 ± 0.25 g) (P < 0.001) or untreated (2.31 ± 0.29 g) (P < 0.05) (Figure 4.5A). In addition, the internodes of canes that had been girdled and shaded (2.06 ± 0.26 g) were significantly (P < 0.05) 36% lighter than those from shoots that were just girdled. The overall effect of stage of development (averaged across vine cultivar and shoot treatment) on internode dry weights was only weakly significant (P = 0.07), with internode dry weights of shoots treated at the early (1.92 ± 0.24 g) 121

158 and mid stage (1.98 ± 0.25 g) of development similar but significantly (P < 0.05) less than internodes from shoots treated during late shoot development (2.58 ± 0.22 g) (Figure 4.5B). Figure 4.5: A, Effects of shoot girdling treatments (averaged over vine cultivar and stage of development) on the dry weight of shoot internodes compared with untreated shoot internodes in winter B, effects of stage of shoot development treatment (averaged over vine cultivar and shoot girdling treatment) on the dry weight of internodes compared with the dry weight of untreated internodes in winter Data are means ± SEM, n =

159 123 Table 4.4: Effect of shoot girdling and stage of development treatments in 2009/2010 season on the dry weight of Cabernet Sauvignon and Hayward buds and internodes compared with that in untreated shoots during winter Data are means ± SEM, n = 3. Shoot treatment Stage of Bud dry weight (mg) Internode dry weight (g) development Cabernet Sauvignon Hayward Cabernet Sauvignon Hayward Untreated 35.5 ± ± ± ± 0.25 Girdled early 28.6 ± ± ± ± 0.95 Girdled/defoliated early 14.8 ± ± ± ± 0.21 Girdled/shaded early 17.7 ± ± ± ± 0.37 Girdled mid 41.4 ± ± ± ± 0.50 Girdled/defoliated mid 41.1 ± ± ± ± 0.34 Girdled/shaded mid 28.2 ± ± ± ± 0.48 Girdled late 33.2 ± ± ± ± 1.18 Girdled/defoliated late 32.8 ± ± ± ± 0.44 Girdled/shaded late 37.2 ± ± ± ± 0.40 P value Vine <0.001 Treatment Stage < Vine*treatment Vine*stage Treatment*stage Vine* treatment*stage

160 4.4.3 Winter bud and internode carbohydrate content In the winter of 2010, the TNC concentration in Cabernet Sauvignon buds (141.1 ± 2.8 mg gdw -1 ) was four-fold higher than that in Hayward buds (32.9 ± 3.3 mg gdw -1 ); however, there were no significant effects of shoot treatments in the previous season on the TNC concentration of buds. The TNC concentration in Cabernet Sauvignon internodes (140.6 ± 2.2 mg gdw - 1 ) was also 12% higher than that in Hayward internodes (125.5 ± 2.5 mg gdw -1 ) (Table 4.5). However, the effects of both the vine cultivar*treatment and vine cultivar*stage of development interactions on internode carbohydrate concentration were weakly significant (P = and P = respectively) due to effects on Hayward canes (Table 4.5). The internodes of Hayward canes that were girdled and defoliated (averaged across stage of development) had significantly (P < 0.05) higher TNC concentrations (140.8 ± 4.7 mg gdw -1 ) than internodes from shoots that were girdled only (117.2 ± 4.9 mg gdw -1 ) or from girdled and shaded shoots (118.0 ± 4.9 mg gdw -1 ). In addition, the TNC concentrations of internodes from Hayward canes treated during the early stage of development were significantly higher (P < 0.05) (133.0 ± 4.2 mg gdw -1 ) than that in canes treated during the late stage of development (115.8 ± 4.1 mg gdw -1 ). The TNC content of buds and their internodes also varied between the two vines, with five-fold more carbohydrate in the buds (5.0 ± 0.5 mg cf. 0.9 ± 0.5 mg) and more than twice as much in the internodes (413 ± 73 mg cf. 181 ± 82 mg) of Cabernet Sauvignon shoots as Hayward shoots (Table 4.5). Only buds from Cabernet Sauvignon canes that had been either girdled and defoliated or girdled and shaded during the earliest stage of development had significantly (P < 0.05) about 50% less carbohydrate compared with buds from untreated canes (Figure 4.6A). There were no significant effects of the treatments on the carbohydrate content of Hayward buds (Figure 4.6B). The comparison between the two vines, however, changed when the carbohydrate content of the internodes was considered (Figures 4.6C and D). Effects of shoot treatment (averaged across all stages of development) 124

161 were significant (P < 0.01) with a 35% decrease in the TNC content of internodes from girdled and defoliated Cabernet Sauvignon canes (265 ± 46 mg) and a 13% decrease in girdled and defoliated Hayward canes (158 ± 55 mg) compared with that in untreated internodes. There was a similar 33% decrease in the TNC content of internodes from girdled and shaded Cabernet Sauvignon internodes (267 ± 43 mg) (P < 0.05) compared with that in untreated canes. 125

162 126 Table 4.5: Effect of shoot girdling and stage of development treatments in 2009/2010 on the carbohydrate content of Cabernet Sauvignon and Hayward buds and internodes during the winter of Data are means ± SEM, n = 3. Shoot treatment Stage of development Bud carbohydrate content (mg / bud) Internode carbohydrate content (mg / internode) Cabernet Sauvignon Hayward Cabernet Sauvignon Hayward Untreated 5.02 ± ± ± ± 26 Girdled early 4.10 ± ± ± ± 120 Girdled/defoliated early 2.09 ± ± ± ± 22 Girdled/shaded early 2.89 ± ± ± ± 46 Girdled mid 5.21 ± ± ± ± 54 Girdled/defoliated mid 4.69 ± ± ± ±46 Girdled/shaded mid 4.13 ± ± ± ± 60 Girdled late 4.30 ± ± ± ± 85 Girdled/defoliated late 4.72 ± ± ± ± 67 Girdled/shaded late 5.28 ± ± ± ± 41 P value Vine <0.001 <0.001 Treatment Stage Vine*treatment Vine*stage Treatment*stage Vine* treatment*stage

163 Figure 4.6: Effects of treatments applied at different stages of development in 2009/2010 on the carbohydrate content of A, Cabernet Sauvignon buds and C, adjacent internodes (untreated, girdled, girdled and defoliated and girdled and shaded ); B, Hayward buds and D, adjacent internodes (untreated, girdled, girdled and defoliated and girdled and shaded ) in winter Data are means ± SEM, n = 3. Note the differences in scale between bud and internode carbohydrate content Effects on spring development Timing and duration of budbreak In spring, budbreak of untreated Cabernet Sauvignon and Hayward canes began at approximately the same date but the Hayward canes reached the mid-point of budbreak earlier than the Cabernet Sauvignon canes in both seasons (21.4 ± 0.5 cf ± 0.4 days after 1 September (DAS) 2010 and 38.6 ± 1.6 cf ± 0.7 DAS 2011) (Figure 4.7). There was, therefore, a significant (P < 0.001) difference in both the timing of budbreak as well as the duration of budbreak (days between 20 and 80% budbreak; 8.0 ± 0.4 days in 2010 cf ± 1.1 days in 2011) between the two seasons for both vines. The stage of development when treatments were applied in the previous season had no significant effect on the subsequent time of budbreak, therefore, shoot treatment means are shown in Figure 4.7. The only 127

164 significant (P < 0.001) effect over the two seasons was an advance in budbreak of girdled and defoliated Cabernet Sauvignon canes (middle point 43.4 ± 0.4 DAS) compared with untreated canes (51.4 ± 0.7 DAS) in 2011 (Figure 4.7B). However, this may have been an artefact of the low number of buds that broke on this treatment (6.6 ± 0.6 buds per cane cf. 8.6 ± 0.6 on untreated canes) with no buds breaking in the later stages. There was no effect of vine cultivar or shoot treatment on the duration of budbreak. Figure 4.7: Time of budbreak in spring for Cabernet Sauvignon (A 2010 and C 2011, untreated, girdled, girdled and defoliated, girdled and shaded and Hayward (B 2010 and D 2011 untreated, girdled, girdled and defoliated, girdled and shaded canes. Budbreak percentage was calculated from the cumulative number of buds that broke at each date and final budbreak of each cane. Data are means ± SEM, n = for each shoot girdling treatment (averaged across stage of development treatments) compared with untreated shoots. The lines are the fit of data to the Boltzmann sigmoid function Amount of budbreak and number of inflorescences produced Seasonal effects on budbreak and inflorescence production were similar for both vines, with significantly (P < 0.001) higher bud productivity in spring 2010 than spring 2011 (Tables 4.6 and 4.7). In 2010, for example, budbreak 128

165 was significantly (P < 0.001) higher in both Cabernet Sauvignon (1.5-fold) and Hayward (1.9-fold) canes and floral budbreak was also significantly (P < 0.05) 1.7-fold higher for both vines in 2010 compared to However, there was no significant seasonal effect on the number of inflorescences produced per floral shoot. The net result of these changes was that the number of inflorescences produced per bud was significantly (P < 0.001) 1.9- fold higher on both Cabernet Sauvignon and Hayward canes in 2010 than in Although the seasonal effects on bud performance were similar for both vines, Cabernet Sauvignon buds consistently produced twice the number of inflorescences (0.94 ± 0.05 in 2010; 0.49 ± 0.05 in 2011) compared with Hayward buds (0.47 ± 0.04 in 2010; 0.25 ± 0.03 in 2011) (Tables 4.6 and 4.7). This difference between vines was due to significantly (P < 0.001) higher budbreak (90.3 ± 5.1 cf ± 5.1 in 2010; 59.2 ± 5.8 cf ± 4.1 in 2011) and a three-fold increase in floral budbreak (71.8 ± 6.9 cf ± 6.4 in 2010; 41.4 ± 5.8 cf ± 4.1 in 2011) on Cabernet Sauvignon than Hayward canes. There were no effects of treatment on the total budbreak of either Cabernet Sauvignon or Hayward canes in either season. Effect of shoot treatments and their stage of application on floral productivity (inflorescences per bud, floral budbreak and inflorescence per shoot) varied between seasons (Tables 4.6 and 4.7). In 2010, the vine cultivar*stage interaction was significant for all three measures of floral productivity (Figure 4.8). For Cabernet Sauvignon vines, the number of inflorescences produced per bud from canes previously treated during mid shoot development (averaged across shoot treatments) (0.76 ± 0.12) was significantly (P < 0.05) and 30% reduced compared with buds from late treated shoots (1.07 ± 0.12), largely due to a significantly (P < 0.05) lower (22%) floral budbreak (56.8 ± 4.2% cf ± 4.6%) (Figures 4.8A and B). By contrast, inflorescence number per Hayward bud was significantly (P < 0.05) 54% lower on early treated canes (0.27 ± 0.09) than mid and late treated canes (0.59 ± 0.10), due to significantly (P < 0.05) lower (49%) floral budbreak (11.6 ± 3.3) than mid 129

166 treated canes (22.5 ± 4.0) and significantly (P < 0.01) fewer (45%) inflorescences per floral shoot (1.03 ± 0.22 cf ± 0.26 mid and 1.83 ± 0.23 late treated canes) (Figures 4.8A - C ). For both vines, the floral productivity of the other two stages of development was similar to that of previously untreated canes. In 2011, the main effect of shoot treatment (averaged across vine and stage of development) and the main effect of stage of development (averaged across vine and shoot treatment) on floral productivity were highly significant (Table 4.7). There was a significant (P < 0.01) reduction in the number of inflorescences per bud on previously girdled and defoliated canes to 59% (0.25 ± 0.04) compared with that on untreated canes (0.42 ± 0.05) and girdled canes (0.48 ± 0.05). This was the result of significantly (P < 0.01) lower floral budbreak (18.7 ± 1.9% cf ± 2.2% and 27.8 ± 2.0%) and number of inflorescence per floral shoot (P < 0.05) (0.95 ± 0.14 cf.1.45 ± 0.16 and 1.48 ± 0.15) (Figures 4.9A, C and E). In terms of the stage of development that shoot treatments were applied in the previous season, there was a significant trend (P < 0.05) for inflorescence production per bud to increase by 40% between early (0.31 ± 0.05) and late application (0.43 ± 0.05) of shoot treatments, due to significantly (P < 0.05) more inflorescences per floral shoot between early (1.04 ± 0.14) and mid and late treatments (1.38 ± 0.14 and 1.42 ± 0.14 respectively) (Figures 4.9B and F). 130

167 Table 4.6: Effects of shoot girdling and stage of development treatments in the 2009/2010 season on the performance of Cabernet Sauvignon and Hayward buds compared with that of untreated shoots during spring 2010/2011. Data are means ± SEM, n = 7-8. Shoot treatment Stage of develop -ment Inflorescence /winter bud Budbreak (%) Floral budbreak (%) Inflorescence /floral shoot Cabernet Sauvignon Untreated 0.98 ± ± ± ± 0.41 Girdled early 0.93 ± ± ± ± 0.44 Girdled/defoliated early 0.91 ± ± ± ± 0.48 Girdled/shaded early 0.79 ± ± ± ± 0.44 Girdled mid 0.90 ± ± ± ± 0.48 Girdled/defoliated mid 0.79 ± ± ± ± 0.44 Girdled/shaded mid 0.66 ± ± ± ± 0.48 Girdled late 1.06 ± ± ± ± 0.41 Girdled/defoliated late 1.15 ± ± ± ± 0.45 Girdled/shaded late 1.03 ± ± ± ± 0.48 Hayward Untreated 0.59 ± ± ± ± 0.39 Girdled early 0.30 ± ± ± ± 0.39 Girdled/defoliated early 0.05 ± ± ± ± 0.39 Girdled/shaded early 0.33 ± ± ± ± 0.39 Girdled mid 0.58 ± ± ± ± 0.44 Girdled/defoliated mid 0.51 ± ± ± ± 0.54 Girdled/shaded mid 0.40 ± ± ± ± 0.48 Girdled late 0.80 ± ± ± ± 0.39 Girdled/defoliated late 0.30 ± ± ± ± 0.41 Girdled/shaded late 0.69 ± ± ± ± 0.41 P value Vine <0.001 <0.001 < Treatment Stage Vine*treatment Vine*stage Treatment*stage Vine*treatment*stage

168 Table 4.7: The effects of shoot girdling treatments and stage of development treatment in the 2010/2011 season on the performance of Cabernet Sauvignon and Hayward buds during spring 2011/2012 compared with that of untreated shoots. Data are means ± SEM, Cabernet Sauvignon n = 7, Hayward n = 11. Shoot treatment Stage Inflorescence / Budbreak Floral Inflorescence / 132 of Cabernet Sauvignon develop -ment winter bud (%) budbreak (%) floral shoot Untreated 0.53 ± ± ± ± 0.42 Girdled early 0.63 ± ± ± ± 0.36 Girdled/defoliated early 0.37 ± ± ± ± 0.36 Girdled/shaded early 0.30 ± ± ± ± 0.47 Girdled mid 0.50 ± ± ± ± 0.38 Girdled/defoliated mid 0.29 ± ± ± ± 0.38 Girdled/shaded mid 0.46 ± ± ± ± 0.38 Girdled late 0.57 ± ± ± ± 0.38 Girdled/defoliated late 0.65 ± ± ± ± 0.38 Girdled/shaded late 0.51 ± ± ± ± 0.38 Hayward Untreated 0.32 ± ± ± ± 0.30 Girdled early 0.23 ± ± ± ± 0.30 Girdled/defoliated early 0.06 ± ± ± ± 0.30 Girdled/shaded early 0.09 ± ± ± ± 0.31 Girdled mid 0.40 ± ± ± ± 0.30 Girdled/defoliated mid 0.01 ± ± ± ± 0.28 Girdled/shaded mid 0.33 ± ± ± ± 0.28 Girdled late 0.53 ± ± ± ± 0.28 Girdled/defoliated late 0.11 ± ± ± ± 0.28 Girdled/shaded late 0.32 ± ± ± ± 0.30 P value Vine <0.001 <0.001 < Treatment < < Stage Vine*treatment Vine*stage Treatment*stage Vine*treatment*stage

169 Figure 4.8: The effects of the stage of development that shoot treatments were applied in 2009/2010 compared with untreated shoots, on floral productivity of buds of Cabernet Sauvignon (untreated treated ) and Hayward (untreated treated ) in spring A, the number of inflorescences produced per bud; B, floral budbreak (%) and C, inflorescences per floral shoot. Data are means ± SEM, n =

170 Figure 4.9: The main effects of shoot girdling treatments (averaged over vine cultivar and stage of development treatment) (A, C and E) and the stage of development treatments (averaged over vine cultivar and shoot girdling treatment) (B, D and F) that were applied to shoots in 2010/2011 on floral productivity of buds in spring 2011/2012. A, and B, the number of inflorescences produced per bud; C, and D, floral budbreak; E, and F, inflorescences per floral shoot. Data are means ± SEM, n = Distribution of floral nodes The distribution of floral nodes along canes followed a similar pattern for both vines, with fewer inflorescences in the basal 20% of nodes on the canes and then an even spread across the remaining 80% of nodes on the cane (Figure 4.10). The only variation to this pattern was in the low flowering 2011 season (Tables 4.6 and 4.7), when there was a significant (P < 0.05) 42% decrease in the number inflorescences in the sub-apical portion (61-80% of nodes) compared with other nodes of Cabernet Sauvignon canes. Between the two vines, there were almost no inflorescences/winter bud in the basal 20% of Hayward canes (0.08 ± 0.07 in 2009/2010; 0.06 ± 0.05 in 2010/2011) compared with Cabernet Sauvignon canes (1.46 ± 0.08 in 2009/2010; 0.82 ± 0.05 in 2010/2011). For floral distribution along canes, there was a significant (P < 0.05) vine cultivar*shoot treatment interaction in both seasons (Figure 4.10). Inflorescence production was almost eliminated from the distal 60% of nodes 134

171 of girdled and defoliated Hayward canes, with significantly (P < 0.01) only 0.13 ± 0.07 inflorescences cf ± 0.23 inflorescences per section of untreated canes, in both seasons. There was a smaller reduction in inflorescence numbers in these nodes of girdled and defoliated Cabernet Sauvignon canes, to 77% of that on untreated canes (2.14 ± 0.18 cf ± 0.14 in 2010; 0.93 ± 0.18 cf ± 0.22 in 2011), but effects were not statistically significant. Inflorescence production was also lower at some nodes of girdled and shaded Hayward canes, reducing inflorescence production in the distal 60% of nodes to 0.52 ± 0.21 in 2010 (48%, not significant) and 0.31 ± 0.10 in 2011 (31%, P < 0.05) of untreated canes. This treatment also reduced Cabernet Sauvignon inflorescence numbers to 72% (1.93 ± 0.29 in 2010 and 0.91 ± 0.17 in 2011) of untreated canes, but again effects were not statistically significant. Figure 4.10: The effects of shoot girdling treatments (averaged over stage of development treatments) in the previous season on the number of inflorescences produced in each 20% of nodes from the base (0%) to the apex (100%) of canes. Treatment means are shown for Cabernet Sauvignon canes in A, 2010 and C, 2011 (untreated, girdled, girdled and defoliated and girdled and shaded ); and for Hayward canes in B, 2010 and D, 2011 (untreated, girdled, girdled and defoliated and girdled and shaded ). Data are means ± SEM, n =

172 4.4.5 Relationship between carbohydrate content, dry weight and bud performance To evaluate the relationship between bud performance in spring and the carbohydrate pools of the previous season a general linear model (GLM) was used. In particular inflorescence numbers per bud, budbreak, floral budbreak, and inflorescences per shoot in spring were tested against bud and internode dry weight and carbohydrate content in the previous winter, for the two vines (Figure 4.11). Vine cultivar explained the most variance (92%) in both budbreak and floral budbreak. None of the variables tested had a significant effect on inflorescence number per shoot, implying the model was a poor fit and the number of inflorescence per floral shoot depended on other factors. However, an additive model including both bud and internode carbohydrate content explained a 72% of the variance in inflorescence number per dormant bud (P < 0.001) (Table 4.8, Figure 4.11). Although vine cultivar explained a significant amount of variance in inflorescence number per bud on its own (54 %), it did not improve the fit of the model when added to bud and internode carbohydrate content. Table 4.8: The statistics of fitting the additive general linear model to bud performance data. The P value and r 2 for the model and the F value for each dependent variable are presented for each attribute tested. Statistics Inflorescences/ Budbreak Floral Inflorescences/ dormant bud (%) budbreak (%) floral shoot Overall P value <0.001 <0.001 < r Vine cultivar F value Bud carbohydrate content F value Internode carbohydrate content F value <0.001 <0.001 < <0.001 <0.001 < <

173 137 Figure 4.11: The relationship between bud dry weight (A - D), internode dry weight (E - H), bud carbohydrate content (I - L) and internode carbohydrate content (M - P), on inflorescence number per bud, total budbreak (%), floral budbreak (%) and inflorescence number per floral shoot. Data are for Cabernet Sauvignon ( ) and Hayward ( ) shoots in winter and spring of

174 4.5 Discussion The objective of the work described in this chapter was to modify shoot carbohydrate supply on inflorescence initiation during axillary bud development in the current season and, hence, floral production in the subsequent spring. To achieve this, the carbohydrate supply in shoots was manipulated by altering the photosynthate supply, either by defoliation (reduced leaf area) or shading (reduced photosynthesis rate), of shoots that had been phloem girdled to restrict import/export of carbohydrate at three separate stages (early, mid and late) of shoot development ( DAB). In the winter after treatments were applied, the TNC content of the axillary buds of Cabernet Sauvignon canes was reduced by 50% when photosynthate production had been reduced by either girdling and shade or girdling and defoliation during the earliest stage of bud development. By contrast, the carbohydrate content of Hayward buds in winter was not affected by shoot treatments in the previous season. However, reducing photosynthate supply by shoot girdling and defoliation of both vines or by girdling and shading of Cabernet Sauvignon shoots at all stages of inflorescence initiation caused a 75% reduction in internode carbohydrate content in winter. Thus, although the treatments did not consistently alter all the TNC pools in canes of both vines during winter, floral production was eliminated or markedly reduced in distal nodes of Hayward canes and reduced in Cabernet Sauvignon canes in the subsequent spring. Therefore, modifying the carbohydrate supply of shoots during inflorescence initiation and shoot development in the current season definitely modified the floral productivity in the following spring. In addition to altering the carbohydrate concentrations in buds and canes, there were important effects of treatments and seasons on the growth of shoots and hence their capacity to store carbohydrate later in the season. The shoots of Cabernet Sauvignon vines generally grew more vigorously than the Hayward canes. Reducing carbohydrate supply to Cabernet Sauvignon canes through girdling and shading or defoliation consistently reduced shoot growth throughout the period of rapid shoot development from 30 to 110 DAB in both seasons. However, only the growth of Hayward 138

175 shoots that were girdled or girdled and defoliated during early shoot growth (19-57 DAB) in the second year was decreased by reduced carbohydrate supply. Growth of shoots of both vines was 80 % greater in the second experimental season (2010/2011) than the first season (2009/2010) Effects of shoot treatments on inflorescence initiation Floral initiation in perennial plants can be disrupted by reducing the supply of mobile floral signals (FLOWERING LOCUS T (FT) protein and carbohydrate) and carbohydrate for meristem development from leaves (Corbesier et al. 1998). In the present study, production of inflorescences was almost eliminated (inflorescence number was 12% of that in untreated canes) on the distal 60% of girdled and defoliated Hayward canes, suggesting that inflorescence initiation had been disrupted. Snelgar and Manson (1992) have previously shown that girdling and defoliation can prevent flowering in Hayward canes but that study did not compare defoliation, shading, and the effects of girdling alone or examine floral productivity of individual buds. Elimination of inflorescence production on this portion of girdled and defoliated Hayward canes indicated that floral signals were apparently derived from leaves subtending axillary buds, as flowering occurred at the basal nodes where four leaves were retained but signals were not translocated distally from the basal leaves. There was no reduction in floral production in shoots that were only phloem girdled suggesting signals were not normally imported into shoots via phloem. Therefore, inflorescence initiation in Hayward buds required translocation of floral signals from subtending leaves. By contrast, inflorescence initiation did occur in Cabernet Sauvignon shoots that had been defoliated and girdled. Inflorescence production was reduced to 77% (not statistically significant) of that in untreated canes but not eliminated, suggesting that inflorescence initiation signals were most probably transmitted to buds from the basal leaves or other tissues in Cabernet Sauvignon shoots. Both temperature and light have been shown to act as the cues for floral initiation in grapevines and several shoot tissues in addition to leaves, can sense temperature cues and produce floral signals 139

176 (Buttrose 1969; Srinivasan and Mullins 1981). Thus, it was possible that either light or temperature could have induced inflorescence initiation in the partially defoliated Cabernet Sauvignon shoots. However, in earlier studies the response of grape cultivars to defoliation during inflorescence initiation has been variable, for example Bennett et al. (2005) showed a small reduction in inflorescence production by Chardonnay vines and similar to that for Cabernet Sauvignon vines in the current study. By contrast, Lavee et al. (1967) found that defoliation eliminated flowering from selected buds of Alphonse, Lavallee or Sultana grapevines. These results suggest, therefore, that inflorescence initiation in the Cabernet Sauvignon buds resulted from floral signals supplied to developing buds from basal leaves, other tissues in the shoot or were transported via xylem. Shading shoots reduced the quantity of light and hence carbohydrate supply from photosynthesis (see Chapter 6) (Greer et al. 2011). Although there was a general reduction (30-50%) in the number of inflorescence produced on the distal 60% of girdled and shaded Hayward and Cabernet Sauvignon canes, the effect of shade was only statistically significant for Hayward canes in 2010/2011 when inflorescence numbers on girdled and shaded shoots were reduced by 70%. Responses of shoots to shade in the current experiment were consistent with many other studies of both grape (Antcliff and Webster 1955; May and Antcliff 1963; Buttrose 1969; Hopping 1977) and kiwifruit (Fabbri et al. 1991; Snelgar et al. 1991; Snelgar et al. 1992) where subsequent floral production of buds from shaded shoots was reduced but not eliminated. This suggested that carbohydrate supply in heavily shaded Hayward and Cabernet Sauvignon did not limit inflorescence initiation but carbohydrate pools were not adequate to support floral morphogenesis in the following spring. Inflorescence initiation occurs in grape and kiwifruit canes during periods of rapid shoot growth when there is a large demand for carbohydrates. In 2009/2010, Hayward shoots largely grew during the first 60 DAB, and this is when defoliation eliminated inflorescence initiation in the distal 60% of nodes. However, in the 2010/2011 season, Hayward shoots grew rapidly for a 140

177 longer period (100 DAB) and the defoliation treatments eliminated inflorescence initiation at the distal 60% of nodes throughout this period. Consistent with this, a link between inflorescence initiation in Hayward buds and rapid shoot growth has been suggested in an earlier study by Snelgar and Manson (1992). However, there has been considerable debate about when inflorescence initiation in Hayward buds does occur (Davison 1990; Fabbri et al. 1991; Snelgar and Manson 1992; Snowball 1996; Walton et al. 2001). The results of the present study may infer that the apparent variation in the timing of inflorescence initiation in Hayward vines could be partially explained by seasonal variation in the period of shoot growth. Thus, inflorescence initiation in Hayward vines did occur when shoot growth was at a maximum and competition for carbohydrate supply was high. As defoliation did not eliminate inflorescence production in Cabernet Sauvignon buds the timing of inflorescence initiation could not be resolved. However, maximum shoot growth occurred during the early and mid shoot treatment stages in both seasons, when inflorescence initiation is reported to occur in other grape cultivars (Lavee et al. 1967; Swanepoel and Archer 1988). Therefore it is likely that inflorescence initiation in Cabernet Sauvignon buds occurred during rapid shoot growth when competition for carbohydrate supply was high Seasonal effects on shoot growth and floral development As indicated above there was a large difference in the growth of shoots between the two experimental years. The final length of shoots of both vines was two-fold greater in the 2010/2011 season than that recorded in the 2009/2010 season. Therefore, shoot growth was considerably faster during the period of inflorescence initiation in 2010/2011 than in 2009/2010. Shoot vigour during inflorescence initiation appeared to influence the subsequent production of inflorescences by both Cabernet Sauvignon and Hayward buds in the following spring. Canes produced from vigorous growth in 2010/2011 subsequently produced 50% less inflorescences in spring 2011/2012 than less vigorous shoots produced in 2009/2010 that flowered in spring 2010/2011. This suggests that the vigour of shoot growth during 141

178 inflorescence initiation influences the number of inflorescences produced in the subsequent spring. During the period of rapid shoot growth, minimum temperatures were 2.5 o C higher on average in December 2010 and January 2011 than the previous season (Appendix 2) and 2 o C higher than the 30-year average for this site. These higher minimum temperatures coincided with two-fold higher growth rates for Cabernet Sauvignon shoots and almost five-fold higher rates for Hayward shoots than those recorded during the previous season and highly suggestive of a temperature-induced increase in the growth rates of both vines. Altering temperatures by a similar amount between budbreak and anthesis also resulted in a two-fold increase in Cabernet Sauvignon vines in a recent study (Keller and Tarara 2010). Similar changes in minimum temperatures also resulted in a 34-63% increase in shoot growth rates of Pinot Noir and Riesling vines (Hendrickson et al. 2004). In general, high temperatures (28-30 o C) have also increased shoot growth in Hayward vines by 5-9-fold (Greer and Jeffares 1998; Richardson et al. 2004). Therefore it was likely that elevated growth rates of both Cabernet Sauvignon and Hayward shoots in 2010/2011 compared with the previous 2009/2010 season were due to an increase in minimum temperatures during the period of rapid shoot growth. The doubling of shoot growth of both vines and the concomitant reduction in inflorescence production in the subsequent growing season was consistent with a cumulative effect of competition for carbohydrate between the shoot and inflorescence growth. This conclusion conformed with similar results where vigorous shoot growth during inflorescence initiation was also associated with a reduction in inflorescence production in the following year of both grapes (Carbonneau and Casteran 1979; Vasconcelos et al. 2009) and Hayward vines (Richardson et al. 2004). Plant hormones, for example gibberellic acid (GA) and auxin, are generally known to stimulate vegetative growth and inhibit reproductive growth (Wilkie et al. 2008; Bangerth 2009) and may account for the above results. In the wild, this phytohormone control would have ensured that vines remain vegetative while they climbing 142

179 through the canopy and became floral only when sunlight was reached at the top of the canopy (Boss et al. 2003). Also consistent with this conclusion, during vigorous vegetative growth of grapevines, stem elongation and tendril formation are stimulated by GA and auxin, at the expense of inflorescence development (Srinivasan and Mullins 1980; Boss and Thomas 2002; Keller 2010). Applications of GA to Bruno kiwifruit also stimulated stem elongation (Lionakis and Schwabe 1984). Furthermore, a recent study of floral induction in citrus has shown that GA can reduce the expression of key flowering genes and regulate genes controlling metabolic pathways, for example increasing the carbohydrate supply needed to support rapid vegetative growth (Goldberg-Moeller et al. 2013). Therefore, the reduced inflorescence production in spring 2011 was likely caused by the increased minimum temperatures stimulating shoot growth in the previous season and perhaps GA production impacting on the carbohydrate supply and inhibiting inflorescence initiation Effects of treatments on shoot development and inflorescence production In the present study, reducing carbohydrate supply in Cabernet Sauvignon shoots by girdling and defoliation or girdling and shading during rapid shoot development had an immediate effect, reducing shoot growth by 30-40%. This was reflected in reduced bud and internode dry weight and carbohydrate content at the end of shoot development. Early (during rapid shoot growth) and mid development (anthesis to rapid berry growth) defoliation treatments also reduced Hayward shoot growth, but early girdling and shading increased Hayward shoot growth. At the end of shoot growth, internode dry weight and to a lesser extent internode carbohydrate content, were reduced on shoots that were girdled and defoliated or girdled and shaded during rapid shoot growth. In the season after treatments were applied to shoots, there was a positive relationship between the internode dry weights and carbohydrate pools in winter and the floral performance of buds in canes of both the vines in spring. The greater numbers of inflorescences produced by Cabernet Sauvignon 143

180 buds than Hayward buds reflected the much greater size and the carbohydrate pool in Cabernet Sauvignon internodes. Earlier studies of Sultana grapes have suggested that carbohydrate supply in the cane adjacent to the bud may influence the fruitfulness of buds (Thomas and Barnard 1937; Sommer et al. 2000; Goffinet 2004) and cane size has been linked with floral productivity in Hayward (Walton et al. 2000; Thorp et al. 2003). Therefore carbohydrate supply during the rapid growth of shoots in spring influenced internode dry weight at the end of the season and hence the floral performance of that cane in the following spring. Carbohydrate supply also influenced carbohydrate pools of the buds of both vines and there was also a good correlation between the amount of carbohydrate stored in buds and their subsequent floral development. Elsewhere, in both peach (Maurel et al. 2004; Bonhomme et al. 2005) and birch (Rinne et al. 2001) buds carbohydrate reserves support the very early stages of budswell and budbreak, before the buds became fully reconnected to the vascular system and hence able to access supplies of carbohydrate from elsewhere in the plant. In both Cabernet Sauvignon and Hayward buds, floral morphogenesis began during the early stages of budswell and budbreak and was likely influenced by the bud carbohydrate status. The good correlation between the bud and internode carbohydrate pools and the subsequent production of inflorescences suggested the very early stages of floral morphogenesis were influenced by localised supplies of carbohydrate and indicated a close coordination between carbohydrate source and inflorescence sink development (Smith and Stitt 2007; Stitt et al. 2007). Although this is not a direct effect of carbohydrate supply on inflorescence initiation, it is an important influence of carbohydrate supply during this period on the subsequent floral development Summary Modifying the carbohydrate supply of Cabernet Sauvignon and Hayward shoots during inflorescence initiation and rapid shoot growth, did not consistently alter cane carbohydrate status in winter, but nevertheless significantly influenced production of inflorescences in the following spring. 144

181 Preventing carbohydrate import through girdling and reducing photosynthesis by defoliation of Hayward shoots prevented inflorescence initiation in the distal 60% of nodes. However, preventing import and reducing photosynthesis by girdling and shading reduced but did not consistently eliminate flowering of either Hayward or Cabernet Sauvignon vines. This suggested that for Hayward vines floral cues were leaf-derived but were not primarily carbohydrate. A doubling of growth of both Cabernet Sauvignon and Hayward shoots in the 2010/2011 season was associated with a substantial reduction in inflorescence production in the following spring, and implied that factors associated with increased vegetative growth suppressed inflorescence initiation. Furthermore, there was a strong correlation between the carbohydrate pool in canes and their subsequent floral production. This suggests that shoot internode development during the period of inflorescence initiation was an important determinant of the carbohydrate pool available for early floral development of buds in spring. As the actual carbohydrate pools in Cabernet Sauvignon canes were higher than those in Hayward canes, this may have influenced the difference in their floral productivity. There was no evidence therefore, for a direct effect of carbohydrate supply on inflorescence initiation of either vine cultivar; however, there was evidence that altering carbohydrate supply during the period when inflorescence initiation occurred affected the carbohydrate pools that were available for floral morphogenesis in the following spring. 145

182 146

183 5 Modifying carbohydrate supply during autumn: effect on the early stages of floral morphogenesis in grape and kiwifruit vines 5.1 Introduction Plant species have evolved a wide range of strategies to optimise the balance between vegetative growth and reproductive success (Eckardt 1977). In general, annual plants rapidly develop their vegetative structure to maximise the capture of carbon through photosynthesis, creating a reserve to support reproductive development (Eckardt 1977). In deciduous perennial plants reserves allow reproductive development to occur before the canopy becomes fully functional, for example in species like willow, birch, maple, stonefruit, grapes and kiwifruit (Loescher et al. 1990; Kozlowski 1992). Grape and kiwifruit vines are specialised deciduous perennial plants that have adapted their pattern of growth to an environment with variable light and to rely on other plants for structural support. As part of this strategy, these vines contain large amounts of carbohydrate in storage pools (Loescher et al. 1990; Smith et al. 1992; Candolfi-Vasconcelos et al. 1994; Holzapfel et al. 2010) to support budbreak and an extended period of early vegetative and reproductive development in spring and to ensure survival and re-establishment in adverse conditions. For grapevines, it has been shown that 67% of carbohydrate is found in the roots, 23% in the trunk and 10% in canes during winter (Lebon et al. 2008). In kiwifruit, carbohydrate storage pools have not been quantified but in winter roots make up around 50% of the biomass of dormant vines compared with 15% in the trunk and 35% of biomass in the canopy structure, that includes canes and perennial wood (Smith et al. 1992). As starch concentrations are much higher in kiwifruit roots in winter ( mg gdw -1 ) than those in perennial wood (30-40 mg gdw -1 ) (Smith et al. 1992) it follows that kiwifruit roots also contain a high proportion of vine carbohydrate reserves in winter. In grape and kiwifruit vines, as in many other woody species, the development of flowers occurs over two seasons. Inflorescence initiation 147

184 occurs in the first season and floral organ morphogenesis occurs between budswell and anthesis in the following spring (Pratt 1971; Brundell 1975c; Srinivasan and Mullins 1981). Grape and kiwifruit vine mobilise reserves to support a large portion of new shoot development through until anthesis (Chapter 3), thus vines are heterotrophic during much of spring growth. The net carbon balance of newly developed Semillon grape (Greer and Sicard 2009; Greer et al. 2011) and Hayward shoots (Greer et al. 2003) is negative for at least DAB and shoots often don t become fully autotrophic until just prior to anthesis. At this time carbohydrate stored in perennial tissues of both kiwifruit and grapevines has reached minimum annual concentrations (Smith et al. 1992; Holzapfel et al. 2010) and as shown in Chapter 3. Generally the transition between heterotrophic and autotrophic status of shoots coincides with the critical stage of meiosis in floral development and low carbohydrate supply may be associated with abscission of grape flowers (Lebon et al. 2004b). Thereafter, the carbohydrate supply from photosynthesis peaks and tissue concentrations increase; however, storage occurs in competition with root and fruit growth (Chapter 3 and Greer et al. (2011)). Rapid increases in shoot (later canes), perennial wood and root reserves occur after veraison/the beginning of seed maturation and concentrations continue to increase until leaf senescence (Chapter 3 and Holzapfel et al. (2010)). Since both grape and kiwifruit vines depend on reserves to support rapid shoot growth and flower morphogenesis over an extended period of spring development, the reserve status of the vines can have a significant effect on floral productivity. For instance, decreasing carbon assimilation through leaf removal (Candolfi-Vasconcelos and Koblet 1990; Snelgar and Manson 1992; Bennett et al. 2005), by shading vines in autumn (Grant and Ryugo 1984; McArtney and Ferree 1999b; Sanchez and Dokoozlian 2005) or by water stress and high crop load (Dayer et al. 2013) has been associated with reduced flowering and vegetative growth of vines during the following spring. It has been suggested that reserves in several organs of grapevines including roots, trunks and canes are essential to support yield in the subsequent season (Weaver and McCune 1960; Candolfi-Vasconcelos and 148

185 Koblet 1990; Goffinet 2004; Bennett et al. 2005; Holzapfel and Smith 2012). In addition, overwintering grape and kiwifruit buds have high concentrations of carbohydrates which may be related to several processes, including protecting dormant meristems from low temperature injury (Wample and Bary 1992; Jones et al. 1999; Richardson et al. 2007) and/or supporting early floral development. However, no studies have thus far linked manipulation of carbon acquisition in autumn with the carbohydrate status of perennial and annual storage tissues and the subsequent development of buds in spring for both kiwifruit and grapevines. In the present study the objectives, therefore, were to manipulate the supply of carbohydrate to potential storage pools and to assess the effect on reproductive growth and development in the following season. Although the vines were grown under different management systems, this study examined basic physiological properties of vines. The carbohydrate supply to the perennial root and canopy wood was altered by reducing carbon assimilation through shading of grape and kiwifruit vine canopies in autumn (from approximately veraison/beginning of seed maturation to leaf fall; DAB). In addition, the supply of carbohydrate to the annually replaced canes and subtended buds was altered in autumn through a combination of shade and phloem girdling treatments to restrict export of carbohydrate from the canes. 149

186 5.2 Materials and methods Plant material This experiment was carried out on mature 24-year-old Hayward vines on Bruno rootstocks and nine-year-old Cabernet Sauvignon vines on SO4 rootstocks over three seasons (2008/2009, 2009/2010, 2010/2011). No Shiraz vines were available for this experiment. Mature Hayward vines were grown in the research orchard and Cabernet Sauvignon vines were grown in large pots at the Kerikeri Research Centre, where winter pruning could be managed to retain autumn treated canes for assessment in the subsequent spring. Full details of the site, vines and their management are given in Chapter 2. Briefly, all vines were cane-pruned with 16 canes (1.8 m in length and 20 buds per cane) on mature Hayward vines and two canes (1.0 m and 13 buds per cane) on the Cabernet Sauvignon vines. Vines were managed according to standard practices, including regular drip irrigation and additions of compound, slow-release fertiliser. The treatments were applied to the same vines in autumn in each of three seasons (2008/2009, 2009/2010 and 2010/2011) and the effect of modified carbohydrate reserve on inflorescence and shoot development was monitored in spring of each of the subsequent seasons (2009/2010, 2010/2011 and 2011/2012) Experimental design Eight mature Hayward kiwifruit vines and 12 Cabernet Sauvignon vines were chosen within one row for this experiment and treatments randomly allocated within blocks according to row position. A split plot design was used with vine light exposure (fully exposed or 85% shade) as the main plots and cane girdling treatments (intact or girdled) as the sub-plots within each vine. Shade treatments were randomly allocated to vines (plots) within each block in the first year of the experiment and treatments were applied to the same vines over all three seasons. Four Hayward and six Cabernet Sauvignon vines were shaded on 24 February 2009, 10 February 2010 and 15 February 2011, ( DAB Hayward and DAB Cabernet Sauvignon ) while the remaining vines were fully exposed to natural radiation. 150

187 Whole vines were shaded by covering them in neutral density shade cloth (Solarshade TM, Permathene, New Zealand) (Figures 5.1A and B). Calibration measurements showed that the shade cloth reduced average PAR levels by 85% (and PFD to 83 cf. 991 µmol m -2 ) and gave a slight 0.7% (0.3 o C) increase in mean temperature (Appendix 4). Shade treatments remained on the vines until leaf fall ( Cabernet Sauvignon DAB and Hayward vines DAB). Well developed canes were selected on each vine (six on Hayward vines and four on Cabernet Sauvignon vines) for cane treatments. Half of the selected canes on each vine (three on Hayward and two on Cabernet Sauvignon vines) were randomly chosen to be phloem-girdled and the remainder were left intact. Girdles (2-3 mm wide) were applied around 50 mm from the base of the cane (Figures 5.1C and D) using a scissor girdling tool. When necessary, girdles on kiwifruit canes were recut to maintain the girdles open until at least the end of March each year. Approximately 80% of Hayward and 50% of Cabernet Sauvignon girdles had healed by leaf fall, and all girdles had healed by budbreak in the following spring. 151

188 A B C D Figure 5.1 Autumn shade and girdling treatments applied to canes. A, shaded Hayward vine, B, shaded Cabernet Sauvignon vine, C, girdled Hayward cane and D, girdled Cabernet Sauvignon cane Measurements Phenology In spring, the times of budbreak and flowering were monitored three times a week as described in Chapter 2 on canes treated in the previous autumn. Prior to anthesis, all buds on the treated canes were assessed to determine budbreak (%), floral budbreak (%), inflorescence number per shoot, inflorescence number per winter bud (see Chapter 2) and the number of lateral flowers in Hayward inflorescences. Three Cabernet Sauvignon inflorescences per cane were covered with a net bag (0.5 mm mesh) during anthesis to collect flower caps (calyptras) which were subsequently counted to determine total flower number per inflorescence in 2009/2010 and 2011/2012. In 2010/2011, Cabernet Sauvignon flower caps were destroyed in a laboratory fire. At 50% anthesis in 2009/2010 and 2010/2011, a subsample of inflorescences (one for Cabernet Sauvignon and four for 152

189 Hayward ) were harvested from each treated cane and their length and fresh weight measured. At the same time, the length of three shoots on each cane was determined Carbohydrate analysis At the end of winter dormancy (early August) in 2009 and 2010, samples of perennial cordon and fine roots (< 2 mm) were collected from each Hayward and Cabernet Sauvignon vine and bud and internode samples were collected from both an intact and girdled cane within each vine. On each cane, five buds in the middle of canes (distal to node 10) and a portion of adjacent internode were collected. A sample of mm diameter wood was removed from the cordon of each Cabernet Sauvignon vine and a cordon sample was obtained from Hayward vines by drilling four holes (5 mm diameter and 40 mm deep) per vine and collecting the material from each drilling. All bud, internode and cordon tissue were immediately frozen in liquid nitrogen and stored at -80 o C for carbohydrate analysis. A root core 26 mm in diameter and 250 mm in length was sampled approximately 200 mm from the trunk of each Cabernet Sauvignon vine. A cube of soil with 150 mm sides was collected from the herbicide strip approximately 600 mm from each Hayward vine trunk. Root material was kept on ice for a short period, then it was washed and a sample of the fine roots (< 2 mm) was frozen in liquid nitrogen before storing at -80 o C. Carbohydrate analyses were carried out as described in Chapter 2. Carbohydrate concentrations from canes and buds were converted into carbohydrate content using bud fresh weight and cane fresh weight per unit length data, combined with cane length and DM% data derived during carbohydrate analyses. Root (14640 g vine -1 ) and cordon (7250 g vine -1 ) biomass data of Smith et al. (1992) for Hayward and root (2316 g vine -1 ) and cordon (1325 g vine -1 ) biomass data for Shiraz vines from Field (2013) were also used, as in Chapter 3, to calculate carbohydrate content for these tissues Statistical analysis All data was analysed using GenStat 14 th edition for Windows (VSN International Ltd., Hemel Hempstead, Hertfordshire, United Kingdom) using 153

190 REML linear mixed models assuming a fully randomised split plot design, to analyse treatment differences in shoot carbohydrate composition, budbreak, flowering, inflorescence size and shoot growth. Differences between treatment means were determined at the appropriate level of significance, and estimated means and standard errors were presented. Where necessary, data was log-transformed for analysis and back transformed for presentation. All data was plotted using Origin version 8.5 (Originlab Corporation, Northhampton, Massachusetts, USA) and curves were fitted to budbreak, flowering and shoot growth data across time using the Boltzmann sigmoid function (Seleznyova and Greer 2001). See Chapter 2 for further details of statistical analysis. 154

191 5.3 Results Total carbohydrate concentrations and content There was a generally consistent trend whereby TNC concentrations were higher in tissues of Cabernet Sauvignon than Hayward vines in winter of both 2009 and 2010 (Tables 5.1 and 5.2). For example, average TNC concentrations in axillary buds were more than two-fold higher in the Cabernet Sauvignon vines (120.0 ± 2.2 to ± 4.7 mg gdw -1 ) than the Hayward vines (58.1 ± 2.6 to 44.6 ± 5.2 mg gdw -1 ) in 2009 and 2010 respectively. Similar differences were observed for the perennial cordon wood in Cabernet Sauvignon vines (140.9 ± 4.7 to ± 4.9 mg gdw -1 ) compared with Hayward vines ( ± 5.5 mg gdw -1 ) and the roots of Cabernet Sauvignon (113.3 ± 8.0 to ± 6.8 mg gdw -1 ) compared with Hayward (54.8 ± 8.5 to 59.4 ± 7.6 mg gdw -1 ) vines. In the winter of 2010, TNC concentrations in Cabernet Sauvignon canes (107.7 ± 3.0 mg gdw -1 ) were also slightly higher concentration than those in Hayward canes (95.8 ± 3.3 mg gdw -1 ). Seasonal affects were small compared with the differences between vines. The shade (reduced vine carbon assimilation) and cane girdling (reduced cane phloem transport) treatments consistently influenced TNC concentrations in canes of both vines (Tables 5.1 and 5.2). For example, exposed canes had significantly (P = 0.001) higher TNC concentrations than shaded canes in both winters (exposed ± 4.2 to ± 3.2 mg gdw -1 cf. shade ± 4.3 to mg gdw -1 ) (averaged across vine cultivar and girdling treatment for 2009 and 2010, respectively) (Figures 5.2A and C). Girdled canes also had significantly (P = 0.001) and consistently higher TNC concentrations than those of intact canes in both winters (140.2 ± 3.5 to ± 3.3 mg gdw -1 cf ± 4.8 to 94.8 ± 3.1mg gdw -1 ) (averaged across the vine cultivar and light exposure for 2009 and 2010, respectively) (Figures 5.2B and D). However, there were no significant effects of either shade or girdling on TNC concentrations of the axillary buds, in either season. By contrast, shading vines in autumn significantly reduced the TNC concentration in roots in winter (P = 0.053) by 26% in 2009 and (P < 0.01) by 37% in 2010 (Figure 5.3). However, there were no significant effects of 155

192 shade treatments on concentrations of TNC in the cordon wood of vines in either season. Table 5.1: The effects of shoot treatments applied during autumn (February - June/July) 2009 on the TNC concentration of Cabernet Sauvignon and Hayward tissues during winter of the same season (August 2009). Data are means ± SEM, n = 4-6. Shoot treatment TNC concentration (mg gdw -1 ) Cabernet Sauvignon Cane Bud Cordon Root Exposed ± ± ± ± 12.0 Exposed/girdle ± ± Shade ± ± ± ± 10.7 Shade/girdle ± ± Hayward Exposed ± ± ± ± 12.0 Exposed/girdle ± ± Shade ± ± ± ± 12.0 Shade/girdle ± ± P value Vine cultivar <0.001 <0.001 <0.001 Shade Girdle Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

193 Table 5.2: The effects of shoot treatments applied during autumn (February - June/July) 2010 on the TNC concentration of Cabernet Sauvignon and Hayward tissues during winter of the same season (August 2010). Data are means ± SEM, n = 4-6. Shoot treatment TNC concentration (mg gdw -1 ) Cabernet Sauvignon Cane Bud Cordon Root Exposed ± ± ± ± 9.6 Exposed/girdle ± ± Shade 94.4 ± ± ± ± 9.6 Shade/girdle ± ± Hayward Exposed 97.7 ± ± ± ± 10.7 Exposed/girdle ± ± Shade 76.9 ± ± ± ± 10.7 Shade/girdle 96.1 ± ± P value Vine cultivar <0.001 <0.001 <0.001 Shade < Girdle Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

194 Figure 5.2: TNC concentrations in shaded and exposed canes (averaged across vine cultivar and girdling treatment) in A, winter 2009, C, winter 2010, and in girdled and intact canes (averaged across vine cultivar and shading treatment) in B, winter 2009 and D, winter Data are means ± SEM, n = 20. Figure 5.3: TNC concentrations in roots in response to light treatments (averaged across vine cultivar) in A, winter 2009 and B, winter Data are means ± SEM, n =

195 The greater biomass of the Hayward vines resulted in the TNC content of all organs, except axillary buds, being consistently higher than in Cabernet Sauvignon vines (Tables 5.3 and 5.4). For example, the TNC content of Hayward canes for 2009 and 2010, respectively (9.46 ± 0.29 to 7.88 ± 0.23 g) was fold higher than Cabernet Sauvignon canes (5.68 ± 0.25 to 5.64 ± 0.21 g), the TNC content of Hayward roots (802 ± 48 to 870 ± 68 g) was more than three-fold higher than the roots of Cabernet Sauvignon vines (262 ± 41 to 243 ± 61 g) and cordon wood (320 ± 13 to 379 ± 29 g) was fold higher than in Cabernet Sauvignon vines (187 ± 11 to 154 ± 24 g). By contrast, the TNC content of Cabernet Sauvignon axillary buds (4.91 ± 0.18 to 4.69 ± 0.19 mg) was fold higher than in Hayward buds (2.13 ± 0.20 to 1.63 ± 0.21 mg). In both seasons, there was a significant (P < 0.01) vine cultivar*shade interaction on the TNC content of roots whereby the roots of shaded Hayward vines were only 62% (2009) or 54% (2010) of that in exposed vines (Tables 5.3 and 5.4; Figures 5.4A and B). In 2010, the TNC content of Hayward cordons was also significantly (P < 0.001) different and only again 61% of that of exposed vines (Tables 5.3 and 5.4; Figures 5.4C and D). There was a significant (P < ) decrease in the TNC content of shaded compared to exposed canes in 2009 and 2010, respectively (7.02 ± 0.27 to 6.13 ± 0.22 g cf ± 0.27 to 7.39 ± 0.22 g) averaged over both vine cultivar and girdling treatment (Figures 5.4A and C). There was also an (P <0.01) increase in the TNC content of girdled compared with intact canes in 2009 and 2010 (8.21 ± 0.23 to 7.25 ± 0.22 g cf ± 0.31 to 6.27 ± 0.22 g) averaged over vine cultivar and light treatment (Figures 5.4B and D). There was however, no effect of either shade or girdling treatments on the carbohydrate content of axillary buds in either season. 159

196 Table 5.3: The effects of shoot treatments applied during autumn (February - June/July) 2009 on the TNC content of Cabernet Sauvignon and Hayward tissues during winter (August 2009). TNC concentrations from Tables 5.1 and 5.2 were combined with bud and cane fresh biomass, cane length and DM% data from carbohydrate analyses to calculate bud and cane TNC content. Biomass data from Smith et al. (1992) for Hayward and Field (2013) for Shiraz together with TNC concentrations were used to calculate cordon and root TNC content. Data are means ± SEM, n = 4-6. Shoot treatment TNC content Cabernet Sauvignon Cane (g) Bud (mg) Cordon (g) Root (g) Exposed 6.11 ± ± ± ± 60.4 Exposed/girdle ± 4.99 ± Shade 4.45 ± ± ± ± 67.6 Shade/girdle 5.52 ± ± Hayward Exposed 9.27 ± ± ± ± 67.6 Exposed/girdle ± ± Shade 7.88 ± ± ± ± 67.6 Shade/girdle ± ± P value Vine cultivar <0.001 <0.001 <0.001 <0.001 Shade Girdle Vine*shade < Vine*girdle Shade*girdle Vine*shade*girdle

197 Table 5.4: The effects of shoot treatments applied during autumn (February - June/July) 2010 on the TNC content of Cabernet Sauvignon and Hayward tissues during winter (August 2010). TNC concentrations from Tables 5.1 and 5.2 were combined with bud and cane fresh biomass, cane length and DM% data from carbohydrate analyses to calculate bud and cane TNC content. Biomass data from Smith et al. (1992) for Hayward and Field (2013) for Shiraz together with TNC concentrations were used to calculate cordon and root TNC content. Data are means ± SEM, n = 4-6. Shoot treatment TNC content Cabernet Sauvignon Cane (g) Bud (mg) Cordon (g) Root (g) Exposed 5.79 ± ± ± ± 66.7 Exposed/girdle 6.49 ± ± 0.42 Shade 4.95 ± ± ± ± 66.7 Shade/girdle 5.37 ± ± 0.38 Hayward Exposed 8.03 ± ± ± ± 75 Exposed/girdle 9.26 ± ± 0.42 Shade 6.32 ± ± ± ± 74.6 Shade/girdle 7.90 ± ± 0.42 P value Vine cultivar <0.001 <0.001 <0.001 <0.001 Shade < <0.001 Girdle Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

198 Total carbohydrate content (g) 1200 A Root 2009 B Root C Cordon 2009 D Cordon Shade Exposed Shade Exposed Figure 5.4: TNC content in roots and cordons of Cabernet Sauvignon and Hayward in response to light treatments. A, roots winter 2009, B, roots winter 2010, C, cordon winter 2009 and D, cordon winter Note the difference in scale between A, B, and C, D. Data are means ± SEM, n = Figure 5.5: TNC content of exposed and shaded canes (averaged across vine cultivar and girdling treatment) in A, winter 2009, C, winter 2010, and in girdled and intact canes (averaged across vine cultivar and shading treatment) in B, winter 2009 and D, winter Data are means ± SEM, n =

199 5.3.2 Amount of budbreak and flowering The number of inflorescences produced per bud varied both between vines and seasons (Tables ). Hayward vines produced more inflorescences per bud in both springs (0.87 ± 0.11 to1.37 ± 0.07) than the Cabernet Sauvignon vines (0.60 ± 0.11 to 0.88 ± 0.11). These differences between the vines were influenced by consistently higher (50%) total and floral budbreak on the Cabernet Sauvignon canes but higher inflorescence production per floral shoot (25-53%) on the Hayward canes (Tables ). Altering carbohydrate supply within the vines through reduced carbon assimilation (85% shade) or preventing phloem export from canes (girdling) consistently affected the productivity of buds in the subsequent season (Tables ). In the spring of both the 2009 and 2010 seasons, there was a significant interaction (P = ) between vine cultivar*shade on the number of inflorescences produced per dormant bud (Figures 5.6A and B). In both seasons, Hayward canes that had been well exposed in the previous autumn produced significantly (P < 0.001) more inflorescences per winter bud (1.29 ± 0.15 to 1.90 ± 0.10) (averaged across girdling treatments) than shaded canes (0.42 ± 0.14; to 0.84 ± 0.10), but there was no significant effect of shading on the inflorescence production of Cabernet Sauvignon canes (0.61 ± 0.09 to 0.89 ± 0.21). However in the final spring, after three seasons of autumn shade, inflorescence production was significantly reduced (P < 0.001) across both vines, with well exposed canes producing 0.61 ± 0.05 inflorescences per bud compared with 0.35 ± 0.05 inflorescences per bud on the shaded canes (Figure 5.6E). Reducing shoot carbohydrate export in autumn by girdling canes significantly (P < 0.05) increased the number of inflorescences produced per bud (averaged across the vines) in both 2009/2010 (girdled 0.92 ± 0.08 cf. intact 0.55 ± 0.13) and 2011/2012 (girdled 0.61 ± 0.05 cf. intact 0.34 ± 0.05) (Figures 5.6B and F). However, in 2010/2011, there were significant interactions between both vine*girdle (P < 0.001) and shade*girdle (P < 0.05) on the number of inflorescences produced per bud. Again, girdled Hayward 163

200 canes produced more inflorescences per bud (1.94 ± 0.10) (averaged across shading treatments) than intact canes (0.81 ± 0.10) (Figure 5.6D), but there was no effect on Cabernet Sauvignon canes. However, when averaged across vine cultivar, shaded intact canes produced significantly (P < 0.001) fewer inflorescences per bud (0.42 ± 0.12) than canes from the three other treatments. The numbers of inflorescences produced per bud on canes that had been either shaded and girdled (1.23 ± 0.13) or exposed and intact (1.20 ± 0.12) were similar, but significantly (P < 0.05) fewer than for those canes that were exposed and girdled (1.66 ± 0.14) in the previous autumn. Figure 5.6: The effect of shade treatments on the number of inflorescences produced per winter bud in A, 2009/2010 (averaged across girdling treatment), C, 2010/2011 (averaged across girdling treatment) and E, 2011/2012 (averaged across vine cultivar and girdling treatments). The effect of girdling treatments on the number of inflorescences produced per winter bud in B, 2009/2010 (averaged across vine cultivar and shade treatments), D, 2010/2011 (averaged across shade treatments) and F, 2011/2012 (averaged across vine cultivar and shade treatments). Data are means ± SEM, n = In A, C and D data are for Cabernet Sauvignon and Hayward. Data are means ± SEM, n = Contributing to these differences in inflorescences per bud were consistent and significant effects of both shade and girdling treatments on the total budbreak, floral budbreak and the number of inflorescences per shoot 164

201 (Figures ; Tables ). Shading canes in the previous season consistently reduced the total budbreak, with significant effects (P < 0.01) averaged across vines in spring 2009 and 2011 (exposed 56.1 ± 2.9 to 48.7 ± 1.7% cf. shaded 43.0 ± 2.9 to 37.2 ± 1.8%) (Figures 5.7A and E). Effects of shade on floral budbreak were similar to those on total budbreak, with significant effects (P < 0.01) when averaged across vine cultivar (exposed 40.6 ± 3.2 to 33.6 ± 1.6% cf. shaded 28.9 ± 3.2 to 22.1 ± 1.7%) (Figures 5.8B and F) in spring 2009 and 2011 respectively. However, there were no significant effects of shade on total budbreak in the spring 2010, but there was a significant interaction between shade*girdle, whereby canes that were shaded but intact had significantly (P < 0.001) lower floral budbreak (28.9 ± 3.1%) than that averaged across the other three treatments (52.4 ± 3.3%). Figure 5.7: The effect of shade treatments (averaged across vine cultivar and girdling treatments) on total budbreak A, 2009/2010 and C, 2011/2012 and floral budbreak B, 2009/2010 and D, 2011/2012. Data are means ± SEM, n = 10. Girdling of canes also affected both the total budbreak and floral budbreak (Figure 5.8; Tables 5.7 and 5.8). Canes that were girdled had significantly (P < 0.01) higher total budbreak than intact canes in both the 2009 and 2011 springs (girdled 55.5 ± 2.2 to 44.7 ± 1.7% cf. intact 43.5 ± 1.7 to 41.2 ± 1.7%) (Figures 5.8A and E). Consistent with this, girdled canes also had higher (P 165

202 < 0.01) floral budbreak in both years (girdled 41.5 ± 2.4 to 32.2 ± 1.7% % cf. intact; 28.0 ± 2.4 to 23.5 ± 1.7%) (Figures 5.8B and F). In spring 2010, girdled Hayward canes had significantly (P < 0.001) higher total budbreak (51.1 ± 2.5%) than intact canes (33.3 ± 2.5%) and higher floral budbreak (girdled 44.5 ± 2.6% cf. intact 20.1 ± 2.6%) (Figures 5.8C and D). There were no significant effects of girdling on Cabernet Sauvignon canes. Figure 5.8: The effect of girdling treatments on total budbreak (A, 2009, C, 2010 and E, 2011) and floral budbreak (B, 2009, D, 2010 and F 2011). A, B, E and F averaged across vine cultivar and shade treatment. C and D averaged across shade treatment, and data are for Cabernet Sauvignon and Hayward. Data are means ± SEM, n = In terms of the number of inflorescences produced per shoot, there were significant vine cultivar*shade*girdle interactions in both 2009 (P < 0.05) and 2010 (P < 0.01) (Figures 5.9A and B); however, the effect of treatments was only significant on the Hayward vines. The number of inflorescences produced on shaded but intact Hayward shoots was significantly (P < 0.001) fewer during both spring 2009 (0 inflorescences per shoot) and spring 2010 (1.31 ± 0.17 inflorescences per shoot) than that of other three treatments. Canes that had been shaded and girdled in both seasons (2.99 ± 0.23 to 3.85 ± 0.22 inflorescences per shoot) also had significantly (P < 0.05) fewer 166

203 inflorescences per shoot than exposed shoots in the same seasons (3.83 ± 0.52 to 4.67 ± 0.22 averaged across the girdling treatments). In the final spring (2011), there were significant main effects (averaged across vine cultivar) of both shade (exposed 1.98 ± 0.17 cf. shaded 1.68 ± 0.19 inflorescences per shoot) and girdling (girdled 2.14 ± 0.18 cf. intact 1.53 ± 0.18 inflorescences per shoot) on the numbers of inflorescences produced for each shoot (Figure 5.9C). Figure 5.9: Effects of shade and girdling treatments on inflorescence numbers per floral shoot. A, spring 2009, B, spring 2010 and C, spring Data are for Cabernet Sauvignon and Hayward vines. Data are means ± SEM, n =

204 168 Table 5.5: The effect of treatments applied to vines in February 2009 on the performance of Cabernet Sauvignon and Hayward buds during spring Data are means ± SEM, n = 4-6. Shoot treatment Inflorescence / winter bud Budbreak (%) Floral budbreak (%) Inflorescence / floral shoot Flowers / inflorescence Cabernet Sauvignon Exposed 0.47 ± ± ± ± ± 29 Exposed/girdle 0.67 ± ± ± ± ± 46 Shade 0.52 ± ± ± ± ± 37 Shade/girdle 0.75 ± ± ± ± ± 62 Hayward Exposed 1.16 ± ± ± ± ± 0.08 Exposed/girdle 1.42 ± ± ± ± ± 0.02 Shade 0 ± ± ± 0 0 ± 0 0 ± 0 Shade/girdle 0.84 ± ± ± ± ± 0.01 P value Vine cultivar <0.001 <0.001 <0.001 <0.001 Shade Girdle Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

205 169 Table 5.6: Effects of treatments applied to vines in February 2010 on the performance of Cabernet Sauvignon and Hayward buds during spring Data are means ± SEM, n = 4-6. Shoot treatment Inflorescence / winter bud Budbreak (%) Floral budbreak (%) Inflorescence / floral shoot Flowers / inflorescence 1 Cabernet Sauvignon Exposed 0.86 ± ± ± ± Exposed/girdle 1.05 ± ± ± ± Shade 0.76 ± ± ± ± Shade/girdle 0.85 ± ± ± ± Hayward Exposed 1.54 ± ± ± ± ± 0.06 Exposed/girdle 2.26 ± ± ± ± ± 0.06 Shade 0.07 ± ± ± ± ± 0.08 Shade/girdle 1.61 ± ± ± ± ± 0.06 P value Vine cultivar <0.001 <0.001 <0.001 < Shade < <0.001 < Girdle <0.001 <0.001 <0.001 < Vine*shade < Vine*girdle < Shade*girdle < Vine*shade*girdle Cabernet Sauvignon flower caps were destroyed in a laboratory fire. 169

206 170 Table 5.7: The effects of treatments applied to vines in February 2011 on the performance of Cabernet Sauvignon and Hayward buds during spring Data are means ± SEM, n = 4-6. Shoot treatment Inflorescence / winter bud Budbreak (%) Floral budbreak (%) Inflorescence / floral shoot Flowers/ inflorescence Cabernet Sauvignon Exposed 0.55 ± ± ± ± ± 31 Exposed/girdle 0.69 ± ± ± ± ± 68 Shade 0.36 ± ± ± ± ± 39 Shade/girdle 0.53 ± ± ± ± ± 29 Hayward Exposed 0.35 ± ± ± ± ± 0.12 Exposed/girdle 0.84 ± ± ± ± ± 0.13 Shade 0.10 ± ± ± ± ± 0.09 Shade/girdle 0.39 ± ± ± ± ± 0.06 P value Vine cultivar <0.001 <0.001 <0.001 <0.001 Shade <0.001 <0.001 < Girdle < <0.001 < Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

207 5.3.3 Time of budbreak and flowering In spring of the three seasons, budbreak of Cabernet Sauvignon and Hayward canes began at approximately the same date but the mid-point of budbreak (when 50% of budbreak had occurred) was earlier by 7-11 days on Cabernet Sauvignon than on Hayward canes (Table 5.8; Figure 5.10). There was no consistent difference in the duration of budbreak on canes of the two vines. In the spring of 2009 there were significant shade*girdle (P < 0.01) and vine*girdle (P < 0.001) interactions on the mid-point of budbreak. The midpoint of budbreak on shaded intact canes (averaged across vines) was significantly different and 13 days later (33.4 ± 2.1 DAS) than that on canes from other treatments (20.1 ± 1.6 DAS) (Table 5.8). In addition for Hayward canes, the mid-point of budbreak was significantly (P < 0.01) earlier on girdled than on intact canes (Table 5.8). In both spring 2010 and 2011, there was a significant (P < ) vine cultivar*shade*girdle interaction on the mid-point of budbreak. The mid-point of budbreak on exposed and girdled Cabernet Sauvignon canes was 14.6 days earlier than the mean mid-point for other cane treatments in spring 2010 and 10 days earlier than that of shaded and girdled canes in spring In both seasons the midpoint of budbreak on girdled Hayward canes was similar across light treatments and significantly (P < ) nine days earlier than on exposed and intact canes and 22 days earlier than shaded intact canes. In spring 2009 and 2010, there were significant (P < 0.05) vine*girdle effects on the duration of budbreak (Table 5.8). For instance the duration of budbreak on girdled Hayward canes (14.7 ± 1.2 to 7.3 ± 0.8 days) was significantly less than that on intact canes (27.6 ± 2.5 to 10.3 ± 0.8 days), but there was no significant effect of treatments on Cabernet Sauvignon canes. 171

208 Figure 5.10: Time of budbreak for Hayward (A, spring 2009 B, spring 2010 and C, spring 2011) and Cabernet Sauvignon (D, spring 2009, E, spring 2010 and F, spring 2011) canes. Budbreak (%) is the number of buds that broke at each date as a percentage of the final number of buds that broke. Treatments are Hayward exposed and shaded, and Cabernet Sauvignon exposed and shaded with intact canes represented by a solid line and girdled canes by a dashed line. Data points are means ± SEM, n =

209 173 Table 5.8: The effects of treatments applied during reserve storage in autumn on the mid-point and duration of budbreak of Cabernet Sauvignon and Hayward canes during the following spring, calculated from fitting the Boltzmann sigmoid function to data. Data are presented for three seasons and are means ± SEM, n = 4-6. Shoot treatment Mid-point of budbreak (days after 1 September) Duration of budbreak (days from 20-80%) Spring 2009 Spring 2010 Spring 2011 Spring 2009 Spring 2010 Spring 2011 Cabernet Sauvignon Exposed 12.8 ± ± ± ± ± ± 3.9 Exposed/girdle 17.4 ± ± ± ± ± ± 2.9 Shade 23.9 ± ± ± ± ± ± 4.2 Shade/girdle 17.6 ± ± ± ± ± ± 4.7 Hayward Exposed 29.2 ± ± ± ± ± ± 2.7 Exposed/girdle 20.8 ± ± ± ± ± ± 2.9 Shade 42.8 ± ± ± ± ± ± 2.8 Shade/girdle 23.6 ± ± ± ± ± ± 3.3 P value Vine cultivar <0.001 <0.001 < Shade <0.001 < Girdle <0.001 <0.001 <0.001 < Vine*shade Vine*girdle < Shade*girdle Vine*shade*girdle <

210 In the spring of both 2010 and 2011, Hayward canes reached the mid-point of flowering three - five days earlier than Cabernet Sauvignon canes (Table 5.9; Figure 5.11). In 2009, there was a significant (P < 0.01) vine*girdle interaction on the mid-point of flowering (Table 5.9) whereby Hayward canes that had been girdled in the previous autumn reached the mid-point of flowering significantly (P < 0.05) earlier by four days than intact canes. In addition, canes that were girdled (averaged across vine cultivar and shade treatment) reached the mid-point of flowering significantly (P < and 0.05) earlier by two - four days in 2010 and 2011 than intact canes. In the final spring, there was also a significant (P < 0.05) shade*girdling interaction whereby canes that had been well exposed and girdled (averaged across vine cultivar) reached the mid-point of flowering five days earlier than those on intact exposed canes (Table 5.9). The duration of flowering (the period between 20% and 80% flowering) was two days less in Hayward canes in both 2010 and 2011 than Cabernet Sauvignon canes (Table 5.9). In 2009 there was a significant (P < 0.01) interaction between shade*girdle on the duration of flowering whereby flowering of shaded and girdled canes (averaged across vine cultivar) was three days shorter than that in shaded intact canes. There were also significant effects (P < 0.01) of the previous girdling treatments on the duration of flowering in both 2010 and 2011, with the intact canes flowering over a shorter period ( days) than girdled canes. In the 2010, the duration of flowering on shaded canes (3.5 ± 0.4 days) was also significantly (P < 0.05) shorter than that for the well exposed canes (4.7 ± 0.3 days). 174

211 Figure 5.11: Time of flowering for Hayward (A, spring 2009 B, spring 2010 and C, spring 2011) and Cabernet Sauvignon (D, spring 2009, E, spring 2010 and F, spring 2011) canes. Flowering (%) is the number of flowers open at each date as a percentage of the total flower number at the end of anthesis. Treatments are Hayward exposed and shaded, and Cabernet Sauvignon exposed and shaded, with intact canes represented by a solid line and girdled canes by a dashed line. Data are means ± SEM, n =

212 176 Table 5.9: The effects of treatments applied during reserve storage in autumn on the mid-point and duration of flowering of Cabernet Sauvignon and Hayward canes during the following season, calculated from fitting the Boltzmann sigmoid function to data. Data are presented for three seasons and are means ± SEM, n = 4-6. Shoot treatment Mid-point of flowering (DAB) Duration of flowering (days from 20-80%) Cabernet Sauvignon Spring 2009 Spring 2010 Spring 2011 Spring 2009 Spring 2010 Spring 2011 Exposed 80.3 ± ± ± ± ± ± 1.02 Exposed/girdle 80.1 ± ± ± ± ± ± 1.14 Shade 79.4 ± ± ± ± ± ± 1.32 Shade/girdle 80.5 ± ± ± ± ± ± 1.02 Hayward Exposed 83.4 ± ± ± ± ± ± 0.66 Exposed/girdle 77.9 ± ± ± ± ± ± 0.69 Shade 82.1 ± ± ± ± ± ± 0.11 Shade/girdle 79.2 ± ± ± ± ± ± 0.69 P value Vine cultivar < < Shade Girdle < Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

213 Inflorescence size The nature of the inflorescences on each vine cultivar meant that Cabernet Sauvignon inflorescences were two - three-fold longer and carried 200-fold more flowers than Hayward inflorescences (Table 5.10). However, the weights of inflorescences from the two vines were similar. The effect of the vine*girdle interaction on inflorescence weight was significant (P < 0.05) in the spring of 2009 and less so (P = 0.09) in the spring of 2010 (Figure 5.12). Inflorescences on shoots from previously girdled Cabernet Sauvignon canes were significantly (P < 0.01) and two-fold heavier (3.9 ± 0.3 g) than those from intact canes (2.0 ± 0.4 g) in 2009 and 54% (P < 0.05) heavier (4.3 ± 0.4 g) than those from intact canes (2.8 ± 0.5 g) in There was also a significant vine cultivar*shade interaction (P < 0.05) on flower number per inflorescence in 2009, whereby previously shaded Cabernet Sauvignon canes produced more inflorescences with 377 ± 50 flowers than 285 ± 37 flowers on inflorescences from the fully exposed canes (Table 5.10). In 2010 Cabernet Sauvignon caps were lost in a laboratory fire prior to counting. There were no significant effects of shade or girdling treatments in the previous season on the length of inflorescences in either 2009 or 2010 (Table 5.10). 177

214 178 Table 5.10: The effects of treatments applied to vines in February on the size of Cabernet Sauvignon and Hayward inflorescences during the following spring in the 2009 and 2010 seasons. Data are means ± SEM, n = 4-6. Shoot treatment Inflorescence length (mm) Inflorescence weight (g) Flowers / inflorescence Cabernet Sauvignon Spring 2009 Spring 2010 Spring 2009 Spring 2010 Spring 2009 Spring Exposed 104 ± ± ± ± ± 29 - Exposed/girdle 134 ± ± ± ± ± 46 - Shade 134 ± ± ± ± ± 37 - Shade/girdle 136 ± ± ± ± ± 62 - Hayward Exposed 55.6 ± ± ± ± ± ± 0.06 Exposed/girdle 52.5 ± ± ± ± ± ± 0.06 Shade 50.2 ± ± ± ± ± ± 0.08 Shade/girdle 53.9 ± ± ± ± ± ± 0.06 P value Vine cultivar <0.001 < < Shade Girdle Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle Flower caps destroyed in a laboratory fire. 178

215 Figure 5.12: Effects of girdling treatments (averaged across shade treatments) on inflorescence weight in A, spring 2009 and B, spring Data are from Cabernet Sauvignon and Hayward canes. Data are means ± SEM, n =

216 5.3.4 Shoot growth The growth of shoots between budbreak and anthesis was greater for the Hayward vines (1160 ± 92 mm) compared to Cabernet Sauvignon vines (702 ± 89 mm) in the 2009 spring (Table 5.11). However there was no difference in the 2010 spring, when as observed in Chapter 4, the average growth of shoots for both vines was % greater than in the 2009 spring. In the 2010 spring, the growth of shoots from canes that had been shaded in the previous autumn (1277 ± 113 mm) (averaged over vine cultivar) was significantly (P < 0.05) (24%) less than that of shoots from continually exposed canes (1670 ± 111 mm) (Figure 5.13). There were no significant effects of shading in the previous autumn on shoot growth in 2009 spring, or girdling treatments on shoot growth in either season (Table 5.11). Table 5.11: The effects of treatments applied to vines in February (autumn) on the length of Cabernet Sauvignon and Hayward shoots at anthesis in the subsequent spring. Data are means ± SEM, n = 4-6. Shoot treatment Shoot length (mm) Spring 2009 Spring 2010 Cabernet Sauvignon Hayward Cabernet Sauvignon Hayward Exposed 928 ± ± ± ± 233 Exposed/girdle 580 ± ± ± ± 233 Shade 836 ± ± ± ± 233 Shade/girdle 466 ± ± ± ± 269 P value Vine cultivar < Shade Girdle Vine*shade Vine*girdle Shade*girdle Vine*shade*girdle

217 Figure 5.13: The effect of shade on canes in autumn of the previous growing season (2009/2010) on the growth of shoots between budbreak and anthesis in the spring of the 2010/2011 season. Data for light exposure treatments are averaged across vine cultivar and girdling treatment. Data are means ± SEM, n =

218 5.4 Discussion The objective of this chapter was to determine what effects reducing vine carbon acquisition in autumn through shading vines or modifying carbohydrate import/export from canes by phloem girdling, had on floral development of vines in the subsequent spring. After reducing the carbon acquisition of vines in autumn through shading for two years, the carbohydrate pool of Hayward roots was reduced by 521 g vine -1 and Cabernet Sauvignon roots by 107 g vine -1. The carbohydrate pool of each Hayward cane was also reduced by 1.5 g (24 g vine -1 ) and each Cabernet Sauvignon cane by 1.0 g (2.0 g vine -1 ) after two seasons of autumn shade. Restricting carbohydrate export from canes in autumn through phloem girdling, by contrast, consistently increased the carbohydrate pool of Hayward canes by 1.6 g (25.6 g vine -1 ) and Cabernet Sauvignon canes by 0.7 g (1.4 g vine -1 ). Thus, these treatments had the desired effect of changing the carbohydrate status of both vines; however, with the exception of buds, all other organs of Hayward vines had significantly and markedly larger carbohydrate pools (root 1130 g, cordon 378 g, 128 g vine -1 ) than the Cabernet Sauvignon vines (root 296 g, cordon 187 g, 12 g vine -1 ). The reduction in carbon acquisition in autumn was consistently associated with reduced floral productivity of Hayward vines in the subsequent spring; however, inflorescence production was only reduced in Cabernet Sauvignon vines after three consecutive seasons of treatment. Reduced carbohydrate export from canes was also consistently associated with increased floral productivity of Hayward canes and with increased floral productivity in Cabernet Sauvignon canes in two out of the three seasons. It would appear, therefore, that canes and roots are the major sinks for carbohydrate storage in autumn and changing pools in these sinks had demonstrable effects on floral morphogenesis in the subsequent spring Effect of treatments on total carbohydrate accumulation Manipulating carbon acquisition and transport within vines in autumn had a significant effect on the TNC content of roots of the dormant vines in winter, suggesting changes in reserve accumulation had occurred. Generally the TNC concentrations in the roots of untreated vines of both vines increased 182

219 (Chapter 3) throughout the autumn treatment period ( DAB). Heavy shade during this period would have reduced photosynthesis rates to about 18% of that of exposed vines (exposed 8.7 ± 0.1 µmol m -2 s -1 cf. shaded 1.6 ± 0.1 µmol m -2 s -1 ; see Chapter 6). This resulted in a substantial reduction in the TNC content of roots of both vines, with a similar 40-50% reduction for Hayward vines in both seasons but a similar reduction in root carbohydrate content on Cabernet Sauvignon vines only after the second season of treatment. Although root growth was not measured in this experiment, significant growth of fine roots on untreated vines occurred during autumn (see Chapter 3). It is possible that this growth may also have been reduced in vines subject to lower carbon acquisition, thereby increasing the effect of this treatment on vine root reserves. This would have been particularly so for Cabernet Sauvignon vines with a smaller root biomass (Holzapfel et al. 2010; Field 2013). Reducing carbon acquisition of grapevines by shading vines in the previous season has, however, been established to cause a reduction in root dry weight and also the concentration of soluble sugars in xylem extracts during the subsequent spring (McArtney and Ferree 1999b). Unfortunately, xylem carbohydrate concentrations in spring are generally only 5% of those in phloem (Ferguson et al. 1983; Keller 2010) and hence probably not indicative of changes in carbohydrate supply from the root reserves. Therefore, it can be concluded that changes in root reserves from a shade induced reduction in carbon acquisition in autumn would have influenced the subsequent reproductive development of vines in spring. Altering the carbohydrate supply of canes through shade or girdling treatments, by contrast, had no effect on the TNC content of axillary buds in this study. The seasonal patterns of variation in concentration in axillary buds (Chapter 3) indicated very limited change during the treatment period ( DAB). This suggested the axillary buds had completed their development at this late stage of shoot development and were not influenced by the treatments. Nevertheless, reduced carbon acquisition (shade) had decreased the TNC content of the canes while reduced carbohydrate export (girdling) had increased the cane TNC content of both vines notably, by about equivalent amounts. It was apparent from seasonal changes in TNC 183

220 concentrations in the shoots/canes of both Hayward and Cabernet Sauvignon vines (Chapter 3), that carbohydrate concentrations increased steadily during late summer and autumn, once shoot growth was complete. Consistent with this, Hayward shoots/canes that had developed in heavy shade throughout the season had lower dry matter accumulation (Grant and Ryugo 1984). Furthermore, in other grape studies where photosynthate production was decreased by defoliation in autumn, starch reserves in canes were concomitantly reduced during the winter (Candolfi-Vasconcelos and Koblet 1990; Koblet et al. 1993; Goffinet 2004). Other studies have also shown that girdling grape shoots around anthesis increased the dry matter content at the base of the shoot (Caspari et al. 1998) and soluble sugars in the bark (Hunter and Ruffner 2001). Thus, altering both carbon acquisition and export from canes in the present study clearly perturbed the typical pattern of TNC accumulation and storage in canes of both vines. However, the effect of reduced carbohydrate acquisition on the carbohydrate content of perennial cordon wood was only significant for the Hayward vines in the first season. Whereas, root carbohydrate content was substantially reduced in both seasons for Hayward vines and substantially reduced for Cabernet Sauvignon vines after two seasons of treatment. By comparison, reducing carbon acquisition of Gordo, DeChaunac and Seyval grapevines by shade had a larger and more consistent impact on the root and cane dry weight of vines than on trunk dry weights (Buttrose 1966; McArtney and Ferree 1999b). Furthermore, reducing carbohydrate acquisition of Semillon vines by defoliation generally reduced TNC in roots more than trunks and cordons (Smith and Holzapfel 2009). Together, the implication of these studies was that cordon carbohydrates were either a less mobile pool or a higher priority pool than those in either canes or roots. In the present study, modifying carbon acquisition in autumn did not have a large effect on carbohydrate pools in the cordon compared with that on the root and cane carbohydrate pools and thus, supports the hypothesis above. 184

221 5.4.2 Effects of autumn shade and girdling on subsequent flower production In the present study, reducing the carbohydrate acquisition of vines in autumn through shade influenced several vine attributes that, during the very early stages of shoot development in spring consequently influenced reproductive capacity and ultimately determined final yield. These attributes included a decrease in both total and floral budbreak, as well as fewer inflorescences per shoot on both vines. Reduced total budbreak and floral budbreak as well as a lower number of inflorescences number per shoot also occurred in several studies, when Hayward vines were grown under artificial shade or canes were grown under the canopy in autumn (Grant and Ryugo 1984; Fabbri et al. 1991; Snelgar et al. 1991; Snelgar et al. 1992). Furthermore, autumn shade over several cultivars also reduced bud fruitfulness in the following season (Hopping 1977; Sanchez and Dokoozlian 2005). However the carbohydrate status of annual or perennial organs in response to treatments was not measured in these studies and thus, a link between the effects on reproductive attributes in spring with the size of carbohydrate pools in autumn could not be made. By contrast, a strong relationship shown between root starch reserves, after defoliation in the previous season, and floral productivity of Chardonnay vines in the subsequent spring was demonstrated by Bennett et al. (2005). Similarly, starch mobilisation from Pinot Noir roots supported a large portion of shoot development in spring (Zapata et al. 2004b). However, Bennett et al. (2005) and Dayer et al. (2013) also indicated a relationship between the trunk starch pools and floral production in grapevines, which was not apparent in the present study. This suggested root reserves were more important than trunk reserves for both Hayward and Cabernet Sauvignon vines than occurred in Chardonnay vines. However, results from both shade and girdling treatments in the current study also suggested that cane reserves influence floral productivity of both vines. Girdling of canes to prevent export of carbohydrate from Hayward and Cabernet Sauvignon canes consistently increased the floral productivity of 185

222 both vines in the subsequent spring. This was evident from increased total and floral budbreak and inflorescence numbers per shoot. In an earlier study, autumn girdling of Hayward canes (Snelgar and Manson 1992) also increased the amount of budbreak and the number of inflorescences per shoot, but no data were presented on the carbohydrate composition of canes. However, in other grape cultivars, cane girdling in spring or around anthesis was used to examine effects on budbreak and shoot growth (Goffinet 2004; Eltom et al. 2013) or effects on fruit set and fruit quality (Coombe 1959; Brown et al. 1988; Caspari et al. 1998) rather than effects of carbohydrate supply in autumn on floral development in the following season. Thus, the conclusion that a positive relationship occurred between local cane carbohydrate and bud and floral development during early spring is advanced here for the first time. The effects of manipulating vine carbohydrate reserve accumulation in the previous season may be cumulative over several growing seasons (Candolfi- Vasconcelos and Koblet 1990; Smith and Holzapfel 2009; Holzapfel and Smith 2012; Dayer et al. 2013). This was certainly the case for Cabernet Sauvignon vines in the present study, as flower production was not reduced until after the third season and consistent with a progressive reduction in root TNC content in shaded vines over time. In contrast, effects of reserve manipulations on floral productivity of Hayward vines were apparent each season, but did not accumulate over the three seasons. A similar lack of cumulative treatment effects on vine floral productivity was also evident in an earlier study of Actinidia chinensis Planch. var. chinensis Hort16A, where vine carbohydrate reserves were restricted by high fruit numbers and by heavy pruning of vines over four seasons (Boyd and Barnett 2011) and also after heavy defoliation of Sultana grapevines in autumn (Scholefield et al. 1978). The lack of cumulative effect on floral productivity in Hayward vines in the present study may suggest that cane reserves had a greater impact on floral productivity or that there was adequate carbohydrate in the perennial organs to support floral development compared with that in Cabernet Sauvignon vines. However, there was a cumulative effect of treatments on the vegetative shoot growth of both Hayward and Cabernet Sauvignon 186

223 vines, with a 24% decrease in shoot growth in the spring after two seasons of autumn treatment. Similar reductions in shoot growth occurred after DeChaunac and Seyval grapevines were shaded in the previous season (McArtney and Ferree 1999b). Thus, for Hayward vines, the consistent effects occurring between seasons suggest that cane reserves had a major impact on floral productivity. However, the cumulative impact of reduced carbon acquisition on floral productivity of Cabernet Sauvignon and the vegetative growth of both vines suggest that root pools were also important. Thus, the conclusion that carbohydrate pools occurring in the vines in autumn were crucial to the development of the floral attributes and ultimately yield of both vines was strongly supported by this evidence Effect of girdling on other metabolites Phloem girdling not only limits carbohydrate export from canes but also affects the movement of other metabolites via the phloem (Goren et al. 2004). In the current study, TNC concentrations in girdled canes were higher (15%) than those in intact canes of both vines and consistent with little to no export occurring from the cane. Limiting export from canes also eliminated the negative effects of shade on subsequent inflorescence production of both vines and increased of floral productivity of girdled Cabernet Sauvignon canes. This result supports the conclusion that cane carbohydrate supply affected the floral productivity of Cabernet Sauvignon vines. However, the interaction between shade and girdling treatments on Hayward vines, whereby flowering was almost eliminated on shaded intact canes, more or less maintained on shaded girdled canes and markedly increased (140%) in exposed girdled canes, was altogether a much greater effect than occurred on the Cabernet Sauvignon vines. In addition, phloem girdling of Hayward canes also advanced budbreak and flowering. This result suggested that reducing transport via phloem may also have influenced the movement of other key metabolites in Hayward canes more so than in Cabernet Sauvignon canes. Certainly this difference in the response of the two vines was comparable with the earlier results (see Chapter 4) where girdling and defoliation during early shoot growth in the previous season almost 187

224 eliminated flowering in Hayward but had much less effect on the flowering of Cabernet Sauvignon vines in the subsequent spring. Phloem sap can transport sugars, amino acids, mrna, organic ions, proteins and phytohormones, including ABA and auxin, in Chardonnay grapevines (Glad et al. 1992). Other studies have shown an increase in ABA occurred in the leaves and apex of girdled grape shoots (Soar et al. 2004) and this may affect leaf gas exchange (Williams et al. 2000), reduced cytokinin export from roots of girdled grapevines (Skene 1972) and increased IAA and GA concentrations in girdled longan (Dimocarpus longan Lour.) shoots (Hegele et al. 2008). Any of these phytohormones could have influenced the floral development in vines (Vasconcelos et al. 2009). The results from the present study suggested that girdled Hayward canes may have influenced flower production by preventing signalling between shoots and roots from occurring via phloem over and above the effects of altered phloem transport on carbohydrate supply. By contrast, flower production of the Cabernet Sauvignon canes was apparently less sensitive to changes in other metabolites and simply dependent on the carbohydrate pool Summary To examine the hypothesis that floral development of both grape and kiwifruit in spring depended on reserves accumulated in the previous autumn, the carbohydrate supply was manipulated through vine shading to reduce carbon acquisition and girdling to reduce phloem transport from canes in autumn, and effects on floral development were monitored in the subsequent spring. Although the vines were grown under different management systems, this study examined basic physiological properties of vines. Reducing carbon acquisition to less than 20% of that in exposed vines markedly reduced the TNC content in both canes and roots in both vines. Phloem girdling to prevent carbohydrate export, by contrast, elevated the TNC of canes but not of axillary buds. This suggested that both canes and roots were the main storage sites for carbohydrate reserves during the second half of the growing season. Hayward vines were markedly more sensitive to reduced carbohydrate acquisition in autumn than Cabernet Sauvignon, with large reductions in flower numbers (60%) each season, but as effects were not 188

225 cumulative and could be prevented by girdling, non-carbohydrate factors within the cane also appeared to have a large impact on floral productivity. Furthermore, the strong interaction between light exposure and phloem transport in Hayward vines, ranging from complete elimination of flowering (by shading) to a magnitude increase in floral production (from girdling), also support the conclusion that, in addition to the effects on cane carbohydrate reserves, signalling via phloem had also been modified. By contrast, there was a cumulative effect of reduced carbohydrate acquisition in autumn on Cabernet Sauvignon vines, whereby shade only reduced flowering by 30% after three seasons of treatment, suggesting root reserves effected floral productivity of this cultivar. Reduced export via phloem also increased floral production on Cabernet Sauvignon vines by 30% and is consistent with the conclusion that increased cane carbohydrate content also supported floral productivity of this cultivar. To conclude, both cane and root reserves are the major sites of reserve storage in both vines in autumn; however, cane reserves appeared to be more important in Hayward vines but both cane and root reserves of Cabernet Sauvignon vines influenced floral development in spring. 189

226 6 Carbohydrate supply during floral sex organ development 6.1 Introduction Grape and kiwifruit vines differ from perennial tree fruit crops in that floral morphogenesis begins during the early stages of budbreak and continues over an extended period until anthesis, rather than occurring before the onset of winter dormancy (Pratt 1971; Brundell 1975c; Srinivasan and Mullins 1981; Polito and Grant 1984; Reinoso et al. 2002; Dennis 2003). Kiwifruit and grape inflorescences grow rapidly and accumulate significant amounts of carbohydrate between budbreak and anthesis (Kliewer 1965; Brundell 1975c; Lebon et al. 2004b). In grape flowers, differentiation of sexual organs occurs when shoots are about mm in length (Lebon et al. 2004b) and formation of the gynoecial plateau in kiwifruit flowers begins around DAB (Brundell 1975c), when shoots are around 300 mm in length (Piller et al. 1998). Floral primordia present in dormant grape and kiwifruit buds may either fail to develop or abort at any stage during floral morphogenesis (Brundell 1975c; Matthews and Anderson 1989; Skinner and Matthews 1989; Vasconcelos et al. 2009). In spring, rapid shoot growth results in significant competition for carbohydrate between leaves (the dominant sink) (Brundell 1975a; Mullins and Rajasekaran 1981), the shoot axis and the inflorescences (Keller and Koblet 1994; Piller et al. 1998; Foster et al. 2003; Lebon et al. 2004a). Interrupting carbohydrate supply to inflorescences during floral morphogenesis can cause abnormal flower development in kiwifruit (Piller et al. 1998) and result in coulure or abscission of up to 80% of grape flowers after anthesis (Mullins et al. 1992; Lebon et al. 2004b). Altering carbohydrate supply by shading (Ferree et al. 2001), defoliation (Coombe 1962; Caspari et al. 1998), girdling (Brown et al. 1988; Caspari et al. 1998; Piller et al. 1998), or pruning (May 2004) of shoots during the later stages of floral morphogenesis may also affect early berry growth. These effects of reduced carbohydrate supply during floral morphogenesis on subsequent fruit growth may be due to effects of reduced cell division during this period and hence reduced capacity of fruit to grow (Lai et al. 1990; Petrie and 190

227 Clingeleffer 2005) and/or effects of reduced carbohydrate supply on shoot development during floral morphogenesis and hence the ability of the shoots to produce carbohydrate to support subsequent fruit development. Carbohydrate to support early bud and shoot development comes from vine reserves laid down in the previous season (Buwalda and Smith 1990; McArtney and Ferree 1999b; Greer et al. 2003; Greer and Sicard 2009). However, new leaves become autotrophic from around 30 DAB (Greer and Jeffares 1998; Greer 2001; Greer and Sicard 2009; Greer et al. 2011). This transition from heterotrophy to autotrophy occurs during the critical stage of meiosis in grape flowers and ovule development in kiwifruit flowers (Piller and Meekings 1997; Lebon et al. 2004b), thus any delay in transiting to the autotrophic stage could be detrimental to floral morphogenesis. Therefore, the importance of carbohydrate supply during this extended period of shoot and flower development in spring is very important, as has been suggested earlier (Piller et al. 1998; Lebon et al. 2004b). However, the effects of altering photosynthate supply within shoots whilst the shoot was isolated from carbohydrate reserves has not apparently been examined thus far, nor more generally have the effect of treatments been followed through to fruit harvest. In this study, therefore, the hypothesis that the amount of carbohydrate supplied from newly developed leaves is critical to the development of flowers and their ability to develop into harvestable berries was examined. Although the vines were grown at different sites and under different management this study examined basic physiological properties of vines. Accordingly, treatments were applied to shoots in spring to modify their carbohydrate supply; carbohydrate supply from current photosynthesis was reduced by shade and import of carbohydrate into shoots was prevented by phloem girdling during the later stages of floral morphogenesis. Treatments were applied from the period when leaves become autotrophic until anthesis, and effects were monitored both during this period and throughout the remainder of the growing season. 191

228 6.2 Materials and methods Plant material This experiment was carried out on mature 25-year-old Hayward kiwifruit vines on Bruno rootstocks (2009/2010 and 2010/2011 seasons), six-yearold Shiraz grapevines on m470 rootstock both growing in Kerikeri, New Zealand (2009/2010 and 2010/2011 seasons), and 22-year-old Cabernet Sauvignon grapevines on SO4 rootstock growing in Kaikohe, New Zealand (2010/2011 season only). Cabernet Sauvignon vines were added in the final season for consistency with other chapters. Full details of vines and their management are given in Chapter 2. Briefly, Hayward vines were trained on a horizontal pergola, Shiraz vines were trained on a split canopy, Smart-Dyson vertical trellis and Cabernet Sauvignon vines were trained on a vertical shoot position trellis. All vines were cane pruned. Each Hayward vine had 16 canes, each 1.8m in length with 20 buds. Shiraz vines had two canes and Cabernet Sauvignon vines four canes, each 0.9 m length with 10 buds. As part of standard commercial vineyard practice, Shiraz and Cabernet Sauvignon shoots were summer pruned in December and January at the end of rapid shoot growth to maintain the shoot lengths of 1.5 m or 1.1 m respectively and control lateral shoot growth. Hayward shoots used in this experiment did not require pruning, as the growth of most shoots terminated naturally (Foster et al. 2003) prior to anthesis. Three leaves were plucked from immediately around Shiraz and Cabernet Sauvignon bunches in December for disease management, but leaf plucking was not used on Hayward shoots as berry skins are prone to sun damage (Sale and Lyford 1990). Hayward flowers were pollinated by bees in both seasons and were also hand pollinated with newly-opened male flowers in 2010/2011. Shiraz and Cabernet Sauvignon flowers were wind pollinated Experimental design and treatments Twenty vines of each cultivar were selected for this experiment. Treatments were applied to four shoots on each vine, comprising of a pair of fruiting shoots selected from the middle of two adjacent canes. A split plot design was used with light exposure, either fully exposed or shaded, randomly 192

229 allocated to each shoot pair within a vine. Within each pair, shoots were randomly assigned to be either phloem girdled or remain intact (subplots). Shoots were shaded by enclosing them in bags made from shade cloth (Solarshade TM, Permathene, New Zealand) (Figure 6.1C). Calibration measurements showed that the shade cloth reduced PAR levels by 85% and PFD to 83 cf. 991 µmol m -2, on average, and gave a slight (0.7%, 0.3 o C) increase in mean temperature (see Appendix 4). The underside of Hayward shade bags were opened at anthesis to allow bee access, but the shoot remained shaded. Phloem girdles were applied to shoots using small sidecutter pliers to completely sever the phloem near the base of each shoot (Figure 6.1D). A small piece of bamboo was then taped across the girdle to support the shoots, and this remained in place until anthesis, when the girdle had healed. Treatments were applied to vines in the early stages of shoot development (E - L 13-15) when shoots had 6-8 leaves and inflorescences were just visible. Shade was removed and girdles had healed at the end of anthesis (E - L 26) (Figures 6.1A and B). The timing of these phenological events varied slightly between vine cultivar so treatments were applied between DAB for Shiraz, DAB for Cabernet Sauvignon and DAB for Hayward vines (Table 6.1). Shiraz and Cabernet Sauvignon shoots carried one inflorescence per shoot and Hayward ± 0.1 inflorescences per shoot. 193

230 A B C D Figure 6.1: Spring shoots and applied treatments. A, Shiraz shoot, B, Hayward shoot, C, shaded Shiraz shoot pair and D, girdled Hayward shoot. Table 6.1: Dates and DAB that treatments were applied to Shiraz, Hayward and Cabernet Sauvignon shoots in 2009/2010 and 2010/2011. Vine cultivar Season Treatment period Start (DAB) End (DAB) Shiraz 2009/ October (25) 9 December (83) Shiraz 2010/ October (21) 6 December (70) Hayward 2009/ October (21) 30 November (61) Hayward 2010/ November (27) 6 December (62) Cabernet Sauvignon 2010/ November (31) 15 December (72) 194

231 6.2.3 Measurements Shoot measurements The length, basal diameter and node number of shoots were measured at the beginning of the experiment and at anthesis. The number of inflorescences on each Hayward shoot was counted at the beginning of anthesis. Shiraz and Cabernet Sauvignon shoots had one inflorescence per shoot that was covered with a net bag (0.5 mm mesh) during anthesis to collect flower caps (calyptras) which were then counted to determine total flower number. In 2010/2011, flower caps collected from Shiraz shoots were lost prior to counting when the laboratory was destroyed by fire. Flower numbers and the number of berries harvested at the end of fruit development were used to calculate the percentage of flowers that set harvestable fruit (Dry et al. 2010). A subsample of shoots (n = 8 in 2009/2010 and n = 4 in 2010/2011) from each treatment was harvested at anthesis and at fruit harvest. The leaf area of each shoot was measured using a LI3100 leaf area meter (Li-Cor, Lincoln, Nebraska, USA). The stem, leaves and flowers/berries from each shoot were weighed and then the tissues dried at 65 o C for 48 hours or until the tissue weight was constant, and the dry weight determined. At fruit maturity, all bunches/berries were harvested from the shoots. Each Shiraz and Cabernet Sauvignon bunch was weighed before berries were removed from the rachi, counted and weighed. A subsample of five typical berries was collected from each bunch and each berry was weighed then dissected into flesh and seeds, the number of seeds counted, and the fresh weight of flesh and seeds and dry weight of each component determined after drying at 65 o C for 48 hours. Mature Hayward fruit from each shoot were counted and weighed. A two mm thick equatorial slice was removed from one fruit from each shoot, weighed, dried for 48 hours at 65 o C and reweighed to determine fruit dry matter content. The pulp from the median weight Hayward fruit from each shoot was removed, immersed in a pectinase (Lafase 60, 5g/L) solution for 48 hours and then washed through a sieve (mesh size 30) to extract seeds. Seeds were dried at 65 o C for

232 hours, weighed and a subsample of 50 seeds counted and weighed in order to calculate total seed number per fruit and mean seed weight Gas exchange In the 2010/2011 season, gas exchange measurements were made using the Li-Cor 6400 open path photosynthesis system (Li-Cor, Lincoln, Nebraska, USA) fitted with a CO 2 mixer as described in Chapter 2. Light response curves were measured on four exposed and four shaded Hayward, Shiraz and Cabernet Sauvignon leaves at the end of the treatment period as described in Chapter 2. Gas exchange measurements were also made on leaves from four shoots from each treatment on each vine cultivar. All measurements were made between 10 am and 3 pm under natural light, temperature and vapour pressure deficit conditions. Leaves were enclosed in the chamber at their in situ orientation, readings were allowed to stabilise and then three consecutive measurements were made. Every second leaf along each shoot was measured, until leaves at the apical end were too small to be sealed within the chamber. Measurements were made at fortnightly intervals during the treatment period and again three weeks after shade treatments were removed from shoots Carbohydrate analysis In the 2010/2011 season, leaf and internode tissue was collected from the four shoots per treatment sampled at both anthesis and harvest. All tissues were immediately frozen in liquid nitrogen and stored at -80 o C before being analysed for carbohydrates using the method described in Chapter Net carbon balance The total carbon fixed by shoots per day during both the treatment period until anthesis and during the remainder of the season until fruit harvest was calculated using the method of Greer et al. (2011). Net photosynthetic rates were integrated over a day assuming that the measured rates represented a daily value during that period (Cartechini and Palliotti 1995). The leaf area per shoot per day was calculated using the measured values from harvested shoots and assuming a constant rate of leaf area development over the 196

233 treatment period and no change thereafter. The net CO 2 fixation per shoot was calculated from daily photosynthetic rates and the calculated leaf area per shoot, and these values were then summed over each period of development. Rates of respiration were calculated using values from light response data, the calculated leaf area per shoot and summing them over the shoots and each period in a similar manner to CO 2 fixation. These values were converted into net carbon acquisition using the molecular fraction of carbon in CO 2 (Greer et al. 2003). The amount of carbon that accumulated in each shoot over both the treatment period and the remainder of the growing season were determined by converting shoot biomass at the end of each period to elemental carbon assuming biomass contained 45% carbon (Walton and Fowke 1995). The net carbon balance over each period was then calculated as the difference between carbon acquisition and carbon accumulation in each shoot Statistical analysis All data was analysed using GenStat 14 th edition for Windows (VSN International Ltd., Hemel Hempstead, Hertfordshire, United Kingdom) using REML linear mixed models, assuming a fully randomised split plot design to analyse treatment differences in shoot growth, inflorescence measurements, berry set, yield, shoot composition, gas exchange and shoot carbon balance data. Differences between treatment means were determined at the appropriate level of significance, and estimated means and standard errors were presented. Due to the difference in flower number, fruit number and fruit weight between vines, these data were log transformed before statistical analysis and then back transformed for presentation. Relationships between carbon assimilation, light and gas exchange parameters were examined using the GLM procedure in Genstat. All data were plotted and nonlinear curves were fitted to light response data using a hyperbolic tangent function (Greer and Halligan 2001) with Origin version 8.5 (Originlab Corporation, Northhampton, MA, USA). See Chapter 2 for further details. 197

234 6.3 Results Shoot measurements Shoot growth The major portion of extension growth of Shiraz (80%), Hayward (63%) and Cabernet Sauvignon (73%) fruiting shoots occurred during the treatment period (not shown but see Figure 4.3 in Chapter 4) from 20 to 30 DAB until anthesis (60-80 DAB). There was no significant change in shoot length between anthesis and fruit harvest due to termination of growth in Hayward shoots and pruning of Shiraz and Cabernet Sauvignon shoots. The basal diameter of shoots from all vines also increased rapidly during the treatment period so that the shoot basal diameter at anthesis was 30% greater than when treatments were applied (Tables 6.2 and 6.3). During the remainder of the growing season, there was a further 5% increase in shoot basal diameter in both Shiraz and Cabernet Sauvignon shoots and a larger, 11% increase in the basal diameter of Hayward shoots. There was a significant interaction between vine cultivar*shade treatments on shoot length and node number in both seasons (Figure 6.2). Exposed Shiraz shoots were significantly longer than shaded shoots in both seasons (P < ; 1384 ± 63 to 1715 ± 71 mm cf ± 62 to 1126 ± 67 mm) and had more nodes (P < 0.001; 17.9 ± 0.7 to 20.9 ± 0.8 nodes cf ± 0.7 to 14.5 ± 0.8 nodes) than shaded shoots (Figures 6.2A - D; Tables 6.2 and 6.3). In 2010/2011, exposed Cabernet Sauvignon shoots also had significantly (P < 0.01) more nodes (20.2 ± 0.7) than shaded shoots (17.2 ± 0.7) (Figure 6.2D). By contrast, exposed Hayward shoots were significantly (P < 0.05) shorter (777 ± 55 mm cf. 939 ± 61 mm) than shaded shoots in 2009/2010 (Figure 6.2A), but there was no effect of shade on Hayward shoot extension in the 2010/2011 season. In addition, shade reduced the basal diameter of all shoots in 2010/2011, with highly significant (P < 0.001) effects on both Shiraz (exposed; 11.0 ± 0.2 mm cf. shaded; 9.2 ± 0.2 mm) and Hayward (exposed; 9.0 ± 0.2 mm cf. shaded; 8.0 ± 0.2 mm) and a significant difference (P < 0.05) for Cabernet Sauvignon shoots (exposed; 9.7 ± 0.2 mm cf. shaded; 9.1 ± 0.2 mm) (Figure 6.1E; Table 6.3). 198

235 The interaction between vine cultivar*girdle treatments significantly affected shoot growth (Figure 6.3). In 2010/2011, phloem girdling significantly (P < 0.05) decreased the length (intact; 1211 ± 56 mm cf. girdled; 1052 ± 58 mm) and node number (intact; 19.6 ± 0.6 cf. girdled; 17.8 ± 0.7) of Cabernet Sauvignon and also Hayward shoots (intact; 740 ± 58 mm cf. girdled; 477 ± 63 mm; intact; 16.4 ± 0.7 nodes cf. girdled; 11.9 ± 0.7 nodes) (Figures 6.3B and C; Table 6.3). By contrast, girdling significantly (P < 0.05) increased the number of nodes on Shiraz shoots (intact; 16.6 ± 0.9 cf. girdled; 18.9 ± 0.9) (Figure 6.3B). In 2009/2010, the basal diameter of intact shoots (9.2 ± 0.2 mm) was significantly (P < 0.05) smaller than that of girdled Shiraz shoots (9.8 ± 0.2 mm) (Table 6.2). By contrast, in the 2010/2011 season the basal diameter of intact shoots (9.83 ± 0.01 mm), averaged across all vines, was significantly (P < 0.001) larger than that of girdled shoots (8.83 ± 0.12 mm) (Table 6.3). 199

236 200 Table 6.2: Effects of altering carbohydrate supply to shoots prior to anthesis on the dimensions of Shiraz and Hayward shoots at anthesis in the 2009/2010 season. Data are means ± SEM, n = 20. Treatment Length (mm) Node number Diameter (mm) Shiraz Hayward Shiraz Hayward Shiraz Hayward Exposed 1368 ± ± ± ± ± ± 0.2 Exposed/girdle 1401 ± ± ± ± ± ± 0.3 Shade 1117 ± ± ± ± ± ± 0.3 Shade/girdle 1081 ± ± ± ± ± ± 0.3 P values Cultivar < <0.001 Shade <0.001 <0.001 Girdle Cultivar*Shade < Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

237 201 Table 6.3: Effects of altering carbohydrate supply to shoots prior to anthesis on the dimensions of Shiraz, Cabernet Sauvignon and Hayward shoots at anthesis. Data are means ± SEM, n= 20. Treatment Length (mm) Node number Diameter (mm) Shiraz Cabernet Sauvignon Hayward Shiraz Cabernet Sauvignon Hayward Shiraz Cabernet Sauvignon Hayward Exposed 1515 ± ± ± ± ± ± ± ± ± 0.2 Exposed/girdle 1914 ± ± ± ± ± ± ± ± ± 0.3 Shade 1192 ± ± ± ± ± ± ± ± ± 0.3 Shade/girdle 1060± ± ± ± ± ± ± ± ± 0.3 P values Cultivar <0.001 <0.001 <0.001 Shade <0.001 <0.001 <0.001 Girdle <0.001 Cultivar*Shade < Cultivar*Girdle < Shade*Girdle Cultivar*Shade*Girdle

238 Figure 6.2: Effects of vine cultivar and light (averaged across girdling treatments) on the dimensions of shoots at anthesis. A, length 2009/2010, B, length 2010/2011; C, number of nodes 2009/2010, D, number of nodes 2010/2011; E, basal diameter 2009/2010; F, basal diameter 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 40. Figure 6.3: Effects of vine cultivar and shoot girdling treatment (averaged across light treatments) on shoot dimensions at anthesis in 2010/2011. A, shoot length; B, number of nodes. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n =

239 6.3.2 Inflorescence and flower measurements Inflorescence weights varied between vine cultivar and season (Tables 6.4 and 6.5). The weight of Hayward inflorescences was almost two-fold larger (2.81 ± 0.26 g) in 2009/2010 than in the 2010/2011 season (1.44 ± 0.36 g). By contrast, Shiraz inflorescences were 30% larger in 2010/2011 (4.47 ± 0.38 g) than in 2009/2010 (3.41 ± 0.27 g). Therefore, the weights of Shiraz and Hayward inflorescences were similar in 2009/2010, but in 2010/2011 both Shiraz and Cabernet Sauvignon (2.66 ± 0.33 g) inflorescences were 4.7-fold and 1.9-fold heavier, respectively, than Hayward inflorescences. There were, however, no significant effects of treatments on inflorescence weight in either experimental season. Due to the nature of their inflorescences, Shiraz and Cabernet Sauvignon shoots carried almost a hundred-fold more flowers than Hayward shoots and, therefore, flower numbers were log transformed for statistical analysis and data back transformed for presentation (Tables 6.4 and 6.5). The interaction between vine cultivar*shade treatments on flower number per shoot was weakly significant (P = 0.057) in 2009/2010 and significant (P < 0.05) in 2010/2011. At anthesis in the 2009/2010 season, the number of flowers on exposed Hayward shoots (3.80 ± 0.17) was significantly (P < 0.001) higher than shaded shoots (2.57 ± 0.25) (Figure 6.4B) but there was no significant difference in flower numbers on Shiraz shoots (Figure 6.4A). By contrast, in 2010/2011, there was no significant difference in flower numbers between exposed and shaded Hayward shoots (Figure 6.4D) but flower numbers on exposed Cabernet Sauvignon shoots (389 ± 22) were significantly (P < 0.01) greater than on shaded shoots (281 ± 22) (Figure 6.4C). The effect of shoot girdling (averaged across vine cultivar and girdling treatment) on flower number per shoot was significant (P < 0.001) in 2009/2010, with 12% more flowers on average on intact compared with girdled shoots (averaged across vine cultivar). 203

240 Table 6.4: Effects of treatments to modify shoot carbohydrate prior to anthesis on inflorescence weight and flower number on Shiraz and Hayward shoots in 2009/2010. Data are means ± SEM, n = 8 inflorescence weight; n = 12 flower number per shoot. Treatment Inflorescence weight (g) Flower number / shoot Shiraz Hayward Shiraz Hayward Exposed 3.07± ± ± ± 0.15 Exposed/Girdle 2.81 ± ± ± ± 0.20 Shade 3.57 ± ± ± ± 0.29 Shade/Girdle 3.47 ± ± ± ± 0.22 P values Cultivar <0.001 Shade <0.001 Girdle <0.001 Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

241 205 Table 6.5: Effects of treatments to modify shoot carbohydrate prior to anthesis on inflorescence weight on Shiraz, Cabernet Sauvignon and Hayward shoots and flower number per Cabernet Sauvignon and Hayward shoot in 2010/2011. Data are means ± SEM, n = 4 inflorescence weight; n = 16 flower number per shoot. Treatment Inflorescence weight (g) Flower number / shoot Shiraz Cabernet Hayward Shiraz 1 Cabernet Hayward Sauvignon Sauvignon Exposed 2.94 ± ± ± ± ± 0.20 Exposed/girdle 3.82 ± ± ± ± ± 0.23 Shade 6.36 ± ± ± ± ± 0.31 Shade/girdle 4.77 ± ± ± ± ± 0.27 P values Cultivar <0.001 <0.001 Shade Girdle Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle In 2010/2011 Shiraz flower caps were destroyed in a laboratory fire. 205

242 Figure 6.4: Effects of vine cultivar and light treatment (averaged across girdling treatment) on flower numbers per shoot A, Shiraz 2009/2010; B, Hayward 2009/2010; C, Cabernet Sauvignon 2010/2011; D, Hayward 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ Fruit set, seed number and seed weight In the 2009/2010 season, the percentage of flowers that produced harvestable fruit was similar on Shiraz (71.4 ± 2.3%) and Hayward (72.3 ± 2.5%) shoots (Tables 6.6 and 6.7). In the following season, Hayward flowers set a higher percentage of harvestable fruit (88.1 ± 2.8%) than Cabernet Sauvignon shoots (40.6 ± 4.5%), but fruit set could not be calculated for Shiraz shoots due to the loss of flower caps in a laboratory fire. In both seasons, there was a significant interaction (P < 0.001; 2009/2010; P < 0.05; 2010/2011) between vine cultivar*shade on fruit set and retention (Tables 6.6 and 6.7). In 2009/2010, fruit set and retention on exposed Hayward shoots (92.2 ± 3.2%) was significantly (P < 0.001) greater than on shaded shoots (52.4 ± 3.9%), but there were no significant effects of shade on Shiraz shoots (Figure 6.5A). In the following season, fruit set and 206

243 retention on exposed Hayward shoots (92.6 ± 3.9%) was again significantly (P < 0.05) different but only 10% greater than on shaded shoots (83.7 ± 4.0%). The difference in fruit set and retention on shaded Hayward shoots between seasons was probably due to the effects of shade on bee pollination in 2009/2010 which was amended by hand pollination in 2010/2011. By contrast, fruit set and retention on exposed Cabernet Sauvignon shoots (32.2 ± 6.2%) was significantly (P < 0.05) lower than that on shaded shoots (49.0 ± 6.6%) (Figure 6.5B). Girdling of shoots did not affect fruit set and retention of any cultivar in either season. Seed number per berry (Figure 6.6) and mean seed dry weight (Figure 6.7) differed between vines because of the berry structure, as Shiraz and Cabernet Sauvignon berries contained few (1-4) large seeds (dry weight mg) and Hayward berries contained a large number ( ) of small seeds (dry weight mg). Therefore, seed numbers per berry and mean seed weight per berry were log-transformed for statistical comparison. There was a significant interaction between vine cultivar and shade treatment on both seed number and mean seed weight per berry (P < 0.001) in 2009/2010 and a weakly significant interaction (P = and respectively) in 2010/2011 (Tables 6.6 and 6.7). In 2009/2010, the mean weight of seeds from berries on shaded Shiraz shoots (18.6 ± 0.4 mg) was significantly (P < 0.001) greater than those from exposed shoots (13.9 ± 0.4 mg). In the following season, the mean weight of seeds from berries on both shaded Shiraz (P < 0.05; 22.8 ± 1.0 mg) and Cabernet Sauvignon (P < 0.001; 36.7 ± 1.0 mg) shoots were greater than those from berries on exposed shoots (Shiraz 19.8 ± 1.0 mg; Cabernet Sauvignon 32.0 ± 0.9 mg). By contrast, there were no effects of treatments on the mean weight of Hayward seeds in either season. However, the number of seeds per berry on shaded Hayward shoots was significantly (P < 0.001) lower and only 60% (741 ± 60) of that in berries from exposed shoots (1210 ± 49) in 2009/2010 and 93% (1191 ± 17) of that in berries from exposed Hayward shoots (1280 ± 15) in 2010/2011. There were no significant effects on seed numbers in Shiraz and Cabernet Sauvignon berries in either season. 207

244 208 Table 6.6: The effects of treatments to modify carbohydrate supply of shoots prior to anthesis on the percentage of flowers that produced harvestable berries, seed number per berry and mean seed weight in 2009/2010. Data are means ± SEM, n= 12. Treatment Fruit set and retention (%) Seed number / berry Mean seed dry weight (mg) Shiraz Hayward Shiraz Hayward Shiraz Hayward Exposed 70.5 ± ± ± ± ± ± 0.04 Exposed/Girdle 73.3 ± ± ± ± ± ± 0.05 Shade 69.9 ± ± ± ± ± ± 0.05 Shade/Girdle 72.2 ± ± ± ± ± ± 0.07 P values Cultivar <0.001 <0.001 Shade <0.001 <0.001 <0.001 Girdle Cultivar*Shade <0.001 <0.001 <0.001 Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

245 209 Table 6.7: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the percentage of Shiraz, Cabernet Sauvignon and Hayward flowers that set harvestable berries, seed number per berry and mean seed weight in 2010/2011. Data are means ± SEM, n = 16. Treatment Fruit set and retention (%) Seed number/ berry Mean seed dry weight (mg) Shiraz 1 Cabernet Sauvignon Hayward Shiraz Cabernet Sauvignon Hayward Shiraz Cabernet Sauvignon Hayward Exposed ± ± ± ± ± ± ± ± 0.03 Exposed/Girdle ± ± ± ± ± ± ± ± 0.04 Shade ± ± ± ± ± ± ± ± 0.04 Shade/Girdle ± ± ± ± ± ± ± ± 0.03 P values Cultivar <0.001 <0.001 <0.001 Shade Girdle Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade* Girdle In 2010/2011 Shiraz flower caps were destroyed in a laboratory fire. 209

246 Figure 6.5: Effects of vine cultivar and light treatment (averaged across girdling treatment) on fruit set (%) and retention in A, 2009/2010; B, 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/2011. Figure 6.6: Effects of vine cultivar and light treatment (averaged over girdling treatment) on seed number per berry of A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between Shiraz/Cabernet Sauvignon and Hayward data. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/

247 Figure 6.7: Effects of vine cultivar and light treatment (averaged over girdling treatment) on mean seed dry weight of A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between Shiraz/Cabernet Sauvignon and Hayward data. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/ Yield Overall, Hayward shoots produced higher yields (303 ± 16 g) than Shiraz shoots (196 ± 13 g) in 2009/2010, but this was reversed in 2010/2011, when Shiraz yields per shoot (552 ± 26 g) were higher than those of Hayward (376 ± 24 g) or Cabernet Sauvignon shoots (279 ± 24 g) (Tables 6.8 and 6.9). The increase in yield of Shiraz shoots was due to more (295 ± 12) and larger berries (1.87 ± 0.08 g) in 2010/2011 season than the 2009/2010 season (147 ± 5 berries per shoot and 1.37 ± 0.08 g per berry). Treatments applied to shoots from DAB until anthesis had significant carryover effects on fruit yield at harvest. The interaction between vine cultivar*shade*girdle on yield was significant (P < 0.05) in 2009/2010 and weakly significant (P = 0.077) in 2010/2011 (Tables 6.8 and 6.9). Thus, shading significantly (P < 0.001) decreased the yield of both intact and 211

248 girdled Hayward shoots by 50% compared with that of exposed shoots in 2009/2010 (Figure 6.8A). In addition, the yield of exposed and girdled Hayward shoots (359 ± 29 g) was significantly (P < 0.05) reduced by 100 g compared with exposed intact shoots (460 ± 27 g). In 2009/2010, there was no effect of treatments on the yield of Shiraz shoots (Figure 6.8A). However, in the following 2010/2011 season, the yield of shaded and girdled Shiraz shoots (347 ± 49 g) was significantly (P < 0.001) lower (56%) than all other Shiraz shoots (mean 620 ± 51 g) (Figure 6.8B). The yield of shaded and girdled Cabernet Sauvignon shoots was also significantly (P < 0.05) different and only 64% of exposed intact shoots (Figure 6.8B). In addition, the yield of girdled Hayward shoots (210 ± 37 g) was significantly (P < 0.01) reduced and only 60% of that of intact shoots (348 ± 31 g), regardless of shoot light exposure (Figure 6.8B). Berry number per shoot differed between vines as a consequence of the flowering behaviour of the vines, with Hayward shoots producing 1-6 large berries ( g berry -1 ) per shoot, and both Shiraz and Cabernet Sauvignon shoots produced small berries (1-2 g berry -1 ) per shoot (Tables 6.8 and 6.9). Therefore, berry numbers and mean berry weight per shoot were log-transformed for statistical comparison and back-transformed for presentation. Shading shoots significantly (P < 0.001) reduced the number of berries produced by the shoots, 35% on average across vines in the 2009/2010 season and by 25% in the 2010/2011 season (Figure 6.9). Berry numbers per shoot were not significantly affected by girdling treatments in 2009/2010, but in 2010/2011 berry numbers were significantly (P<0.01) different and 25% lower on girdled compared to intact shoots (Figure 6.10). In the 2009/2010 season, there was a significant interaction (P < 0.001) between vine cultivar*shade treatments on berry weight (Figure 6.11). For example, shading Shiraz shoots significantly (P < 0.01) increased the mean berry weight (1.53 ± 0.08 g) compared with that on exposed shoots (1.24 ± 0.05 g). By contrast, shading Hayward shoots significantly (P < 0.001) reduced the mean weight of berries by 25%, to 77.3 ± 10.4 g compared with 212

249 102.4 ± 3.8 g on exposed shoots. The girdling treatments had no significant effects on mean berry weight in 2009/2010. However, in the 2010/2011 season, there was a significant (P < 0.05) vine cultivar*shade*girdle interaction on mean berry weight (Figure 6.11). Thus, berries from shaded intact Shiraz shoots (2.01 ± 0.05 g) were significantly (P < 0.05) heavier by 14% than those from shaded girdled shoots (1.76 ± 0.07 g). In addition, the mean weight of berries from both exposed intact (1.64 ± 0.04 g) and shaded girdled Cabernet Sauvignon shoots (1.66 ± 0.09 g) were significantly (P < 0.05) larger by 16% than berries from exposed girdled shoots (1.42 ± 0.05 g). By contrast, for the Hayward vines, berries from exposed intact shoots (104.7 ± 2.8 g) were significantly (P < 0.05) larger by 19% larger than those from shaded girdled shoots (88.1 ± 4.3 g). 213

250 214 Table 6.8: Effects of altering carbohydrate supply to Shiraz and Hayward kiwifruit shoots prior to anthesis on yield (total fresh berry weight) per shoot and its components mean berry weight and berry number per shoot in the 2009/2010 season. Data for berry number and mean berry weight were log transformed for analysis and back transformed for presentation. Data are means ± SEM, n = 12. Treatment Yield shoot -1 (g) Berry number Mean berry weight (g) Shiraz Hayward Shiraz Hayward Shiraz Hayward Exposed 195 ± ± ± ± ± ± 3.6 Exposed/girdle 209 ± ± ± ± ± ± 4.0 Shade 221 ± ± ± ± ± ± 6.9 Shade/girdle 157 ± ± ± ± ± ± 14.0 P values Cultivar <0.001 <0.001 <0.001 Shade <0.001 < Girdle Cultivar*Shade < <0.001 Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

251 215 Table 6.9: Effects of treatments on Shiraz, Cabernet Sauvignon and Hayward shoots prior to anthesis on yield (total fresh berry weight) per shoot and its components mean berry weight and berry number per shoot in the 2010/2011 season. Data for berry number and mean berry weight were log transformed for analysis and back transformed for presentation. Data are means ± SEM, n = 16. Treatment Yield shoot -1 (g) Berry number Mean berry weight (g) Shiraz Cabernet Sauvignon Hayward Shiraz Cabernet Sauvignon Hayward Shiraz Cabernet Sauvignon Hayward Exposed 618 ± ± ± ± ± ± ± ± ±2.8 Exposed/girdle 619 ± ± ± ± ± ± ± ± ± 2.2 Shade 621 ± ± ± ± ± ± ± ± ± 4.4 Shade/girdle 347 ± ± ± ±11 86 ± ± ± ± ± 4.3 P values Cultivar <0.001 <0.001 <0.001 Shade < Girdle <0.001 < Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

252 Figure 6.8: Effects of vine cultivar, light and girdling treatment on fruit yield per shoot in A, 2009/2010; B, 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, n = 12, 2009/2010, n = 16, 2010/2011. Figure 6.9: Effects of vine cultivar and light treatment (averaged over girdling treatment) on berry number per shoot. A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Note the differences in scale between Shiraz/Cabernet Sauvignon and Hayward berry numbers. Data are means ± SEM, n = 12, 2009/2010, n = 16, 2010/

253 Figure 6.10: Effects of vine cultivar and girdling treatment (averaged across light treatment) on berry numbers per shoot in 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale for Shiraz and Cabernet Sauvignon compared to Hayward. Data are means ± SEM, n = 12, 2009/2010, n = 16, 2010/2011. Figure 6.11: Effects of vine cultivar, light and girdling treatment on mean berry weight. A, Shiraz 2009/2010; B, Shiraz 2010/2011; C, Hayward 2009/2010; D, Hayward 2010/2011; E, Cabernet Sauvignon 2010/2011. Data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale for Shiraz and Cabernet Sauvignon compared to Hayward. Data are means ± SEM, n = 12, 2009/2010; n = 16, 2010/

254 6.3.5 Shoot dry weight and leaf area The distribution of biomass within shoots of the three vines at anthesis and harvest was similar (Tables 6.10 and 6.11). Leaf growth was largely completed by anthesis with no significant changes in the dry weight of leaves per shoot (averaged over vines) between anthesis and fruit harvest in both seasons (11.1 ± 0.4 to 11.5 ± 0.8 g cf ± 1.0 to 13.5 ± 1.8 g) (Tables 6.10 and 6.11). However, the dry weight of the whole shoot increased significantly in both seasons (P < 0.001) by 5-6-fold between anthesis and harvest (18.9 ± 0.7 to 20.9 ± 1.3 g cf ± 4 to ± 13.9 g). This was due to marked increases in both the dry weight of shoot stems (7.0 ± 0.4 to 8.0 ± 3.1 g cf ± 1.4 to 28.3 ± 4.7 g; averaged over vines) and the inflorescence / fruit components (0.82 ± 0.04 to 1.42 ± 0.12 g cf ± 2.9 to 93.0 ± 8.9 g; averaged over vines). Contributions of individual organs to shoot biomass were, therefore, consistent over the two seasons. By contrast, vegetative proportions of the stem biomass decreased between anthesis and fruit harvest. Leaves made up 57% and the stem 38% of the shoot biomass at anthesis but only 11% and 23% respectively at fruit harvest. However, the reproductive component of shoots increased from 5% at anthesis to 65% at harvest. In both experimental seasons, shading shoots from between 20 and 30 DAB until the end of anthesis significantly (P < 0.001) reduced shoot biomass at anthesis to about 60% of that in exposed shoots (13.7 ± 1.0 to 16.5 ± 1.5 g cf ± 1.0 to 25.4 ± 1.5 g) when averaged over vine cultivar and girdling treatment (Figures 6.12E and F). There was also a significant shade*vine cultivar interaction (P < 0.01) on leaf dry weight at anthesis in both experimental seasons. Shading developing Shiraz shoots significantly reduced leaf biomass at anthesis to between 49% and 64% of that of exposed shoots (P < < 0.01; 8.36 ± 0.76 to 6.25 ± 1.37 g cf ± 0.81 to ± 1.37 g) (Figures 6.12A and B). For Shiraz shoots leaves generally made up around 53% of the total shoot biomass at anthesis, but in 2010/2011 shading reduced the leaf biomass to 40% of the shoot weight. Furthermore, shading Hayward shoots significantly (P < 0.001) reduced leaf dry weight at anthesis by similar amounts (60%) in both seasons (6.27 ±

255 to 8.32 ± 1.37 g cf ± 0.73 to ± 1.37 g). On Hayward shoots, leaves made up around 73% of the total shoot biomass at anthesis, but on shaded shoots leaves made up 53% of shoot biomass in 2009/2010 and 70% in 2010/2011. In 2009/2010, the dry weight of Shiraz stems at anthesis (6.73 ± 0.73 g) was significantly (P < 0.001) different and also only 60% of that in exposed shoots (11.36 ± 0.78 g) (Figure 6.12C), but in both treatments stems made up around 45% of total biomass. However, there were no significant effects of shade on the stem biomass of Hayward shoots at anthesis in 2009/2010, or on stem biomass of any cultivar in the 2010/2011 season. In the 2009/2010 season, the shade treatments (averaged over vines) significantly (P < 0.01) reduced the biomass of inflorescences (0.70 ± 0.06 g; 5.1% of total biomass) to 75% of that on exposed shoots (0.94 ± 0.06 g; 3.6% of total biomass) (Figure 6.12D) but there was no significant difference in the 2010/2011 season. Furthermore, there were no significant effects of girdling treatments on shoot biomass at anthesis in either season. It was noteworthy that effects of treatments persisted after the shade was removed and the girdles had healed, with significant differences in shoot biomass apparent at fruit harvest. For instance, in 2009/2010, the total biomass of shaded shoots (69.7 ± 6.0 g averaged across vines) was significantly (P < 0.001) different and reduced by 30% compared with the biomass of exposed shoots (102.7 ± 5.8 g) (Figure 6.13G). However, in the following 2010/2011 season, the vine cultivar*shade*girdle interaction on total shoot biomass was significant (P < 0.05) (Figure 6.13H). Thus, previously shaded and girdled Shiraz shoots produced significantly (P < 0.01) less total biomass (50%) than shoots from other treatments. In addition, the biomass of shaded and girdled Cabernet Sauvignon shoots was significantly (P < 0.05) lower by 66% of that of exposed and intact shoots. By contrast, there were no significant effects of treatments on the total biomass of Hayward shoots (mean 78.7 ± 9.6 g) at harvest in 2010/

256 There was also a significant vine cultivar*shade interaction on leaf biomass (P < 0.05) at harvest in 2009/2010 (Figure 6.13A). For example, the leaf biomass (5.83 ± 0.81 g; 8.3% of total shoot biomass) of shaded Shiraz shoots was significantly (P < 0.01) lower and only 35% of that of exposed shoots (16.35 ± 1.76 g; 16.6% of total shoot biomass), probably due to early leaf fall. However, the shade treatment had no effect on leaf biomass of Hayward shoots. Furthermore in 2010/2011, there was a large and highly significant (P < 0.001) reduction in leaf biomass at harvest, whereby, when averaged across vines, the leaf dry weight of shaded shoots (7.42 ± 1.21 g; 9.2% of total shoot biomass) was less than 50% of that of exposed shoots (16.62 ± 1.22 g; 11.6% of total shoot biomass) (Figure 6.13B). In both experimental seasons, the vine cultivar*shade interaction had a significant (P < 0.01) effect on stem biomass at harvest (Figures 6.13C and D). The dry weight of shaded Shiraz shoots was significantly (P < 0.001) lower in both seasons, and 55% or 36% of that of exposed shoots (17.00 ± 2.84 to 17.1 ± 4.6 g; 24.2 and 12.6% of total shoot biomass cf ± 2.91 to 48.8 ± 4.4 g; 31.4 and 21.8% of total shoot biomass). However, there were no significant effects of treatment on stem biomass of the other vines in either season with the average weight of Hayward stems in both seasons 14.9 ± 2.5 to 7.3 ± 1.3 g and Cabernet Sauvignon in 2010/2011 was 31.7 ± 3.9 g. In terms of fruit biomass, the vine cultivar*shade interaction was also significant (P < 0.01) in 2009/2010 (Figure 6.13E). The dry weight of fruit on shaded Hayward shoots (37.8 ± 5.9 g; 55.3% of total shoot biomass) was significantly (P < 0.001) reduced, (by 52%) compared with that of exposed shoots (79.3 ± 4.9 g; 69.6% of total shoot biomass). However, there were no significant effects of shade on fruit biomass of Shiraz shoots. By contrast in 2010/2011, the vine cultivar*shade*girdle interaction on fruit biomass was significant (P < 0.05) (Figure 6.13F). Thus, previously shaded and girdled Shiraz shoots produced significantly (P < 0.05) 50% less fruit (fruit 80.7% of total shoot biomass) than Shiraz shoots from other treatments (fruit 73.6% of total shoot biomass). In addition, the fruit biomass of shaded and girdled 220

257 Cabernet Sauvignon shoots (fruit 42.1% of total shoot biomass) was significantly (P < 0.05) reduced, by about 80% compared with that of exposed and intact shoots (fruit 59.6% of total shoot biomass). However, there were no significant effects of treatment on the fruit biomass of Hayward shoots in the 2010/2011 season. 221

258 222 Table 6.10: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the dry weight of Shiraz and Hayward leaves and stems at anthesis and at fruit harvest in the 2009/2010 season. Data are means ± SEM, n = 8. Treatment Anthesis Harvest Leaf (g) Stem (g) Inflorescence (g) Total (g) Leaf (g) Stem (g) Fruit (g) Total (g) Shiraz Exposed 13.0 ± ± ± ± ± ± ± ± 12.5 Exposed/girdle 13.0 ± ± ± ± ± ± ± ± 11.7 Shade 9.19 ± ± ± ± ± ± ± ± 14.8 Shade/girdle 7.52 ± ± ± ± ± ± ± ± 11.0 Hayward Exposed 17.0 ± ± ± ± ± ± ± ± 9.5 Exposed/girdle 16.5 ± ± ± ± ± ± ± ± 10.0 Shade 6.31 ± ± ± ± ± ± ± ± 11.0 Shade/girdle 6.23 ± ± ± ± ± ± ± ± 12.5 P values Cultivar < < Shade < < < < < < Girdle Cultivar*Shade < Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

259 223 Table 6.11: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the dry weight of Shiraz, Cabernet Sauvignon and Hayward shoots at anthesis and at fruit harvest in the 2010/2011 season. Data are means ± SEM, n = 4. Treatment Anthesis Harvest Leaf (g) Stem (g) Inflorescence (g) Total (g) Leaf (g) Stem (g) Fruit (g) Total (g) Shiraz Exposed 11.3 ± ± ± ± ± ± ± ± 22 Exposed/girdle 13.9 ± ± ± ± ± ± ± ± 22 Shade 6.99 ± ± ± ± ± ± ± ± 19 Shade/girdle 5.51 ± ± ± ± ± ± ± ± 16.4 Cabernet Sauvignon Exposed 13.4 ± ± ± ± ± ± ± ± 23 Exposed/girdle 11.0 ± ± ± ± ± ± ± ± 10.8 Shade 9.57 ± ± ± ± ± ± ± ± 6 Shade/girdle 8.35 ± ± ± ± ± ± ± ± 11.9 Hayward Exposed 21.9 ± ± ± ± ± ± ± ±13 Exposed/girdle 19.5 ± ± ± ± ± ± ± ± 10.6 Shade 8.10 ± ± ± ± ± ± ± ± 6.0 Shade/girdle 8.53 ± ± ± ± ± ± ± ± 9.3 P values Cultivar < <0.001 <0.001 <0.001 Shade < <0.001 < Girdle Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

260 Figure 6.12: Effects of vine cultivar and light treatment (averaged across girdling treatment) on the dry weight of shoot components at anthesis. A, leaf 2009/2010; B, leaf 2010/2011; C, stem 2009/2010; D, stem 2010/2011; E, inflorescence 2009/2010; F, inflorescence 2010/2011; G, total shoot weight 2009/2010; H, total shoot weight 2010/2011. E, F and G, show light treatment means averaged across vine cultivar and girdling treatment. In A, B, C, D and H data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between the shoot components. Data are means ± SEM, A, C and F n = 16, E and G n = 32, B, D and H n =

261 Figure 6.13: Effects of vine cultivar and light treatment (averaged over girdling treatment) on the dry weight of shoot components at harvest A, leaf 2009/2010; B, leaf 2010/2011; C, stem 2009/2010; D, stem 2010/2011; E, fruit 2009/2010; F, fruit 2010/2011; G, total shoot weight 2009/2010; H, total shoot weight 2010/2011. B and G, show light treatment means averaged across vine cultivar and girdling treatment. In A, C - F and H data are shown for Shiraz, Hayward and Cabernet Sauvignon. Note the differences in scale between the shoot components. Data are means ± SEM, F and H n = 4, D n = 8, A, C, and E n = 16, B n = 24, G n = 32. At anthesis, the leaf area of Hayward shoots was 27% less than that of Shiraz shoots in 2009/2011 (1440 ± 79 cm 2 cf ± 84 cm 2 ) and 23% less than Cabernet Sauvignon shoots in 2010/2011 (1510 ± 119 cm 2 cf ± 119 cm 2 ) (Tables 6.12 and 6.13). However, at fruit harvest at the end of the season, there were no differences in the leaf area of shoots between vines in either experimental year. Shade treatments reduced the leaf area of shoots throughout the growing season. At anthesis, previously shaded shoots (averaged across cultivar and girdling treatment) had significantly (P < < 0.01) smaller leaf area (20-35%) in both seasons (1351 ± 81 to 1568 ± 97 cm 2 cf ± 83 to 1970 ± 97 cm 2 ) (Tables 6.12 and 6.13). At the end of the 2009/2010 season, there was a significant vine cultivar*shade interaction, whereby the leaf area of shaded Shiraz shoots was significantly (P < 0.01) 53% smaller than on exposed shoots (822 ± 180 cm 2 cf ± 172 cm 2 ) but there was no difference in the leaf area of previously shaded or exposed Hayward shoots. In the following 2010/2011 season, shade 225

262 treatments (averaged across vine cultivar and girdling treatment) resulted in a significant (P < 0.001) reduction (35%) in leaf area compared that on continuously exposed shoots (932 ± 104 cm 2 cf ± 106 cm 2 ). Table 6.12: Effects of treatments to modify shoot carbohydrate supply prior to anthesis on the leaf area of Shiraz and Hayward shoots at anthesis and at fruit harvest in the 2009/2010 season. Data are means ± SEM, n = 8. Treatment Leaf area (cm 2 ) Anthesis Harvest Shiraz Hayward Shiraz Hayward Exposed 2391 ± ± ± ± 172 Exposed/girdle 2398 ± ± ± ± 179 Shade 1692 ± ± ± ± 188 Shade/girdle 1385 ± ± ± ± 210 P values Cultivar < Shade < Girdle Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

263 227 Table 6.13: Effects of treatments to modify shoot carbohydrate supply on the leaf area of Shiraz, Hayward and Cabernet Sauvignon shoots at anthesis and at harvest in the 2009/2010 seasons. Data are means ± SEM, n = 4. Treatment Leaf area (cm 2 ) Anthesis Shiraz Hayward Cabernet Sauvignon Harvest Shiraz Hayward Cabernet Sauvignon Exposed 1876 ± ± ± ± ± ± 268 Exposed/girdle 2213 ± ± ± ± ± ± 268 Shade 1701 ± ± ± ± ± ± 268 Shade/girdle 1529 ± ± ± ± ± ± 268 P values Cultivar Shade Girdle Cultivar*Shade Cultivar*Girdle Shade*Girdle Cultivar*Shade*Girdle

264 6.3.6 Shoot total carbohydrate concentrations and content At anthesis in 2010/2011, the TNC concentrations in the leaves and stems of the three vines were similar (Table 6.14). However, the TNC content in the stems of both Cabernet Sauvignon (0.72 ± 0.05 g shoot -1 ) and Shiraz (0.55 ± 0.06 g shoot -1 ) shoots were higher than in Hayward shoots (0.27 ± 0.06 g shoot -1 ), due to differences in shoot biomass at anthesis. At the end of the growing season, the TNC concentrations of Cabernet Sauvignon leaves (185.5 ± 7.7 mg g -1 DW) were higher than either Shiraz (113.1 ± 8.7 mg g -1 DW) or Hayward (85.1 ± 7.7 mg g -1 DW) leaves. At this time, Cabernet Sauvignon leaves also had a higher TNC content (1.85 ± 0.15 g shoot -1 ) compared with Hayward leaves (1.05 ± 0.15 g shoot -1 ) (Table 6.14). By contrast, at the end of the season, the TNC concentration in the stems of Shiraz shoots were higher (137 ± 3.1 mg g -1 DW) than in Cabernet Sauvignon (120.5 ± 2.8 mg g -1 DW) or Hayward stems (109.6 ± 2.8 mg g -1 DW). In addition, the TNC content of both Cabernet Sauvignon (3.78 ± 0.30 g shoot -1 ) and Shiraz (3.15 ± 0.33 g shoot -1 ) stems were greater than those of Hayward stems (0.78 ± 0.11 g shoot -1 ). Shading of the shoots caused a significant (P < 0.001) decrease in the TNC concentration in leaves (83.8 ± 6.2 mg g -1 DW) compared with leaves of exposed shoots at anthesis (119.1 ± 6.2 mg g -1 DW) averaged across all vines and girdling treatments (Figure 6.14A). These differences contributed to a significant (P < 0.05) interaction between vine cultivar*shade on the TNC content of leaves at anthesis (Figure 6.15A). For instance, shade significantly reduced the carbohydrate content of Shiraz (0.68 ± 0.12 g shoot -1 ) and Hayward leaves (0.55 ± 0.13 g shoot -1 ) compared with those leaves from exposed shoots ( Shiraz 1.53 ± 0.26 g shoot -1 ; Hayward 2.27 ± 0.26 g shoot -1 ). By contrast, there were no significant effects of treatment on the carbohydrate concentrations in leaves at the end of the growing season. However, as the reduction in leaf biomass due to shade persisted until fruit harvest, so the reduction in TNC content in leaves from shaded shoots (0.82 ± 0.12 g shoot -1 ) (averaged over vine cultivar and girdling treatment) was significantly (P < 0.001) lower than leaves of exposed shoots (1.94 ± 0.13 g shoot -1 ) at this time (Figure 6.15B). 228

265 At anthesis, there was a significant interaction between vine cultivar*shade*girdle on the carbohydrate concentrations in stems (Figure 6.14B). The TNC concentration of shaded Shiraz stems (56.4 ± 4.0 mg g -1 DW) was significantly (P < 0.001) reduced compared with exposed stems (74.4 ± 2.3 mg g -1 DW), regardless of girdling treatment. By contrast, for the Cabernet Sauvignon shoots only shaded and intact shoots (52.6 ± 5.1 mg g - 1 DW) had significantly (P < 0.001) lower TNC concentrations than the stems from other treatments (mean 70.8 ± 2.5 mg g -1 DW). Hayward shoots that had been exposed and girdled (95.2 ± 4.6 mg g -1 DW) had significantly (P < 0.001) higher TNC concentrations than shoots from other treatments (mean 55.4 ± 2.9 mg g -1 DW). In terms of the TNC content of stems at anthesis, there was a significant (P < 0.001) reduction in the content of shaded stems (0.42 ± 0.07 g shoot -1, averaged over vine cultivar and girdling treatment) compared with that in exposed shoot stems (0.61 ± 0.05 g shoot -1 ) (Figure 6.15C). At the end of the growing season, the stems of shaded shoots had significantly higher (P < 0.05) TNC concentrations (125.8 ± 2.3 mg g -1 DW, averaged over vines and girdling treatments) than exposed shoots (119.0 ± 2.5 mg g -1 DW) (Figure 6.14C). However, there was a significant (P < 0.05) vine cultivar*shade interaction on stem TNC content at harvest. Shaded Shiraz shoots had significantly (P < 0.001) and markedly lower TNC content (1.95 ± 0.42 g shoot -1 ) than exposed shoots (4.35 ± 0.51 g shoot -1 ), but there were no significant effects in other vines (Figure 6.15D). In addition, TNC concentrations in the stems of previously girdled shoots (125.8 ± 2.5 mg g -1 DW, averaged over vine cultivar and light treatment) were slightly but significantly (P < 0.05) higher than those of intact shoots (119.0 ± 2.3 mg g -1 DW) Figure 6.16A). However, the reduction in biomass of girdled stems resulted in lower total stem TNC content in girdled (1.90 ± 0.26 g shoot -1 ) compared with intact shoots (3.24 ± 0.24 g shoot -1 ) at fruit harvest (Figure 6.16B). 229

266 230 Table 6.14: Effects of treatments to modify carbohydrate supply applied prior to anthesis on the TNC concentration and content of leaves and stems of shoots at anthesis and at fruit harvest in the 2010/2011 season. Data are means ± SEM, n = 4. Treatment TNC at anthesis TNC at fruit harvest Concentration (mg g -1 DW) Content (g shoot -1 ) Concentration (mg g -1 DW) Content (g shoot -1 ) Leaf Stem Leaf Stem Leaf Stem Leaf Stem Shiraz Exposed 127 ± ± ± ± ± ± ± ± 0.53 Exposed/girdle 145 ± ± ± ± ± ± ± ± 0.83 Shade 88.7 ± ± ± ± ± ± ± ± 0.44 Shade/girdle 95.7 ± ± ± ± ± ± ± ± 0.19 Cabernet Sauvignon Exposed 99.6 ± ± ± ± ± ± ± ± 0.70 Exposed/girdle 121 ± ± ± ± ± ± ± ± 0.67 Shade 78.0 ± ± ± ± ± ± ± ± 0.61 Shade/girdle 108 ± ± ± ± ± ± ± ± 0.33 Hayward Exposed 107 ± ± ± ± ± ± ± ± 0.19 Exposed/girdle 116 ± ± ± ± ± ± ± ± 0.07 Shade 64.4 ± ± ± ± ± ± ± ± 0.11 Shade/girdle 67.5 ± ± ± ± ± ± ± ± 0.06 P values Cultivar <0.001 <0.001 <0.001 <0.001 <0.001 Shade <0.001 <0.001 < < Girdle < <0.001 Cultivar*shade Cultivar*girdle Shade*girdle Cultivar*shade*girdle

267 Figure 6.14: Effects of light treatment (averaged over vine cultivar and girdling treatment A, B, and D) on the TNC concentration of shoot components in 2010/2011. A, leaf at anthesis; B, leaf at harvest; C, stem at anthesis; D, stem at harvest. In C, the bars represent data from Shiraz, Hayward and Cabernet Sauvignon. Data are means ± SEM, C n = 4, A, B and D n =

268 Figure 6.15: Effects of vine cultivar and light treatment (averaged over girdling treatment A and D) and of light treatment (averaged over vine cultivar and girdling treatment B and C) on the TNC content of shoot components in 2010/2011. A, leaf at anthesis; B, leaf at harvest; C, stem at anthesis; D, stem at harvest. In A, and D the bars represent data from Shiraz, Hayward and Cabernet Sauvignon. Note the different scale in D. Data are means ± SEM, A and D n = 8, B and C n = 24. Figure 6.16: Effects of girdling treatment (averaged across vine cultivar and light treatment) on A, the TNC concentration and B, the TNC content of stems at harvest in 2010/2011. Data are means ± SEM, n =

269 6.3.7 Gas exchange Light responses Photosynthetic light responses of exposed and shaded leaves of each vine cultivar were compared at the end of the treatment period (Figure 6.17) and there were significant interactions between vine cultivar*shade on the light saturated maximum rate of photosynthesis, rate of respiration and the apparent photon yield (Table 6.15). The largest reduction in the light saturated rate of photosynthesis was between exposed and shaded Cabernet Sauvignon leaves, where rates of shaded leaves were 44% of those of exposed leaves (P < 0.001) whereas on Hayward leaves (P < 0.001) and Shiraz leaves (P < 0.01) rates on shaded leaves were consistently 60% of those of exposed leaves (Table 6.15). Rates of respiration in shaded leaves were also significantly (P < 0.01) only 60% of those in exposed Hayward leaves, significantly (P < 0.05) 48% of those in exposed Cabernet Sauvignon leaves and only 28% of those in exposed Shiraz leaves. The apparent photon yield was also significantly (P < 0.01) 58% lower in shaded Hayward leaves and 43% lower in shaded Cabernet Sauvignon leaves than that in exposed leaves. Averaged across leaves of all vines, the mean photon flux density that the maximum photosynthetic rate occurred in shaded leaves (752 ±76 µmol (photons) m -2 s -1 ), that is light saturation, was also significantly (P < 0.001) only 70% of that in exposed (1060 ± 60 µmol (photons) m -2 s -1 ) leaves (Table 6.15). 233

270 Figure 6.17: Photosynthetic light responses for the youngest mature leaves on shaded ( ) and exposed ( ) leaves from A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots at anthesis. Data for are means ± SEM, n =

271 Table 6.15: Photosynthetic light responses for Hayward, Shiraz and Cabernet Sauvignon leaves from fully exposed shoots at the end of the treatment period in the 2010/2011 season. P max is the light saturated maximum rate of photosynthesis, Rs is the rate of respiration, φ app is the apparent (CO 2 limited) photon yield and PFD sat is the photon flux density at which photosynthesis was light saturated. Data for each vine cultivar are mean values ± SEM, n = 4. Vine cultivar Light Rs exposure P max µmol (CO 2 ) m -2 s -1 µmol (CO 2 ) m -2 s -1 Φ app mol (CO 2 ) mol (photons) -1 PFD sat µmol (photons) m -2 s -1 Shiraz Exposed 10.1 ± ± ± ± 159 Cabernet Sauvignon Shade 6.3 ± ± ± ± 123 Exposed 15.1 ± ± ± ± 55 Shade 6.7 ± ± ± ± 187 Hayward Exposed 14.1 ± ± ± ± 61 P values Shade 8.7 ± ± ± ± 88 Cultivar < Shade <0.001 <0.001 < Cultivar*Shade Photosynthetic rates of treated leaves During the treatment period, the mean photosynthetic rates of leaves of the three vines (averaged across all treatments) were similar ( Shiraz 5.1 ± 0.1 µmol m -2 s -1 ; Cabernet Sauvignon 4.9 ± 0.1 µmol m -2 s -1 and Hayward 5.5 ± 0.1 µmol m -2 s -1 ) (Figure 6.18). However, there was a significant (P < 0.05) interaction between vine cultivar*shade*girdle on the mean photosynthetic rates. For all vines, the photosynthetic rates of exposed leaves were over five-fold higher than rates for shaded leaves (8.7 ± 0.1 µmol m -2 s -1 cf. 1.6 ± 0.1 µmol m -2 s -1 ). Furthermore, the effect of girdling was only significant for Shiraz, where the photosynthetic rates of leaves from exposed intact shoots (8.92 ±.032 µmol m -2 s -1 ) were significantly (P < 0.05) 13% higher than rates of leaves from exposed girdled shoots (7.77 ± 0.31 µmol m -2 s -1 ). Three weeks after treatments were removed from shoots (~90 DAB), there was still a significant (P < 0.001) interaction between vine cultivar*shade on 235

272 photosynthetic rates. Thus, photosynthetic rates of leaves from previously shaded Shiraz (7.2 ± 0.5 µmol m -2 s -1 ) and Cabernet Sauvignon (5.40 ± 0.49 µmol m -2 s -1 ) shoots had increased once the shade was removed, but were still significantly (P < 0.05) 30% lower than the rates of leaves from continuously exposed shoots ( Shiraz 10.2 ± 0.4 µmol m -2 s -1 ; Cabernet Sauvignon 8.5 ± 0.4 µmol m -2 s -1 ) (Figure 6.18). By contrast, three weeks after treatments were removed the photosynthetic rates of leaves from continuously exposed Hayward shoots (4.9 ± 0.5 µmol m -2 s -1 ) had decreased and rates for leaves from previously shaded shoots increased (4.9 ± 0.5 µmol m -2 s -1 ) so that the two were equivalent. In addition, there was no significant effect of girdling treatments on the photosynthetic rates of leaves from any shoots once the girdles had healed. Figure 6.18: Effects of shade and girdling treatments on mean photosynthetic rates of shaded ( ) and exposed ( ) leaves of A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots in the 2010/2011 season. Solid lines represent intact shoots and dashed lines girdled shoots. Treatments were removed from Shiraz vines 70 DAB, Cabernet Sauvignon vines 72 DAB and Hayward vines 62 DAB. Data were collected from every second leaf on each shoot and the mean rates calculated per shoot. Data are means ± SEM, n =

273 Stomatal conductance of treated leaves The average stomatal conductance of Shiraz (0.147 ± mol m -2 s -1 ), Cabernet Sauvignon (0.139 ± mol m -2 s -1 ) and Hayward (0.139 ± mol m -2 s -1 ) leaves were also similar during the treatment period (Figure 6.19). However, there was a significant interaction (P < 0.001) between vine cultivar*shade*girdle on stomatal conductance. The stomatal conductance of exposed Shiraz (0.202 ± mol m -2 s -1 ) and Cabernet Sauvignon leaves (0.187 ± mol m -2 s -1 ), were significantly (P < 0.001) higher and almost double that of shaded leaves (0.090 ± mol m -2 s -1 and ± mol m -2 s -1 ) (Figure 6.19). However, girdling significantly reduced the stomatal conductance of exposed Cabernet Sauvignon leaves (P < 0.05; girdled; ± 0.010; shoots cf. intact; ± mol m -2 s -1 ) and also exposed Hayward leaves (P < 0.001; girdled; ± cf. intact; ± mol m -2 s -1 ) possibly due to metabolite accumulation in the leaves. When shade treatments were removed from shoots, rates of stomatal conductance increased over the three week post treatment period. However, there were still significant interactions between vine cultivar*shade (P < 0.001) and between vine cultivar*girdle (P < 0.05) treatments on the leaf stomatal conductance. For instance, the stomatal conductance of leaves on both previously shaded Shiraz (0.203 ± mol m -2 s -1 ) and Cabernet Sauvignon (0.251± mol m -2 s -1 ) leaves were still significantly lower than those from continuously exposed leaves (P < 0.001; Shiraz ± mol m -2 s -1 ; P < 0.05; Cabernet Sauvignon ± mol m -2 s -1 ). By contrast, the stomatal conductance of leaves from intact Hayward shoots (0.147 ± ± mol m -2 s -1 ) was significantly (P < 0.01) lower than that of leaves from girdled shoots (0.187 ± ± mol m -2 s -1 ). 237

274 Figure 6.19: Effects of shade and girdling treatments on stomatal conductance of shaded ( ) and exposed ( ) leaves of A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots in the 2010/2011 season. Solid lines represent intact shoots and dashed lines girdled shoots. Treatments were removed from Shiraz vines 70 DAB, Cabernet Sauvignon vines 72 DAB and Hayward vines 62 DAB. Data were collected from every second leaf on each shoot and the mean rate calculated per shoot. Data are means ± SEM, n = Transpiration rates of treated leaves During the treatment period, the mean transpiration rate of both Shiraz (2.38 ± 0.07 mmol m -2 s -1 ) and Cabernet Sauvignon (2.31 ± 0.06 mmol m -2 s -1 ) leaves were higher than transpiration rates of Hayward leaves (2.15 ± 0.06 mmol m -2 s -1 ) (Figure 6.20). There was also a significant (P < 0.001) interaction between vine cultivar*shade*girdle on leaf transpiration rates during the treatment period. For example, for both Shiraz and Cabernet Sauvignon, the transpiration rates were significantly (P < 0.001) and 50% lower on shaded than exposed leaves, but girdling treatments had no significant effect on transpiration rates. For Hayward leaves, there was a similar 50-60% reduction in transpiration rates between leaves from the exposed intact shoots and leaves from both shaded intact and shaded girdled shoots. In addition, the transpiration rates of leaves on exposed and 238

275 girdled shoots were significantly (P < 0.001) lower and only 70% of those of leaves on exposed intact shoots. Three weeks after shade treatments were removed, there was still a significant interaction between vine cultivar*shade*girdle on leaf transpiration rates (Figure 6.20). The transpiration rate of previously shaded Shiraz leaves (4.23 ± 0.14 mmol m -2 s -1 ) remained significantly (P < 0.05) lower than the rate for exposed leaves (5.29 ± 0.12 mmol m -2 s -1 ). However, for Hayward shoots, leaves on previously shaded and girdled shoots had transpiration rates that were significantly higher (P < 0.01) 25-50% higher than those of leaves from other treatments. Figure 6.20: Effects of shade and girdling treatments on mean transpiration rates of shaded ( ) and exposed ( ) leaves of A, Shiraz ; B, Cabernet Sauvignon ; and C, Hayward shoots in the 2010/2011 season. Solid lines represent intact shoots and dashed lines girdled shoots. Treatments were removed from Shiraz vines 70 DAB, Cabernet Sauvignon vines 72 DAB and Hayward vines 62 DAB. Data were collected from every second leaf on each shoot and the mean rate calculated per shoot. Data are means ± SEM, n =

276 Assimilation model From the general linear model (GLM) procedure, the factors that had a large and significant (P < 0.001) effect on photosynthetic assimilation rates (A) were photon flux density (PFD), internal CO 2 concentration (c i ) and stomatal conductance (g s ). The effect of vine cultivar was significant (P < 0.001) but explained less than 2% of the variance and was, therefore, excluded from the model. Shade treatment also had a significant effect, but was already accounted for in the model by the photon flux density measurements. The effect of girdling treatment was not significant and thus was not included in the model. When the key factors of photon flux density, internal CO 2 concentration and stomatal conductance were fitted to an additive model, a large portion of the variance was accounted for (r 2 = 0.80). However, using a multiplicative model: A = f(pfd x g s x c i ) improved the amount of variance accounted for (r 2 = 0.99) (Table 6.15). In this model, photon flux density explained 49% of the variance in photosynthetic assimilation, stomatal conductance and internal CO 2 concentration each explained 16%, and the interaction between the two another 17% of the variance. Together, the interactions between photon flux density and stomatal conductance or internal CO 2 concentration and across all three parameters accounted for only 2% of the variance. Table 6.16: Coefficients and statistics of fitting a multiplicative model of photon flux density, stomatal conductance and internal CO 2 concentration to photosynthetic assimilation data. The percentage of error mean square (EMS) accounted for by each of the terms and the interaction between terms is presented. The model was highly significant (P < 0.001), n = Intercept PFD (% EMS) g s (% EMS) c i ( (% EMS) g s.c i (% EMS) 3.7 ± ± (49) ± 0.8 (16) -8.6 ± (17) ± 0.01 (18) 240

277 6.3.8 Net carbon balance Shoot carbon acquisition over both the treatment period (up to anthesis) and during the remainder of the growing season from anthesis to fruit harvest, was influenced by a significant vine cultivar*shade*girdle interaction (P < 0.05) (Tables 6.16 and Figure 6.21). During the treatment period, carbon acquisition of exposed shoots of all vines ( Shiraz 23.7 ±1.4 g shoot -1 ; Cabernet Sauvignon 22.3 ±1.4 g shoot-1; Hayward 21.7 ±1.4 g shoot -1 ) was significantly (P < 0.001) different and about 8-10-fold higher than that of shaded shoots ( Shiraz 2.7 ± 1.4 g shoot -1 ; Cabernet Sauvignon 2.7 ± 1.4 g shoot -1 ; Hayward 2.7 ± 1.4 g shoot -1 ). These differences in carbon acquisition were influenced by effects of shade on both the photosynthetic rates of the shoots (Figure 6.18) and the leaf area of shoots (exposed 0.20 ± 0.01 m 2 cf. shaded 0.15 ± 0.01 m 2 ) (Table 6.13). In addition, girdling significantly reduced the carbon acquisition of exposed Cabernet Sauvignon (P < 0.01) and exposed Hayward shoots (P < 0.05); by approximately 30% compared with that of exposed and intact shoots, although effects on either leaf area or photosynthetic rates themselves were not significant. After treatments were removed from shoots at anthesis, carbon acquisition of exposed shoots ( Shiraz ±18.3 g shoot -1 ; Cabernet Sauvignon 62.1 ± 2.5 g shoot -1 ; Hayward 49.1 ± 6.5 g shoot -1 ) remained significantly (P < 0.001) 2-7-fold greater than that of shaded shoots ( Shiraz 45.5 ±7.5 g shoot -1 ; Cabernet Sauvignon 25.8 ± 2.1 g shoot -1 ; Hayward 6.5 ± 4.8 g shoot -1 ) over the remainder of the season (Figure 6.21). This was because of effects on both the photosynthetic rates and shoot leaf area. However, in contrast to effects before anthesis, the carbon acquisition of previously girdled and exposed Shiraz shoots (139.1 ± 18.9 g shoot -1 ) was significantly (P < 0.05) higher than that of exposed and intact shoots (103.6 ± 28.7 g shoot -1 ) during the rest of the growing season. Again, this was attributed to a significantly greater leaf area (girdled 0.26 ± 0.2 m 2 shoot -1 cf. intact 0.15 ± 0.2 m 2 shoot -1 ) of shoots (Table 6.13). At anthesis, there was a weakly significant interaction (P = 0.082) between vine cultivar*shade*girdle on the carbon accumulated structurally as biomass of the shoots (Table 6.16 and Figure 6.21). Shading shoots significantly (P < 241

278 0.001) reduced the amount of carbon accumulated as biomass of both Cabernet Sauvignon and Hayward shoots, by 40% and 65% respectively. However, shade did not reduce the amount of carbon accumulated in the intact Shiraz shoots but there was a significant (P < 0.001) and three-fold reduction in the carbon accumulated in shaded and girdled shoots compared with that in exposed girdled shoots. The interaction between vine cultivar*shade*girdle on the carbon accumulated as biomass of the shoots was also significant (P < 0.05) at harvest (Table 6.16 and Figure 6.21). For Shiraz shoots, the previous girdling treatments again resulted in a significantly (P < 0.05) lower amount of carbon accumulated as biomass, by 46%, in shaded shoots but an increase in biomass accumulation of 40% in exposed shoots. Thus, at the end of the season there was a highly significant (P < 0.001) and three-fold difference between the amount of carbon accumulated as biomass in previously shaded and girdled shoots compared with exposed and girdled Shiraz shoots. During the treatment period, there was a weakly significant (P < 0.069) interaction between vine cultivar*shade*girdle on the net carbon balance of shoots, that is, the net difference between the amount of carbon acquired through photosynthesis and the amount of carbon sequestered as biomass (Table 6.16 and Figure 6.21). For all vines, there was a highly significant difference (P < 0.001) between the negative net carbon balance that occurred with shaded shoots and the positive carbon balance of exposed shoots. However, girdling of shaded Cabernet Sauvignon shoots significantly (P < 0.05) reduced the magnitude of the negative carbon balance to 44% of that in intact but shaded shoots. In addition, girdling also significantly reduced the net carbon balance of both exposed Cabernet Sauvignon (P < 0.001) and exposed Hayward shoots (P < 0.01) by 58% and 47% respectively. Shade and girdling treatments applied to shoots prior to anthesis continued to affect the net carbon balance of shoots throughout the remainder of the season (Figure 6.21). The carbon balance of previously shaded shoots 242

279 generally remained negative and significantly (P < 0.001) lower than the positive balance that occurred with continuously exposed shoots (Figure 6.21). The only exception to this was the positive net carbon balance that occurred with previously shaded and girdled Cabernet Sauvignon shoots. This resulted from a similar acquisition of carbon but comparatively low accumulation of carbon as stem and fruit biomass of these shoots when compared with that of shaded and intact shoots (Table 6.11). On shaded shoots, previous girdling treatments also significantly (P < ) increased the net carbon balance by g compared with intact shoots, even when the net carbon balance was negative. However, on continuously exposed Cabernet Sauvignon shoots, previous girdling treatments increased the net carbon balance by 1.8-fold compared with intact shoots (P < 0.001) because of the effect of girdling on shoot carbon biomass accumulation (Table 6.11). Previous girdling reduced the net carbon balance of exposed Hayward shoots by 60% (P < 0.001) but, in contrast, was caused by effects on net carbon acquisition. 243

280 244 Figure 6.21: Effects of treatments on the net carbon acquisition A, at anthesis and B, fruit harvest; carbon in shoot biomass C, at anthesis and D, fruit harvest; and the shoot net carbon balance E, at anthesis and F, fruit harvest in 2010/2011. In each case, the bars represent data from Shiraz, and Cabernet Sauvignon and Hayward. Data are means ± SEM, n =

281 245 Table 6.17: Effects of altering carbohydrate supply to Shiraz, Cabernet Sauvignon and Hayward kiwifruit shoots prior to anthesis on the net carbon balance of shoots during the treatment period (from DAB to fruit set DAB) and from fruit set until fruit harvest ( DAB) in the 2010/2011 season. Data are means ± SEM, n = 4. Treatment Treatment period Anthesis to harvest Net carbon acquisition (g shoot -1 ) Carbon as biomass (g shoot -1 ) Net carbon balance (g shoot -1 ) Net carbon acquisition (g shoot -1 ) Carbon as biomass (g shoot -1 ) Net carbon balance (g shoot -1 ) Shiraz Exposed ± ± ± ± ± ± 3.10 Exposed/girdle ± ± ± ± ± ± 4.27 Shade 3.56 ± ± ± ± ± ± 1.74 Shade/girdle 1.80 ± ± ± ± ± ± 2.91 Cabernet Sauvignon Exposed ± ± ± ± ± ± 1.82 Exposed/girdle ± ± ± ± ± ± 5.77 Shade 1.52 ± ± ± ± ± ± 3.31 Shade/girdle 1.86 ± ± ± ± ± ± 2.08 Hayward Exposed ± ± ± ± ± ± 2.64 Exposed/girdle ± ± ± ± ± ± 2.20 Shade 0.80 ± ± ± ± ± ± 5.33 Shade/girdle 2.80 ± ± ± ± ± ± 4.27 P values Cultivar <0.001 <0.001 <0.001 Shade <0.001 <0.001 <0.001 < <0.001 Girdle <0.001 Cultivar*Shade <0.001 Cultivar*Girdle <0.001 Shade*Girdle < <0.001 Cultivar*Shade*Girdle

282 6.4 Discussion The objective of the work described in this chapter was to examine the hypothesis that the amount of carbohydrate supplied during floral morphogenesis, from 20 to 30 DAB when leaves generally become autotrophic through to when anthesis occurs, determines whether the development of flowers is completed and if flowers can produce harvestable berries. To test this hypothesis, the carbohydrate supply to Shiraz, Cabernet Sauvignon and Hayward shoots was manipulated by reducing photosynthesis rates through shade and by restricting shoot import via phloem girdling. Although the vines were grown at different sites and under different management, this study examined basic physiological properties of vines. At the end of the treatment period, the TNC content of leaves and stems was reduced by 13-76% on shoots where photosynthesis had been reduced by shade compared with exposed shoots of the three vines. These effects of shade on leaf carbohydrate content persisted throughout the remainder of the growing season until fruit harvest. By contrast, reducing shoot carbohydrate import by phloem girdling had no effect on the carbohydrate content of shoots at anthesis; however, at the end of the growing season, the TNC content of girdled stems was less than that of intact stems. Thus the objective to alter the carbohydrate pools during the critical stages of floral development was achieved by shade treatments and to a lesser extent by girdling. The effect of the carbohydrate pools on floral morphogenesis was the next part of the hypothesis tested and although effects of modified carbohydrate supply on floral morphogenesis were not consistent between vines or seasons, flower numbers were generally reduced by 10-33% and berry numbers were consistently reduced by 10-35% on shoots that had lower carbohydrate supply prior to anthesis. Thus, the hypothesis appeared to be supported and the evidence is now explored in greater detail. Shade reduced carbon acquisition prior to anthesis by 8-10-fold through decreases in both photosynthetic rates and leaf area expansion. This reduced total carbohydrate pools in shaded shoots at anthesis by about 50% across vines, through decreases in both shoot biomass and carbohydrate 246

283 concentrations. Leaf area expansion was completed at anthesis, and therefore, carbon acquisition of previously shaded shoots was 2-10-fold lower than that of exposed shoots for the remainder of the season. Thus, at harvest carbohydrate pools of leaves from shoots shaded prior to anthesis remained 50% less than those of exposed shoots for all vines, but only pools in the stems of shaded Shiraz shoots were reduced by 55% compared with those in exposed shoots. Earlier studies of both Hayward (Grant and Ryugo 1984) and Semillon (Greer et al. 2011) shoots shaded throughout the growing season showed a similar 2-10-fold reduction in carbon acquisition was associated with a 25-50% reduction in shoot dry weights at harvest but carbohydrate data were not presented. Preventing carbohydrate import/export by girdling shoots during the pre-anthesis period did not affect shoot carbohydrate pools at anthesis. However, at harvest the carbohydrate pools in stems of previously girdled shoots were reduced by 40% compared with those in intact stems across vines. Reductions in carbohydrate accumulation were the result of less stem growth prior to anthesis and hence, lower capacity to accumulate carbohydrate in late summer and autumn. Previous studies that have investigated girdling of Sauvignon Blanc either earlier in the season before budbreak (Eltom et al. 2013) or later in the season at fruit set (Caspari et al. 1998) have also reported effects on shoot growth and carbon accumulation. Hence, both the expansion of leaf area and growth of the shoot axis prior to anthesis are critical factors determining the acquisition and accumulation of carbohydrate during the remainder of the season Floral morphogenesis The effects of reduced carbohydrate supply on flower numbers per shoot and the subsequent development of flowers into harvestable berries were not consistent between vines or seasons. In particular, reduced carbon acquisition prior to anthesis decreased the flower numbers on Hayward but not Shiraz shoots in 2009/10, and on Cabernet Sauvignon but not Hayward shoots in 2010/2011. In addition, girdling reduced flower numbers on both Hayward and Shiraz shoots by 12% in 2009/2010 but had no effect in 2010/2011. At the end of the season, berry numbers were consistently 247

284 reduced on shoots subjected to reduced carbon acquisition prior to anthesis but reduced phloem transport only reduced berry numbers of Hayward and Cabernet Sauvignon shoots in 2010/2011. Earlier studies have shown that decreasing photosynthate supply to Sauvignon Blanc (Caspari et al. 1998) or Muller-Thurgau (Keller and Koblet 1994) inflorescences by reducing light or by defoliation at anthesis caused necrosis and abortion of inflorescence branches and inflorescence necrosis. However, effects of shade on fruit set were variable, for example when shade was applied from five days prior to anthesis to five weeks after anthesis fruit set and berry numbers were reduced in grape cultivars like Chambourcin and Seyval but not in DeChaunac or Vidal Blanc (Ferree et al. 2001). Shading Muscat and Grenache grapevines during flowering also had inconclusive effects on fruit set (Coombe 1959). However, altering carbohydrate supply by defoliation, topping or girdling several grape cultivars (Coombe 1959; Caspari et al. 1998) or Hayward kiwifruit (Piller et al. 1998) generally influenced fruit set. Together these results confirm that carbohydrate supply can alter both floral development and fruit set of Cabernet Sauvignon and Hayward but the effects may vary between season and cultivar. In contrast to these varying results, it has been well established that manipulation of grapevine reserves in the previous season influenced both inflorescence number and the number of flowers per inflorescence can be influenced by the supply of reserve carbohydrate to shoots in spring (Candolfi-Vasconcelos and Koblet 1990; Duchene et al. 2003; Bennett et al. 2005; Holzapfel et al. 2006; Smith and Holzapfel 2009) (see also Chapter 5). It was also apparent that the timing and degree of the dependence on reserve supply appeared to vary seasonally. For instance, up to 30 DAB and after the very early stages of shoot development had occurred, the carbon acquisition of shoots had a much greater impact on shoot reproductive capacity than the supply from reserves Shoot development prior to anthesis A large proportion of the annual vegetative growth of fruiting shoots of all vines was completed during the treatment period before anthesis (

285 DAB) had occurred, with berry development occurring thereafter. A similar temporal separation between vegetative and berry growth has been shown to occur in both Semillon grapes (Greer et al. 2011) and Hayward kiwifruit (Greer et al. 2003) and appeared to be a characteristic attribute of vines. During this pre-anthesis period of the current study, leaf area of untreated shoots increased by 3-fold on Hayward shoots, 6-fold on Shiraz shoots and 10-fold on Cabernet Sauvignon shoots. However, the vegetative growth of shoots prior to anthesis was highly regulated by the carbohydrate supply. For instance, shade applied to shoots early in development consistently decreased the total biomass of the shoots at anthesis, by about 50%, across all vines and in both seasons. This decrease was largely a result of decreased leaf biomass. Elsewhere, shading of Semillon grapevines throughout the growing season also reduced leaf biomass (Greer et al. 2011). It can be concluded, therefore, that reduced carbohydrate supply during early shoot development had a major impact on shoot growth, particularly leaf biomass. Consistent with this conclusion, shading of shoots prior to anthesis caused a 5-fold reduction in the photosynthetic rates of leaves of all vines. This is much greater than the 50% reduction observed in previous studies of shade on grape (Cartechini and Palliotti 1995; Greer et al. 2011) and Hayward kiwifruit (Greer 2001). The difference between these studies may be related to the different levels of imposed shade. For example, in the present study shade treatments reduced light exposure by 85% (83 µmol m -2 ) compared with 70% (391 µmol m -2 ) and 40% (785 µmol m -2 ) shade of Sangiovese leaves (Cartechini and Palliotti 1995), 70% shade (400 µmol m -2 ) of Semillon leaves (Greer et al. 2011) and between Hayward leaves grown in a controlled environment at a light intensity of 1100 compared with 250 µmol m -2 s -1 by Greer (2001). Light response curves for both shaded and well exposed leaves in the current study and these earlier investigations have all demonstrated a marked reduction in photosynthetic rates below 200 µmol m -2 (Cartechini and Palliotti 1995; Greer and Halligan 2001; Greer et al. 2011). It is also likely that the imposed shade reduced the photosynthetic contribution from inflorescences (Lebon et al. 2005), although this was not 249

286 measured the current study. Thus, the reduced vegetative growth that occurred in shaded shoots of the present study was consistent with markedly reduced carbon acquisition rates. In addition to lower rates of photosynthesis, reduced carbon acquisition by the shaded shoots was also attributable to a reduction in shoot leaf area. In the current study the leaf area of shaded shoots of all vines was reduced by 20-35% compared with that of exposed shoots, through a decrease in both leaf number and average leaf area. This contrasts with earlier studies where, for example, Greer et al. (2011) found that the leaf area of Semillon vines exposed to 70% shade was 20% larger than the leaf area of exposed vines and in Hayward vines exposed to low light (to 250 µmol m -2 s -1 ) where the maximum size of leaves increased but the total leaf area of the shoot was unchanged (Greer 2001). These increases in leaf area and decrease in leaf thickness in previous studies are more typical shade leaf response (Givnish 1988) than those from the present study. However, at the very low level of light (15%) intercepted by shaded shoots in this study and the concomitant reduction in photosynthesis, it would appear that the shoots were unable to generate sufficient resources to support cell division and expansion to either expand individual leaves or maintain an adequate leaf appearance rate equivalent to that of exposed shoots, with the outcome that light capture was further impeded in a negative feedback. The combined effects of lower photosynthetic rates and the lower leaf area on the shaded shoots of all vines resulted in markedly reduced rates of carbon acquisition (10-fold lower) compared with that of exposed shoots during the period of the treatment. Furthermore, a 3-fold reduction in carbon acquisition for Shiraz and Cabernet Sauvignon shoots and a 6-fold reduction for Hayward shoots occurred over the remainder of the growing season. This is much greater than the 30% reduction in rates of carbon acquisition of shoots from shaded Semillon vines (70% shade) recorded over the growing season by Greer et al. (2011) and the 1.3-fold reduction in Hayward vines grown in low light (250 cf µmol m -2 ) over the growing season (Greer 2001). The greater effects of shade on the rates of carbon 250

287 acquisition in the present study were consistent with the conclusion that the vines were treated with a higher degree of shade (85%) than elsewhere and this had a major impact on carbon acquisition. By contrast, phloem girdling had little effect on the accumulation of biomass in exposed shoots of the three vines during the pre-anthesis treatment period. This suggested that after the first days of growth when the treatments were applied, the contribution of photosynthesis to the shoot carbohydrate supply was much greater than import of reserves from perennial organs and hence biomass accumulation increased. Consistent with this conclusion, well exposed shoots were autotrophic within DAB (Greer 2001; Greer and Sicard 2009; Greer et al. 2011) with the greatest import from reserves occurring in grape shoots at around the 6-10 leaf stage (Yang and Hori 1979; Yang et al. 1980) which occurred prior to about 20 DAB (Greer and Weston 2010). However, by the end of the preanthesis treatment period, carbon sequestration into biomass of shaded Hayward and Cabernet Sauvignon shoots was reduced which could suggest a reallocation of carbon away from vegetative growth or that assimilates from reserve sources were also required to support growth. In a similar study, shade imposed throughout the season also reduced the biomass accumulation of Semillon grape shoots (Greer et al. 2011). By contrast to the other vines, the shaded and intact Shiraz shoots of the present study were apparently able to balance the negative effects of shade on carbon acquisition by importing sufficient carbon to support shoot growth during the treatment period. The corollary to this conclusion was that the biomass accumulation of shaded Shiraz shoots was indeed reduced, when import was prevented by a phloem girdle. Thus, there appeared to be a marked contrast in the ability or the demands of the three vines to utilise carbon from reserves in the pre-anthesis period. Differences between Merlot and Pinot Noir in their requirement for reserves for growth during this period (Zapata et al. 2004a) is further support of there being marked cultivar differences in carbon requirements. Certainly, these results together highlight the predominant effect of reduced carbon acquisition on biomass accumulation during the early stages of shoot development and also that 251

288 vines vary in their ability to balance a reduction in carbon acquisition with import from reserves. Thus, these differences in the ability to balance carbon supply and demand through the utilisation of reserves are likely to be genetic in origin. It was evident that the combination of reduced carbon acquisition with a high demand for carbon being sequestered as shoot biomass resulted in the negative net carbon balance of shaded shoots on all vines until at least anthesis. This compared with the positive net carbon balances that occurred for the well exposed shoots of all vines and previously shown by the export of carbon from grape shoots (Hale and Weaver 1962; Koblet 1969). However, the combination of shade and girdle treatments also caused a lower biomass accumulation to occur compared with shaded and intact shoots and that meant that the net carbon balance was less negative in girdled than in intact shoots. This result provides further support for the conclusion that shaded intact shoots had imported significant amounts of carbon to support shoot growth during the pre-anthesis period. That significant import occurred normally during the early stages of shoot growth is well supported in the literature (Yang et al. 1980; Bennett et al. 2005; Field et al. 2009; Greer et al. 2011). However, the net carbon balance of both exposed and shaded Semillon shoots (Greer et al. 2011) and Hayward shoots exposed to both high and low PFD s (Greer 2001) became positive at around 40 DAB. To conclude, after the first DAB, all vines were highly dependent on photosynthetic carbon acquisition to establish the shoot leaf area and hence the vines productive capacity. However, when carbon acquisition was insufficient to support vegetative growth, such as occurred with shade, shoots were then able to access vine reserves through until anthesis. This ability of the vines to access different carbon pools provides a great deal of flexibility to meet the different environmental demands in the early spring Shoot development after anthesis The pre-anthesis effects of reduced carbohydrate supply to fruiting shoots continued to affect their performance throughout the remainder of the growing season. There was little change in leaf biomass of untreated shoots 252

289 (exposed/intact shoots) between anthesis and harvest but there were large changes in biomass accumulation of both stem ( fold increase) and fruit biomass ( fold increase) during the same period. These changes in the biomass of shoot stems and fruit had a large impact on carbohydrate pools in shoots after anthesis and are similar to the results from Chapter 3. However, these results differ from previous studies of Semillon grapevines where both leaf and stem biomass peaked at about anthesis (Greer and Sicard 2009; Greer et al. 2011) and then fruit biomass increased. These results indicate that leaf development, and hence photosynthetic capacity is determined before anthesis. Across the shoot, the distribution of biomass between leaves, stem and fruit in Shiraz (8, 20 and 72%) and Cabernet Sauvignon (9, 26, 64%) shoots at harvest were similar to that recorded in previous studies of Semillon (14, 17, 69%) (Greer et al. 2011), Pinot Noir (20, 13, 67%) (Petrie et al. 2000) and Cabernet Sauvignon (23, 20, 57%) (Palliotti et al. 2010) grapevines but differed from Sangiovese vines where each of these organs made up around 33% of shoot biomass (Palliotti et al. 2004). The distribution of biomass in leaves, stem and fruit of Hayward shoots at harvest (18, 12, 70%) was also similar to that of Shiraz and Cabernet Sauvignon shoots in this study, but differed from that of Hayward shoots in an earlier controlled environment study where the fruit biomass was very low (47, 40, 13%) (Greer et al. 2003). In general, fruit were the predominant sink in shoots across a range of vines and the few exceptions were when vines produced low crops. The impact of pre-anthesis shade on carbon acquisition was maintained across vines throughout the remainder of the season, despite the shade being removed, and continued the negative carbon balance in previously shaded shoots. The only exception was previously girdled and shaded Cabernet Sauvignon shoots that carried very low crop loads, hence had low demand and were, therefore, able to achieve a positive carbon balance by harvest. Effects of reducing both carbon acquisition and phloem import of shoots prior to anthesis also persisted after the shade was removed and 253

290 girdles had healed, but the effects on shoot components varied between cultivar and season. In the 2009/2010 season, the biomass of previously shaded shoots of both vines was less than that of exposed shoots at harvest, but there were no residual effects of girdling treatments. In the following season, pre-anthesis shade also resulted in reduced leaf biomass accumulation on shoots of all vines and also lower stem biomass of Shiraz shoots at harvest. In addition, the total biomass of all components of shaded and girdled Shiraz and Cabernet Sauvignon shoots was lower than in shoots from other treatments. This suggested that in the second season, when the total biomass of Shiraz shoots was double that of the 2009/2010 season and attributable to higher crop loads and hence a greater demand for assimilates. Furthermore, crop loads of Cabernet Sauvignon shoots were also comparatively high and shaded intact shoots appeared to access vine reserves to maintain fruit growth rates. These results, thus, concurred with the conclusion from Candolfi-Vasconcelos et al. (1994) that crop load can have a major demand on the source of carbohydrate and shoots can draw on reserves if carbon acquisition was inadequate, as shown above. By contrast, the biomass accumulation of Hayward shoots, where the crop loads were similar between seasons, was not affected by phloem import in either season. This suggests that the balance between carbohydrate sourced from current photosynthesis or exported from storage to the shoots may vary with prevailing environmental conditions and vine cultivar as concluded above Pollination and fruit set The large, 40% reduction in fruitset and fruit retention and the low number of seed per fruit on Hayward shoots in 2009/2010 was consistent with a direct effect of shade on reduced bee activity (Hopping and Jerram 1980). However, there was still a small reduction in fruit set and retention (10%), when flowers were pollinated both by bees and by hand in 2010/2011. Lai et al. (1990) noted that differences in fruit set and retention between early and late developing Hayward flowers and concluded that an effect of reduced carbon acquisition accounted for these differences. Thus, the small effect on fruit set and retention in 2010/2011 of the present study may also have occurred through reduced carbon acquisition. By contrast, there was no 254

291 effect of manipulating carbon acquisition on fruit set of Shiraz shoots in 2009/2010 but a 17% increase in fruit set and retention on shaded Cabernet Sauvignon shoots occurred in 2010/2011. This increased fruit set and retention was most likely caused by the 28% reduction in flower numbers on the shaded Cabernet Sauvignon shoots, and therefore, from reduced competition for carbohydrate supply rather than a direct effect of shade on pollination per se. Furthermore, results from earlier studies, where heavy shade was applied to grape shoots for short periods before and after anthesis, were variable, but generally resulted in a decreased fruit set (Ebadi et al. 1996; Ferree et al. 2001) linked to a reduction in carbon supply (Keller and Koblet 1994; Caspari et al. 1998; Rogiers et al. 2011). In addition, low carbohydrate supply during early berry development may also have resulted in increased berry drop (Candolfi-Vasconcelos and Koblet 1990). Thus, the results of the present study have demonstrated that carbon supply in the early stage of the growing season influenced fruit set of both Hayward and Cabernet Sauvignon berries Changes in carbon supply before anthesis influences final yield Final berry size was also influenced by altering shoot carbohydrate supply prior to anthesis. For example, with Shiraz shoots shade did not generally reduce yield because reduced berry numbers were compensated for by increases in berry size. An exception was on Shiraz shoots that were both shaded and girdled in the second season, when berry numbers were reduced without any compensation in berry size. In previous bunch thinning studies, an increase in berry size generally occurred with reduced bunch and hence berry numbers (Bravdo et al. 1984; Guidoni et al. 2002). In the current study, shade treatments imposed prior to anthesis also increased the mean seed weight of both Shiraz and Cabernet Sauvignon berries without influencing berry size and this is consistent with results from an earlier study where water stress was imposed on grapevines during early berry development (Shellie 2010). Lack of complete pollination reduced the size of Hayward berries in the first season of the present study, as indicated by the large reduction in both berry weight and seed number. This result is similar to those from other studies of pollination in Hayward (Hopping and Hacking 255

292 1983; Costa et al. 1993). However, despite hand pollination of all Hayward berries in the second season of the present study, there was still a small reduction in the berry weight and seed number on previously shaded shoots. This could imply there was also an effect of an early limitation in carbohydrate supply on the ovary size and ovule number at anthesis. A similar effect on ovary size and subsequent berry weight occurred between early and late opening Hayward flowers (Lai et al. 1990). Therefore the growth of Hayward berries was more sensitive to effects of reduced carbohydrate supply than Shiraz and Cabernet Sauvignon berries. Hayward shoots were apparently very sensitive to any reduction in shoot carbohydrate supply before anthesis, leading to 40-50% reductions in yield in both seasons. Shiraz and Cabernet Sauvignon shoots appeared more resilient, with no effect on yield of Shiraz shoots in the first season, but a yield reduction on shaded and girdled shoots of both grape cultivars in the second year. This suggested that Shiraz and Cabernet Sauvignon shoots were more readily able to access other carbohydrate pools to support berry development. Shade treatments consistently reduced berry number per shoot for all vines in both seasons while the girdling treatments only reduced berry numbers in the second season. Reductions in berry numbers occurred through a combination of reduced flower numbers, fruit set and berry loss suggesting shoot treatments affected processes both prior to and after anthesis. In conclusion, limiting supply of carbohydrates early in development influenced yield at harvest because of the sensitivity of berry size and/or berry numbers to the local availability of carbohydrate for growth Summary That carbohydrate supply to shoots during the later stages of flower development affected flower number and the ability of the flowers to set and develop into berries in all three vines confirmed the hypothesis that carbohydrate supply during this period is critical for floral development. From 20 to 30 DAB carbohydrate from photosynthesis was largely utilised to support rapid shoot growth, however, when carbon acquisition was reduced approximately five-fold by heavy shading, Shiraz and Cabernet Sauvignon 256

293 shoots were still able to import carbohydrate from reserve pools to support the high demand. The main effect of carbohydrate supply prior to anthesis on harvestable yield was via berry numbers through effects on both flower development and fruit set and retention, but there were also some effects on berry size. Manipulating the carbohydrate supply of shoots prior to anthesis through heavy shade also had a large effect on shoot leaf area and photosynthetic rates. As leaf development was completed prior to anthesis, this effect of shade on shoot leaf area subsequently had a large effect on the ability of the shoot to acquire carbon during the remainder of the season and consequently the final yield was severely limited. These results reinforce the conclusion that carbohydrate supply was extremely important in the early stages of shoot development in all vines and any limitations in the supply will have long term consequences for the productivity of the shoot over the remainder of the growing season. 257

294 258

295 7 Conclusions and Integration Fruiting vines are morphologically and physiologically different from many other major fruit crops. For example, kiwifruit and grapevines have different growth and flowering patterns and they produce comparatively few inflorescences, but a high proportion of flowers are pollinated and produce fruit (Coombe 1976; Hopping 1990). Both vines have a protracted period of inflorescence development, beginning with the initiation of floral meristems in spring and summer of the first season, followed by a quiescent period before floral morphogenesis occurs between budswell and anthesis in spring of the second season (Brundell 1975c; Walton et al. 1997; Boss et al. 2003). In their natural environment vines or lianas, do not produce a strong aerial perennial structure, but must seek the support of other plants and grow rapidly to reach suitable positions in the canopy to flower. There are, however, a number of gaps in understanding of how carbohydrate resources are allocated between the vegetative and reproductive growth in commercially important grapes and kiwifruit vines and how this influences their floral productivity. Accordingly, the following hypotheses were proposed to investigate these gaps and advance the understanding of regulation of flowering in kiwifruit and grapevines. 7.1 Hypothesis 1: Seasonal patterns of vine development and associated carbohydrate dynamics of grape and kiwifruit vines are similar during the key stages of development. The seasonal patterns of vine development and associated carbohydrate dynamics of untreated grape and kiwifruit vines were generally similar but the extent of vine growth and the concentrations of carbohydrate in tissues differed between the vines. Overall, the seasonal growth of the vines in this study could be divided into two distinct growth stages (Table 7.1): 1. From budbreak until the end of rapid shoot growth (0-100 DAB) 259

296 2. From the end of shoot growth until vines become dormant ( DAB). The long period between budbreak and anthesis, as well as the production of relatively few inflorescences were features of all vines in this and previous studies (Brundell 1975b; Coombe 1988; McPherson et al. 2001). The timing of budbreak and flowering was similar for all vines in this study (Chapter 3, 4 and 5). However the duration between budbreak and anthesis was shorter in Hayward (55 days) compared to both Shiraz and Cabernet Sauvignon (64 72 days) vines. The amount of budbreak of the Shiraz and Cabernet Sauvignon vines was also 2-fold higher than that of Hayward, but Hayward vines produced twice as many inflorescences per shoot. Therefore the numbers of inflorescences produced per bud were similar across the vines ranging between inflorescence per winter bud. During the first growth stage very high shoot growth rates were a consistent feature of all vines (Chapters 3, 4, 5 and 6) and similar to previous studies (Mullins et al. 1992; Piller and Meekings 1997). Rapid increases in shoot length (10-17 mm day -1 ) and leaf area (16-41 cm 2 day -1 ) occurred together with a 2-fold increase in photosynthetic rates. Temperature strongly influenced the timing and rate of shoot growth of all vines (Chapter 3, 4 and 5) and this is similar to previous studies (Greer and Jeffares 1998; Keller and Tarara 2010). During this period, inflorescences grew rapidly from floral meristems laid down in the previous season and new axillary buds developed and initiated floral meristems for the following season. However, at the end of this first growth stage the biomass of Hayward shoots was at least 2-fold higher than that of Shiraz and Cabernet Sauvignon shoots. This supports the conclusion that vines maintain high growth rates as part of the growth strategy of the liana habit, but the rate of growth varies between vines. During the second stage of vine development ( DAB), fruit and root growth became the dominant processes in all vines (Williams and Matthews 1990). There were however, differences in the extent of growth of these organs between the vines. In particular, new root development occurred 260

297 throughout this 150 day stage in Hayward vines compared with two short day periods at the beginning and end of the growth stage for both Shiraz and Cabernet Sauvignon vines. The period of fruit development also differed, with seed maturation and harvest approximately 30 days earlier for Shiraz and Cabernet Sauvignon vines compared with Hayward vines (Coombe 1976; Hopping 1976). The fruit biomass produced by Shiraz shoots was 73% greater than Hayward and 4-fold higher than Cabernet Sauvignon shoots. Thus, although the overall phenology of the vines was comparable there were differences in the duration of phases and considerable differences in the vigour and partitioning of biomass between the vines. In terms of photosynthesis, the light responses and seasonal trends of the three vines were broadly similar to this and previous studies (Kriedemann 1968; Greer and Halligan 2001). However photosynthetic rates for Hayward leaves were generally higher, rising more rapidly early in the season and then maintaining higher rates through until later in the season (beyond 180 DAB). Photosynthetic rates of Shiraz and Cabernet Sauvignon shoots declined earlier (150 DAB), during fruit maturation. The carbon balance of shoots from all vines became positive around DAB (Greer et al. 2003; Greer et al. 2011) and rates of both carbon acquisition and carbon sequestration as biomass peaked at the end of the rapid shoot growth in the first stage of development. Peak rates of carbon acquisition were higher in the Hayward shoots (2.7 ± 0.4 g C day -1 ) than in the Shiraz (1.2 ± 0.1 g C day -1 ) and Cabernet Sauvignon (0.4 ± 0.1 g C day -1 ) shoots. Therefore, seasonal patterns of shoot carbon acquisition varied between vines according to their leaf area and photosynthetic rates. The very earliest stage of shoot growth of all vines in spring was highly dependent on the supply of carbohydrate from the woody tissues (Chapter 3 and 5) as previously suggested for grapes (Bennett et al. 2005), particularly from the roots and canes. This use of reserves was evident from the decrease in the concentrations of total carbohydrates in roots (2-6-fold) and perennial wood (0.5-3-fold) during rapid shoot growth and also by root 261

298 death in the latter half of the first growth stage in all vines (Chapter 3). The concentrations of total carbohydrate in perennial storage organs of all vines continued to decline after shoots became autotrophic, through until the end of the rapid shoot growth stage. This sustained decline in carbohydrate reserves indicated that insufficient carbohydrate was available from photosynthesis to support the growth and metabolism of the whole vine until the end of the first stage of growth (100 DAB). Hence carbohydrate reserves stored in the previous season supported the development of all vines throughout the first stage of growth. Carbohydrate reserves accumulated in the perennial woody organs (2-3- fold increase in TNC concentration) as well as in the current season s canes (2-fold increase in TNC concentration) and fruit (2.5-8-fold increase in TNC concentration) of all vines throughout the second stage of growth (Chapters 3 and 6). The mild climatic conditions of the region enabled leaves to remain functional for at least 30 days after harvest, suggesting the post-harvest period was an important time for reserve replenishment of both grape and kiwifruit vines and consistent with Holzapfel et al. (2006). Active root growth and increases in cane biomass during this stage were also likely to be associated with increasing the storage capacity of all vines. Increases in carbohydrate concentrations and vine storage capacity supported the conclusion that reserve replenishment is a key feature of the second stage of growth for all vines. Seasonal patterns of vine growth and the dynamics of carbohydrate concentrations were generally similar but the extent of growth and carbohydrate partitioning differed between the three vines. During the first stage of growth carbohydrate, initially supplied from vine reserves and then supplemented by photosynthesis, was committed to the establishment of both the vegetative (leaves, the shoot axis and initiation of new axillary buds) and reproductive (inflorescences) components of the shoots. Thereafter, carbohydrate from photosynthesis supported fruit development, root growth and the replenishment of vine reserves in canes, roots and perennial wood. Carbon acquisition was greater and more carbon was partitioned to shoot 262

299 vegetative biomass and root growth in Hayward than that of the other vines. By contrast, a greater proportion of carbon was allocated to reproductive growth and carbohydrate concentrations in vine tissues were generally higher in Shiraz vines. Overall, the seasonal patterns of vine development and the dependence of each stage on carbohydrate from either reserves or photosynthesis were generally similar between the vines. 7.2 Hypothesis 2: The number of inflorescences produced is determined by carbohydrate supply during floral initiation and early bud development in grape and kiwifruit. To assess this hypothesis, carbohydrate supply to shoots was manipulated by girdling, shading and defoliation treatments during rapid shoot growth (the putative time of inflorescence initiation) and the effects determined from floral production in the subsequent season (Chapter 4) (Table 7.1). Despite the high carbohydrate demand by leaves, stems, inflorescences/fruit and newly initiated axillary buds during this stage (Chapter 3), carbohydrate supply was apparently not a key signal for inflorescence initiation in either Hayward or Cabernet Sauvignon vines. The evidence for this conclusion was that only girdling and defoliation of Hayward shoots virtually eliminated flowering at the defoliated nodes and similar to that reported in a previous study (Snelgar and Manson 1992). By contrast, girdling and shading of Hayward shoots which also reduced photosynthate supply and import of reserves reduced (50% of untreated) but did not eliminate flowering at the treated nodes. It was unlikely that carbohydrate supply itself was a key signal for inflorescence initiation as there should have been a consistent effect of reduced photosynthate supply from defoliation and shading treatments on inflorescence initiation. However, that inflorescence initiation had been prevented at the defoliated Hayward nodes suggested that floral signals were more likely leaf-derived, for example FT protein (Turck et al. 2008). Restricting import and carbohydrate supply (girdling and defoliation or girdling and shading) of Cabernet Sauvignon shoots also reduced (70% of untreated) but did not eliminate flowering and is consistent with earlier 263

300 studies (May and Antcliff 1963; Candolfi-Vasconcelos and Koblet 1990), again supporting the conclusion that high carbohydrate supply was not an essential signal for inflorescence initiation in either vine cultivar. Treatments that reduced shoot carbohydrate supply (girdling and shading or defoliation) during the period of inflorescence initiation reduced (13-35% lower than untreated) the carbohydrate content of internodes of both vines at the end of the growing season. This was due to the reduction (30-40%) in the growth of shoots in spring and hence, reduced capacity of shoots to store carbohydrate in autumn. This change in bud and internode carbohydrate content in winter explained 70% of the variance in inflorescence number produced by shoots in spring and agrees with previous studies of Sultana grapevines (Scholefield et al. 1978; Sommer et al. 2000). Therefore, the hypothesis that carbohydrate supply directly influenced the initiation of floral buds was not supported. However, carbohydrate supply during the rapid shoot growth of both vines did influence the capacity of internodes to store carbohydrate to support floral development in the following spring. 7.3 Hypothesis 3: That carbohydrate supply from reserve pools affects the early stages of floral morphogenesis and hence inflorescence numbers through the amount of carbohydrates available and the location of those carbohydrates. The previous hypothesis confirmed that the carbohydrate supplied by photosynthesis during rapid shoot growth had an effect on the floral development of both vines in the subsequent spring but not on inflorescence initiation per se. The question that remained was how carbohydrates stored in the vine in the previous season influenced floral morphogenesis? The capacity of the vines to store carbohydrate reserves was influenced by the size of internodes and the photosynthetic capacity of shoots was established during rapid shoot growth (Chapter 3 and 4) and the actual accumulation of reserves in woody tissues in late summer and autumn during stage two of vine growth (Chapter 3 and 5). Carbohydrates mobilised from both cane and 264

301 root tissues were essential for inflorescences development during the early stages of floral morphogenesis (0-30 DAB) (Chapter 4 and 5) and before shoots became autotrophic (Chapter 3 and 6) (Table 7.1). The size of shoot internodes (storage capacity) and leaf area (photosynthetic capacity), and hence the capacity for vines to accumulate and store carbohydrate later in the season, was largely determined by DAB. If the supply of carbohydrate from reserves was low during early shoot development (0-30 DAB), then shoot growth was reduced by 40%. However, if carbohydrate supply from photosynthesis was reduced 5-fold through heavy shade after shoots became autotrophic (30 DAB), there was a reduction in both shoot leaf area (30% decrease) and shoot biomass (40% decrease). These reductions in both storage capacity and carbon acquisition capacity of shoots were evident in reduced cane carbohydrate reserves (13-35%) in winter, and were associated with a 30-40% decrease in flowering in the following season (Chapter 4). These results demonstrated, therefore, that the capacity of canes to supply carbohydrate to support early floral morphogenesis in spring was influenced by the growth and development of shoots in the previous spring. The question as to how important restoring vine reserves in autumn are for floral productivity of grape and kiwifruit vines remains. Reducing carbon acquisition of vines 5-fold by shading in late summer and autumn over two seasons caused a 40% decrease in the root carbohydrate concentrations of both Hayward and Cabernet Sauvignon vines in winter (Chapter 5) and is consistent with previous results for Chardonnay vines (Bennett et al. 2005). Heavy shade in late summer and autumn also caused a 15-25% decrease in cane carbohydrate concentrations of both vines in each season (Chapter 5). By contrast, preventing export of carbohydrate from canes (by phloem girdling) during this period increased cane carbohydrate concentrations of both vines in winter by 10-15% (Chapter 5). Hayward vines were the most sensitive to these modifications of carbohydrate acquisition and storage in autumn, with up to a 2-fold reduction 265

302 in inflorescence numbers produced in each season. However, as effects on floral productivity of the Hayward vines were not cumulative over the three seasons it was suggested that cane rather than root reserves were the primary source of carbohydrate for early floral morphogenesis in this cultivar. By contrast, there was a cumulative effect of reduced carbohydrate acquisition in autumn on both root reserves and the inflorescence production by Cabernet Sauvignon vines, whereby shade reduced flowering by 30% after the third season of treatment. This suggested that the cumulative reduction in root carbohydrate influenced floral productivity of this cultivar. Cane carbohydrate status was also important for Cabernet Sauvignon vines as an increase in cane carbohydrate reserves (15%) through girdling was associated with increases in inflorescence production in the following season. These results suggest that the internodes adjacent to axillary buds as well as the roots were the major sources of carbohydrate to support early floral morphogenesis in spring. Thus, the hypothesis that carbohydrate supply from reserve pools affected the early stages of floral morphogenesis and hence inflorescence numbers through the availability of carbohydrates was supported. Furthermore, carbohydrate supply from these sinks could be modified, both during the rapid phase of shoot growth in spring of the previous season, when the capacity of shoots to capture and store carbon was determined, and during summer and autumn of the previous season when the carbohydrate accumulation in sinks actually occurred. 7.4 Hypothesis 4: That the amount of carbohydrates supplied from leaves to flowers during the late stages of flower development determined flower number and the ability of the flowers to set and develop into fruit. As shown in the previous sections, carbohydrate availability had a marked influence on the capacity of the vines to produce inflorescences during the early stages of floral morphogenesis in spring, especially if carbohydrate was limited in the previous season. The question now addressed is does the supply of carbohydrate during the final stage of floral morphogenesis (30 DAB - anthesis) influence the number of flowers produced and ultimately the 266

303 yield of the vine? To address this question, carbohydrate supply from photosynthesis and reserves were manipulated by shade and girdling of shoots during late floral morphogenesis in the second season of floral development (Table 7.1). The rapid increase in shoot growth and leaf area during rapid shoot growth determined the photosynthetic capacity of shoots for the remainder of the growing season. Shading shoots of all vines during this period decreased photosynthetic rates 5-fold and decreased shoot leaf area by 25%, thus dramatically reducing the capacity of shoots to produce carbohydrate from photosynthesis over the remainder of the season. Shoots of all vines that were shaded during late floral morphogenesis maintained a negative carbon balance throughout the entire growing season, despite being well exposed to light after anthesis. This reduction in carbohydrate supply affected flower development before anthesis reducing flowers numbers (10-30% decrease) and/or modifying fruit set (60-130% change) although effects varied between cultivar and season as found in earlier studies (Ferree et al. 2001). Modifying the carbohydrate supply before anthesis also influenced final berry size (70-122% change) of the vines, either through effects on ovary development prior to anthesis or effects on the subsequent capacity of shoots to acquire sufficient carbon to support fruit development after anthesis. This contributed to a 10-30% decrease in berry numbers and a reduction in the fruit yields of all vines. Therefore, shoot growth and leaf area expansion of all vines was a key driver of inflorescence development in late spring and berry development thereafter as well as the carbon economy of the reminder of the vine (Hale and Weaver 1962; Koblet 1969). Therefore, the hypothesis that the amount of carbohydrates supplied from leaves to flowers during the later stages of floral morphogenesis determined flower number and the ability of the flowers to set and develop into a fruit was confirmed. 267

304 268 Vines dormant Vines dormant Table 7.1: Overall summary of the timing of treatments to manipulate carbohydrate supply during floral development and the timing and effect of treatments on floral production relative to the untreated control (%) First season of floral development Second season of floral development Vegetative growth Floral growth Chapter 4 Treatments Floral effect Hayward Cabernet Sauvignon Chapter 5 Treatments Floral effect Hayward Cabernet Sauvignon Chapter 6 Treatment Floral effect Hayward Cabernet Sauvignon Shiraz First stage of growth Second stage of growth First stage of growth Second stage of growth Budbreak - 30 DAB Rapid shoot growth Reserve storage Budbreak -30 DAB Rapid shoot growth Reserve storage Inflorescence initiation Buds quiescent Floral morphogenesis Berry growth Early Late Girdle (G), Girdle and defoliate (GD), Girdle and shade (GS) GD 12% GS 40% Cane girdle (G) x Vine shade (S) GD 77% GS 72 G 200% S 40% G 130% S 70% (yr 3) Girdle (G) x Shade (S) G 92 % (yr 2) S 75 % G 64 % (yr 2) S 70 % G 81 % (yr 2) S 76 % 268

305 7.5 Future studies The storage of carbohydrate in woody vine tissues in summer and autumn and its remobilisation during the early stages of floral morphogenesis in spring had a large effect on the number of inflorescences produced by both kiwifruit and grape vines. Further studies, for example using radiolabelling, enzymes or gene expression will be required to more clearly define how and when these reserves accumulate and are mobilised to support spring inflorescence development. Results from this study have demonstrated that reserves in both canes and roots are essential for early floral morphogenesis in spring. However, the carbohydrate concentrations in the perennial cordons and trunks of the vines also declined during spring but did not appear to be linked to inflorescence development. Therefore, the role of reserves from each of these organs to support the early stages of vine development needs to be more clearly defined. Further studies are also needed to determine how the initial period of rapid growth influences the capacity of shoots (subsequently canes) to store carbohydrates during reserve accumulation in autumn. Finally, to develop a clearer picture of seasonal carbohydrate dynamics in these vines, especially carbohydrate storage and remobilisation, whole vine carbohydrate budgets are needed so that changes in carbohydrate pools in organs at key points in vine development can be examined. The strong interaction between light intensity, leaves and floral signals in Hayward vines appears to be a key factor affecting inflorescence development in that cultivar. During inflorescence initiation, defoliation clearly interrupted the flow of signals from the leaf to the subtending axillary bud and hence eliminated flowering at these nodes. In addition, heavy shade during the period of reserve accumulation in autumn eliminated flowering in intact, but not in girdled canes. Both of these observations suggest that phloem mobile signals are very important for inflorescence development in Hayward vines. These signals require further investigation and molecular studies across a range of Actinidia genotypes may be useful to determine the mechanisms involved. 269

306 A doubling of growth of both Cabernet Sauvignon and Hayward shoots in spring of the first season of floral development was associated with a substantial reduction in inflorescence production in the following spring. This implied that factors associated with accelerated vegetative growth suppressed inflorescence initiation in both vines. The relationship between temperature, vigorous shoot growth and inflorescence initiation in both Hayward and Cabernet Sauvignon vines needs further investigation. This could provide further insights into seasonal variation in inflorescence production of both vines. Molecular studies and hormone analysis of vines that exhibit a range of vegetative vigour may help to develop understanding of the relationship between vigour and floral development as well as the mechanisms involved. 7.6 Overall Summary The seasonal pattern of growth and carbohydrate dynamics of Shiraz and Cabernet Sauvignon grapevines and Hayward kiwifruit vines were generally similar but the duration of some phenological stages and the extent of growth varied between the vines. The vines were characterised by a protracted period of inflorescence development over two seasons, vigorous shoot growth in spring and significant accumulation of carbohydrate reserves in autumn. The extent of vegetative shoot and root growth was greatest in Hayward vines while fruit production was greatest in Shiraz vines. However carbohydrate concentrations were generally greater in Shiraz compared with Hayward and Cabernet Sauvignon vines. Reduced carbohydrate supply affected inflorescence production by both Hayward and Cabernet Sauvignon vines during floral morphogenesis. Altering carbohydrate supply during early shoot growth of the first season did not influence inflorescence initiation directly, but did influence the capacity of shoots to store reserves for the following season. Shading vines in autumn of the first season reduced the storage of carbohydrate in both canes and roots, while girdling canes elevated storage of carbohydrate in canes. Carbohydrate from reserves stored in the previous season was essential to support the early stages of floral morphogenesis and altering this supply had 270

307 large effects on the inflorescence development of particularly Hayward but also Cabernet Sauvignon vines. Modifying carbohydrate supply from photosynthesis during the later stages of floral morphogenesis also influenced the reproductive capacity and yields of Hayward and Cabernet Sauvignon vines. Therefore, carbohydrate supply was essential for floral morphogenesis in both grape and kiwifruit vines, with the early stages dependent on carbohydrate from vine reserves and the later stages on carbohydrate from photosynthesis. Hence, carbohydrate supply is a key factor determining the productivity of both grape and kiwifruit vines. 271

308 272

309 References Abbott DL (1984) 'The apple tree: Physiology and management.' (Grower Books: London) Albrigo LG, Sauco VG (2004) Flower bud induction, flowering and fruit-set of some tropical and subtropical fruit tree crops with special reference to citrus. Acta Horticulturae 632, Aloni B, Karni L, Zaidman Z, Schaffer AA (1996) Changes of carbohydrates in pepper (Capsicum annuum L) flowers in relation to their abscission under different shading regimes. Annals of Botany 78, Aloni B, Pashkar T, Karni L (1991) Partitioning of (C-14) sucrose and acid invertase activity in reproductive organs of pepper plants in relation to their abscission under heat-stress. Annals of Botany 67, Amerine MA, Root GA (1960) Carbohydrate content of various parts of the grape cluster. II. American Journal of Enology and Viticulture 11, Andersen MN, Asch F, Wu Y, Jensen CR, Naested H, Mogensen VO, Koch KE (2002) Soluble invertase expression is an early target of drought stress during the critical, abortion-sensitive phase of young ovary development in maize. Plant Physiology 130, Andres F, Coupland G (2012) The genetic basis of flowering responses to seasonal cues. Nature Reviews Genetics 13, Antcliff AJ, Webster WJ (1955) Studies on the Sultana vine. I. Fruit bud distribution and budburst with reference to forecasting potential crop. Australian Journal of Agricultural Research 6, Asadian G, Pieber K (2005) Morphological variations in walnut varieties of the mediterranean regions. International Journal of Agriculture and Biology 7, Bagnall DJ, King RW (2001) Phytochrome, photosynthesis and flowering of Arabidopsis thaliana: Photophysiological studies using mutants and transgenic lines. Australian Journal of Plant Physiology 28, Baldet P, Hernould M, Laporte F, Mounet F, Just D, Mouras A, Chevalier C, Rothan C (2006) The expression of cell proliferation-related genes in early developing flowers is affected by a fruit load reduction in tomato plants. Journal of Experimental Botany 57, Bangerth K (2009) Floral induction in mature, perennial angiosperm fruit trees: Similarities and descrepancies with annual/biennial plants and the involvement of plant hormones. Scientia Horticulturae 122, Bangerth KF (2006) Flower induction in perennial fruit trees: Still an enigma? Acta Horticulturae 727, Bates TR, Dunst RM, Joy P (2002) Seasonal dry matter, starch, and nutrient distribution in 'Concord' grapevine roots. Hortscience 37, Bennett J, Jarvis P, Creasy GL, Trought MCT (2005) Influence of defoliation on overwintering carbohydrate reserves, return bloom, and yield of mature Chardonnay grapevines. American Journal of Enology and Viticulture 56, Bernier G (1988) The control of floral evocation and morphogenesis. Annual Review of Plant Physiology and Plant Molecular Biology 39, Bernier G, Havelange A, Houssa C, Petitjean A, Lejeune P (1993) Physiological signals that induce flowering. Plant Cell 5, Bernier G, Kinet JM, Sachs RM (1981) 'The physiology of flowering.' (CRC Press: Boca Raton, Florida) Bernier G, Perilleux C (2005) A physiological overview of the genetics of flowering time control. Plant Biotechnology Journal 3, Bertelsen MG, Tustin DS, Waagepetersen RP (2002) Effects of GA(3) and GA(4+7) on early bud development of apple. Journal of Horticultural Science & Biotechnology 77,

310 Bertin N, Gautier H, Roche C (2002) Number of cells in tomato fruit depending on fruit position and source-sink balance during plant development. Plant Growth Regulation 36, Black MZ (2012) The kiwifruit (Actinidia sp.) vine root system: responses to vine manipulations., Waikato University. Black MZ, Minchin PEH, Gould N, Patterson KJ, Clearwater MJ (2012) Measurement of Bremsstralung radiation for in vivo monitoring of 14 C tracer distribution between fruit and roots of kiwifruit (Actinidia arguta) cuttings. Planta 236, Blazquez MA (2000) Flower development pathways. Journal of Cell Science 113, Blazquez MA, Green R, Nilsson O, Sussman MR, Weigel D (1998) Gibberellins promote flowering of Arabidopsis by activating the LEAFY promoter. Plant Cell 10, Bohlenius H, Huang T, Charbonnel-Campaa L, Brunner AM, Jansson S, Strauss SH, Nilsson O (2006) CO/FT regulatory module controls timing of flowering and seasonal growth cessation in trees. Science 312, Bohner J, Bangerth F (1988) Effects of fruit-set sequence and defoliation on cell number, cell-size and hormone levels of tomato fruits (Lycopersicon esculentum Mill.) within a truss. Plant Growth Regulation 7, Boldingh H, Smith GS, Klages KU (2000) Seasonal concentrations of non-structural carbohydrates of five Actinidia species in fruit, leaf and fine root tissue. Annals of Botany 85, Bonhomme M, Rageau R, Lacointe A, Gendraud M (2005) Influences of cold deprivation during dormancy on carbohydrate contents of vegetative and floral primordia and nearby structures of peach buds (Prunus persica L. Batch). Scientia Horticulturae 105, Boss PK, Bastow RM, Mylne JS, Dean C (2004) Multiple pathways in the decision to flower: Enabling, promoting, and resetting. Plant Cell 16, S18-S31. Boss PK, Buckeridge EJ, Poole A, Thomas MR (2003) New insights into grapevine flowering. Functional Plant Biology 30, Boss PK, Thomas MR (2002) Association of dwarfism and floral induction with a grape 'green revolution' mutation. Nature 416, Boyd LM, Barnett AM (2011) Manipulation of whole-vine carbon allocation using girdling, pruning, and fruit thinning affects fruit numbers and quality in kiwifruit. Hortscience 46, Boyer JS, McLaughlin JE (2007) Functional reversion to identify controlling genes in multigenic responses: Analysis of floral abortion. Journal of Experimental Botany 58, Bramley RGV, Hamilton RP (2004) Understanding variability in winegrape production systems 1. Within vineyard variation in yield over several vintages. Australian Journal of Grape and Wine Research 10, Bravdo BA, Hepner Y, Loinger C, Cohen S, Tabacman H (1984) Effect of crop level on growth, yield and wine quality of a high yielding Carignane vineyard. American Journal of Enology and Viticulture 35, Brown K, Jackson DI, Steans GF (1988) Effects of chlormequat, girdling, and tipping on berry set in Vitis vinifera L. American Journal of Enology and Viticulture 39, Brundell DJ (1975a) Effect of defoliation, shading, and tipping of shoots on flower bud development and vegetative growth of chinese gooseberry (Actinidia chinensis Planch.). New Zealand Journal of Agricultural Research 18, Brundell DJ (1975b) Flower development of the chinese gooseberry (Actinidia chinensis Planch.) I. Development of the flowering shoot. New Zealand Journal of Botany 13,

311 Brundell DJ (1975c) Flower development of the chinese gooseberry (Actinidia chinensis Planch.) II. Development of the flower bud. New Zealand Journal of Botany 13, Brunner AM, Nilsson O (2004) Revisiting tree maturation and floral initiation in the poplar functional genomics era. New Phytologist 164, Burge GK, Spence CB, Marshall RR (1987) Kiwifruit - Effects of thinning on fruit size, vegetative growth, and return bloom. New Zealand Journal of Experimental Agriculture 15, Buttrose MS (1966) The effect of reducing leaf area on the growth of roots, stems and berries of Gordo grape vine. Vitis 5, Buttrose MS (1969) Fruitfulness in grapevines: Effects of light intensity and temperature. Botanical Gazette 130, Buwalda JG, Hutton RC (1988) Seasonal changes in root-growth of kiwifruit. Scientia Horticulturae 36, Buwalda JG, Smith GS (1990) Acquisition and utlization of carbon, mineral nutrients, and water by the kiwifruit vine. Horticultural Reviews 12, Candolfi-Vasconcelos MC, Candolfi MP, Koblet W (1994) Retranslocation of carbon reserves from the woody storage tissues into the fruit as a response to defoliation stress during the ripening period in Vitis vinifera L. Planta 192, Candolfi-Vasconcelos MC, Koblet W (1990) Yield, fruit quality, bud fertility and starch reserves of the wood as a function of leaf removal in Vitis vinifera - Evidence of compensation and stress recovering. Vitis 29, Carbonneau A, Casteran P (1979) Irrigation-depressing effect on floral initiation of Cabernet sauvignon grapevines in Bordeaux area. American Journal of Enology and Viticulture 30, 3-7. Carmona MJ, Calonje M, Martinez-Zapater JM (2007a) The FT/TFL1 gene family in grapevine. Plant Molecular Biology 63, Carmona MJ, Chieb J, Martinez-Zapater JM, Thomas MR (2008) A molecular genetic perspective of reproductive development in grapevine. Journal of Experimental Botany 59, Carmona MJ, Cubas P, Calonje M, Martinez-Zapater JM (2007b) Flowering transition in grapevine (Vitis vinifera L.). Canadian Journal of Botany 85, Cartechini A, Palliotti A (1995) Effect of shading on vine morphology and productivity and leaf gas exchange characteristics in grapevines in the field. American Journal of Enology and Viticulture 46, Caspari HW, Lang A, Alspach P (1998) Effects of girdling and leaf removal on fruit set and vegetative growth in grape. American Journal of Enology and Viticulture 49, Castellanos AE (1991) Photosynthesis and gas exchange of vines. In 'The Biology of Vines'. (Eds FE Putz and HA Mooney) pp (Cambridge University Press: Cambridge) Cawoy V, Lutts S, Ledent JF, Kinet JM (2007) Resource availability regulates reproductive meristem activity, development of reproductive structures and seed set in buckwheat (Fagopyrum esculentum). Physiologia Plantarum 131, Chandler JW (2011) The hormonal regulation of flower development. Journal of Plant Growth Regulation 30, Chaumont M, Morot-Gaudry JF, Foyer CH (1994) Seasonal and diurnal changes in photosynthesis and carbon partitioning in Vitis vinifera leaves in vines with and without fruit. Journal of Experimental Botany 45, Chloupek O, Hrstkova P, Schweigert P (2004) Yield and its stability, crop diversity, adaptability and response to climate change, weather and fertilisation over 75 years in the Czech Republic in comparison to some european countries. Field Crops Research 85, Clingeleffer PR (2006) Management practices for Sunmuscat (Vitis vinifera L.): a new drying variety. Australian Journal of Grape and Wine Research 12,

312 Cohen A (1981) Recent developments in girdling of citrus trees. In 'Proceedings International Society of Citriculture'. Okitsu, Japan. (Ed. K Matsumoto) pp Comas LH, Anderson LJ, Dunst RM, Lakso AN, Eissenstat DM (2005) Canopy and environmental control of root dynamics in a long-term study of Concord grape. New Phytologist 167, Coombe B (1976) The development of fleshy fruits. Annual Review of Plant Physiology 27, Coombe B (1988) Grape phenology. In 'Viticulture'. (Eds BG Coombe and D PR) pp (Winetitles: Adelaide) Coombe B (1995) Adoption of a system for identifying grapevine growth stages. Australian Journal of Grape and Wine Research 1, Coombe BG (1959) Fruit set and development in seeded grape varieties as affected by defoliation, topping, girdling, and other treatments. American Journal of Enology and Viticulture 10, Coombe BG (1962) The effect of removing leaves, flowers and shoot tips on fruit-set in Vitis vinifera L. Journal of Horticultural Science 37, Corbesier L, Coupland G (2005) Photoperiodic flowering of Arabidopsis: Integrating genetic and physiological approaches to characterization of the floral stimulus. Plant Cell and Environment 28, Corbesier L, Coupland G (2006) The quest for florigen: a review of recent progress. Journal of Experimental Botany 57, Corbesier L, Havelange A, Lejeune P, Bernier G, Perilleux C (2001) N content of phloem and xylem exudates during the transition to flowering in Sinapis alba and Arabidopsis thaliana. Plant Cell and Environment 24, Corbesier L, Lejeune P, Bernier G (1998) The role of carbohydrates in the induction of flowering in Arabidopsis thaliana: Comparison between the wild type and a starchless mutant. Planta 206, Corbesier L, Prinsen E, Jacqmard A, Lejeune P, Van Onckelen H, Perilleux C, Bernier G (2003) Cytokinin levels in leaves, leaf exudate and shoot apical meristem of Arabidopsis thaliana during floral transition. Journal of Experimental Botany 54, Corelli-Grappadelli L, Magnanini E (1993) A whole-tree system for gas-exchange studies. Hortscience 28, Costa G, Testolin R, Vizzotto G (1993) Kiwifruit pollination - an unbiased estimate of wind and bee contribution. New Zealand Journal of Crop and Horticultural Science 21, Cseke LJ, Cseke SB, Ravinder N, Taylor LC, Shankar A, Sen B, Thakur R, Karnosky DF, Podila GK (2005) SEP-class genes in Populus tremuloides and their likely role in reproductive survival of poplar trees. Gene (Functional Genomics) 358, Davison RM (1990) The physiology of the kiwifruit vine. In 'Kiwifruit Science and Management'. (Eds IJ Warrington and GC Weston) pp (Ray Richards Publishers: Auckland) Dayer S, Prieto JA, Galat E, Perez Pena J (2013) Carbohydrate reserve status of Malbec grapevines after several years of regulated deficit irrigation and crop load regulation. Australian Journal of Grape and Wine Research 19, Dennis FG (1981) Limiting factors in fruit set of 'Delicious' apple. Acta Horticulturae 120, Dennis FG (1986) Apple. In 'Handbook of Fruit Set and Development'. (Ed. SP Monselise) pp (CRC Press Inc: Florida) Dennis FG (2003) Flowering, pollination and fruit set and development. In 'Apples, Botany, Production and Uses'. (Eds DC Ferree and IJ Warrington) pp (CAB International) Dennis FG, Neilsen JC (1999) Physiological factors affecting biennial bearing in tree fruit: The role of seeds in apple. Horttechnology 9,

313 Dokoozlian NK (1999) Chilling temperature and duration interact on the budbreak of Perlette grapevine cuttings. Hortscience 34, Dry PR, Gregory GR (1988) Grapevine Varieties. In 'Viticulture'. (Eds BG Coombe and PR Dry) pp (Winetitles Adelaide) Dry PR, Longbottom S, Mcloughlin S, Johnson TE, Collins C (2010) Classification of reproductive performance of ten winegrape varieties. Australian Journal of Grape and Wine Research 16, Duchene E, Monamy C, Langellier F, Jaegli N, Salber R, Meluc D, Panigai L (2003) Effects of the ripening conditions in the vineyard on growth and yield components in the following season. Journal International des Sciences de la Vigne et du Vin 37, Dunn GM, Martin SR (2000) Do temperature conditions at budburst affect flower number in Vitis vinifera. cv. Cabernet Sauvignon? Australian Journal of Grape and Wine Research 6, Dunn GM, Martin SR (2007) A functional association in Vitis vinifera L. cv. Cabernet Sauvignon between the extent of primary branching and the number of flowers formed per inflorescence. Australian Journal of Grape and Wine Research 13, Durdan SF, Herbert RJ, Rogers HJ, Francis D (2000) The determination time of the carpel whorl is differentially sensitive to carbohydrate supply in Pharbitis nil. Plant Physiology 123, Ebadi A, May P, Coombe BG (1996) Effect of short-term temperature and shading on fruit-set, seed and berry development in model vines of V. vinifera, cvs Chardonnay and Shiraz. Australian Journal of Grape and Wine Research 2, 2-9. Eckardt FE (1977) Physiological behaviour in relation to the environment - a comparison between a crop and various types of natural vegetation. In 'Environmental effects on crop physiology'. (Eds JJ Landsberg and CV Cutting) pp (Academic Press: London) Eichorn K, Lorenz D (1977) Phaenologische Entwicklungstadien der Rabe. Nachrichtenblatt des Deutschen Pflanzenschutzdienstes (Braunschweig) 29, Eissenstat DM, Bauerle TL, Comas LH, Lakso AN, Neilsen D, Neilsen GH, Smart DR (2006) Seasonal patterns of root growth in relation to shoot phenology in grape and apple. Acta Horticulturae 721, Elsner EA, Jubb GL (1988) Leaf area estimation of Concord grape leaves from simple linear measurements. American Journal of Enology and Viticulture 39, Eltom M, Trought MCT, Winefield CS (2013) The effects of cane girdling before budbreak on shoot growth, leaf area and carbohydrate content of Vitis vinifera L. Sauvignon Blanc grapevines. Functional Plant Biology 40, Eriksson S, Bohlenius H, Moritz T, Nilsson O (2006) GA4 is the active gibberellin in regulation of LEAFY transcription and Arabidopsis floral initiation. Plant Cell 18, Escalona JM, Tomas M, Martorell H, Medrano M, Ribas-Carbo M, Flexas J (2012) Carbon balance in grapevines under different soil water supply: importance of whole plant respiration. Australian Journal of Grape and Wine Research 18, Evans LT (1969) The nature of flower induction. In 'The induction of flowering - some case histories'. (The Macmillan Company of Australia: Melbourne) Fabbri A, Lisetti M, Benelli C (1991) Studies on flower induction in kiwifruit. Acta Horticulturae 297, Ferguson AR (1990) Stem, branches, leaves and roots of the kiwifruit vine. In 'Kiwifruit Science and Management'. (Eds IJ Warrington and GC Weston) pp (Ray Richards Publisher: Auckland) Ferguson AR, Eiseman JA, Leonard JA (1983) Xylem sap from Actinidia chinensis: seasonal changes in composition. Annals of Botany 51, Ferguson AR, Huang H (2007) Genetic resources of kiwifruit: domestication and breeding. Horticultural Reviews 33,

314 Fernandez GCJ (1992) Residual analysis and data transformations in statistical analysis. Hortscience 27, Ferree DC, McArtney SJ, Scurlock DM (2001) Influence of irradiance and period of exposure on fruit set of french-american hybrid grapes. Journal of the American Society for Horticultural Science 126, Field SK (2013) Investigating the relationship between reserve nitrogen and carbohydrate on initial grapevine growth and development and the role of root derived cytokinins., Charles Sturt. Field SK, Smith JP, Holzapfel BP, Hardie WJ, Emery RJN (2009) Grapevine responses to soil temperature: Xylem cytokinins and carbohydrate reserve mobilisationfrom budbreak to anthesis. American Journal of Enology and Viticulture 60, Fisher JB, Ewers FW (1991) Structural responses to stem injury in vines. In 'The Biology of Vines'. (Eds FE Putz and HA Mooney) pp (Cambridge University Press: Cambridge, UK) Forshey CG, Elfving DC (1989) The relationship between vegetative growth and fruiting in apple trees. Horticultural Reviews 11, Foster T, Johnston R, Seleznyova A (2003) A morphological and quantitative characterization of early floral development in apple (Malus x domestica Borkh.). Annals of Botany 92, Freeman BM, Smart RE (1976) A root observation laboratory for studies with grapevines. American Journal of Enology and Viticulture 27, 36. Garcia-Luis A, Fornes F, Guardiola JL (1995) Leaf carbohydrates and flower formation in citrus. Journal American Society of Horticultural Science 120, Givnish TJ (1988) Adaptation to sun and shade: a whole plant perspective. Australian Journal of Plant Physiology 15, Glad C, Regnard JL, Querou Y, Brun O, Morotgaudry JF (1992) Phloem sap exudates as a criterion for sink strength appreciation in Vitis-vinifera Cv Pinot-Noir grapevines. Vitis 31, Gladstone EA, Dokoozlian NK (2003) Influence of leaf area density and trellis/training system on the light microclimate within grapevine canopies. Vitis 42, Goffinet MC (2004) Relation of applied crop stress to inflorescence development, shoot growth characteristics, and cane starch reserves in 'Concord' grapevine. Viticulture - Living with Limitations 640, Goldberg-Moeller R, L. S, Shlizerman L, Samuels S, Zur N, Ophir R, Blumwald E, Sadka A (2013) Effects of gibberellin treatment during flowering induction period on global gene expression and the transciption of flowering-control genes in Citrus buds. Plant Science 198, Goldschmidt EE (2013) The evolution of fruit tree productivity: A review.. Economic Botany 67, Goldschmidt EE, Aschkenazi N, Herzano Y, Schaffer AA, Monselise SP (1985) A role for carbohydrate levels in the control of flowering in citrus. Scientia Horticulturae 26, Goldschmidt EE, Golomb A (1982) The carbohydrate balance of alternate-bearing citrus trees and the significance of reserves for flowering and fruiting. Journal American Society of Horticultural Science 107, Goldschmidt EE, Samach A (2004) Aspects of flowering in fruit trees. Acta Horticulturae 653, Goldschmidt EE, Tamim M, Goren R (1997) Gibberellins and flowering in citrus and other fruit trees: A critical analysis. Acta Horticulturae 463, Goren R, Huberman M, Goldschmidt EE (2004) Girdling: physiological and horticultural aspects. Horticultural Reviews 30, Granger AR (2005) The effect of three rootstocks on yield and fruiting of sweet cherry. Acta Horticulturae 667,

315 Grant JA, Ryugo K (1982) Influence of developing shoots on flowering potential of dormant buds of Actinidia chinensis. Hortscience 17, Grant JA, Ryugo K (1984) Influence of within-canopy shading on fruit size, shoot growth, and return bloom in kiwifruit. Journal of the American Society for Horticultural Science 109, Greer DH (1999) Seasonal and daily changes in carbon acquisition of kiwifruit leaves with and without axillary fruit. New Zealand Journal of Crop and Horticultural Science 27, Greer DH (2001) Photon flux density dependence of carbon acquisition and demand in relation to shoot growth of kiwifruit (Actinidia deliciosa) vines grown in controlled environments. Australian Journal of Plant Physiology 28, Greer DH (2012) Modelling leaf photosynthetic and transpiration temperature dependent responses in Vitis vinifera cv. Semillon grapevines growing in hot, irrigated vineyard conditions. AoB PLANTS. Greer DH, Cirillo C, Norling CL (2003) Temperature-dependence of carbon acquisition and demand in relation to shoot and fruit growth of fruiting kiwifruit (Actinidia deliciosa) vines grown in controlled environments. Functional Plant Biology 30, Greer DH, Halligan EA (2001) Photosynthetic and fluorescence light responses for kiwifruit (Actinidia deliciosa) leaves at different stages of development on vines grown at two different photon flux densities. Australian Journal of Plant Physiology 28, Greer DH, Jeffares D (1998) Temperature-dependence of carbon acquisition and demand in relation to shoot growth of kiwifruit (Actinidia deliciosa) vines grown in controlled environments. Australian Journal of Plant Physiology 25, Greer DH, Seleznyova AN, Green SR (2004) From controlled environments to field simulations: leaf area dynamics and photosynthesis of kiwifruit vines (Actinidia deliciosa). Functional Plant Biology 31, Greer DH, Sicard SM (2009) The net carbon balance in relation to growth and biomass accumulation of grapevines (Vitis vinifera cv. Semillon) grown in a controlled environment. Functional Plant Biology 36, Greer DH, Weedon MM, Weston C (2011) Reductions in biomass accumulation, photosynthesis in situ and net carbon balance are the costs of protecting Vitis vinifera 'Semillon' grapevines from heat stress with shade covering. AoB PLANTS. Greer DH, Weston C (2010) Heat stress affects flowering, berry growth, sugar accumulation and photosynthesis of Vitis vinifera cv. Semillon grapevines grown in a controlled environment. Functional Plant Biology 37, Greer DH, Wunsche JN (2003) Late-season temperature effects on the carbon economy and tree performance of 'Royal Gala' apple (Malus domestica) trees. New Zealand Journal of Crop and Horticultural Science 31, Guidoni S, Allara P, Schubert A (2002) Effect of bunch thinning on berry skin anthocyanin composition of Vitis vinifera cv. Nebbiolo. American Journal of Enology and Viticulture 53, Gyllenstrand N, Clapham D, Kallman T, Lagercrantz U (2007) A Norway spruce FLOWERING LOCUS T homolog is implicated in control of growth rhythm in conifers. Plant Physiology 144, Hale CR, Weaver RJ (1962) The effect of developmental stage on direction of translocation of photosynthate in Vitis vinifera. Hilgardia 33, Hamman RA, Dami IE, Walsh TM, Stushnoff C (1996) Seasonal carbohydrate changes and cold hardiness of Chardonnay and Riesling grapevines. American Journal of Enology and Viticulture 47, Hardie WJ (2000) Grapevine biology and adaptation to viticulture. Australian Journal of Grape and Wine Research 6, Harris RW (1992) Root-shoot ratios. Journal of Arboriculture 18,

316 Hegele M, Wunsche JN, Manochai P, Naphrom D, Sruamsiri P (2008) Effect of girdling and defoliation of longan shoots on hormonal changes during flower induction by KClO3. Acta Horticulturae 774, Heidenreich C, Pritts M, Kelley M-J, Demchak K (2008) 'High tunnel raspberries and blackberries.' Department of Hoticulture Publications, Cornell University, New York. Hendrickson L, Ball MC, Wood JT, Chow WS, Furbank RT (2004) Low temperature effects on photosynthesis and growth of grapevine. Plant, Cell and Environment 27, Henriod RE, Jameson PE, Clemens J (2003) Effect of irradiance during floral induction on floral initiation and subsequent development in buds of different size in Metrosideros excelsa (Myrtaceae). Journal of Horticultural Science & Biotechnology 78, Heyer AG, Raap M, Schroeer B, Marty B, Willmitzer L (2004) Cell wall invertase expression at the apical meristem alters floral, architectural, and reproductive traits in Arabidopsis thaliana. Plant Journal 39, Holzapfel BP, Smith JP (2012) Developmental stage and climatic factors impact more on carbohydrate reserve dynamics of Shiraz than cultural practice. American Journal of Enology and Viticulture 63, Holzapfel BP, Smith JP, Field SK, Hardie WJ (2010) Dynamics of carbohydrate reserves in cultivated grapevines. Horticultural Reviews 37, Holzapfel BP, Smith JP, Mandel RM, Keller M (2006) Manipulating the postharvest period and its impact on vine productivity of Semillon grapevines. American Journal of Enology and Viticulture 57, Hopping ME (1976) Structure and development of fruit and seeds in chinese gooseberry (Actinidia chinensis Planch.). New Zealand Journal of Botany 14, Hopping ME (1977) Effect of light intensity during cane development on subsequent bud break and yield of 'Palomino' grape vines. New Zealand Journal of Experimental Agriculture 5, Hopping ME (1990) Floral biology, pollination, and fruit set. In 'Kiwifruit Science and Management'. (Eds IJ Warrington and GC Weston) pp (Ray Richards Publisher: Auckland) Hopping ME, Hacking NJA (1983) A comparison of pollen application methods for artificial pollinationof kiwifruit. Acta Horticulturae 139, Hopping ME, Jerram EM (1979) Pollination of kiwifruit (Actinidia chinensis Planch.): stigma-style structure and pollen tube growth. New Zealand Journal of Botany 17, Hopping ME, Jerram EM (1980) Supplementary pollination of tree fruits. II. Field trials on kiwifruit and Japanese plums. New Zealand Journal of Agricultural Research 23, Hsu C-Y, Adams JP, et al. (2011) FLOWERING LOCUS T duplication coordinates reproductive and vegetative growth in perennial poplar. Proceedings of the National Academy of Sciences of the United States of America 108, Huang S, Cerny RE, Qi Y, Bhat D, Aydt CM, Hanson DD, Malloy KP, Ness L (2003) Transgenic studies on the involvement of cytokinin and gibberellin in male development. Plant Physiology 131, Hunter JJ, Ruffner HP (2001) Assimilate transport in grapevines - effect of phloem disruption. Australian Journal of Grape and Wine Research 7, Irish VF (2010) The flowering of Arabidopsis flower development. The Plant Journal 61, Ito A, Hayama H, Kashimura Y (2002) Sugar metabolism in buds during flower bud formation: a comparison of two Japanese pear Pyrus pyrifolia (Burm.) Nak. cultivars possessing different flowering habits. Scientia Horticulturae 96, Ito A, Hayama H, Kashimura Y (2004a) Possible roles of sugar concentration and its metabolism in the regulation of flower bud formation in Japanese pear (Pyrus pyrifolia). Acta Horticulturae 636,

317 Ito A, Yoshioka H, Hayama H, Kashimura Y (2004b) Reorientation of shoots to the horizontal position influences the sugar metabolism of lateral buds and shoot internodes in Japanese pear (Pyrus pyrifolia (Burm.) Nak.). Journal of Horticultural Science & Biotechnology 79, Jack T (2004) Molecular and genetic mechanisms of floral control. Plant Cell 16, S1- S17. Jackson DI (1991) Environmental and hormonal effects on development of early bunch stem necrosis. American Journal of Enology and Viticulture 42, Jackson DI, McNeil DL (1999) Walnuts. In 'Temperate and Subtropical Fruit Production'. (Eds D Jackson and N Looney) pp (CABI Publishing: Oxford) Jackson DI, Palmer JW (1999) Pome Fruits. In 'Temperate and Subtropical Fruit Production'. (Eds DI Jackson and N Looney) pp (CAB International: Wallingford) Jackson RS (1994) Grapevine structure and function. In 'Wine Science - Principles and applications' pp (Academic Press: London) Jacqmard A, Gadisseur I, Bernier G (2003) Cell division and morphological changes in the shoot apex of Arabidopsis thaliana during floral transition. Annals of Botany 91, Jansson S, Douglas CJ (2007) Populus: A model system for plant biology. Annual Review of Plant Biology 58, Jimenez CM, Diaz JBR (2003) A statistical model to estimate potential yields in peach before bloom. Journal of the American Society of Horticultural Science 128, Jo Y-S, Cho H-S, Liu I-S, Kim W-S (2008) Comparative photosynthetic characteristics among major Actinidia species. Acta Horticulturae 773, Jones KS, Paroschy J, McKersie BD, Bowley SR (1999) Carbohydrate composition and freezing tolerance of canes and buds in Vitis vinifera. Journal of Plant Physiology 155, Jonkers H (1979) Biennial bearing in apple and pear: A literature survey. Scientia Horticulturae 11, Kahana A, Silberstein L, Kessler N, Goldstein RS, Perl-Treves R (1999) Expression of ACC oxidase genes differs among sex genotypes and sex phases in cucumber. Plant Molecular Biology 41, Keller M (2010) 'The science of grapevines: anatomy and physiology.' (Academic Press) Keller M, Koblet W (1994) Is carbon starvation rather than excessive nitrogen supply the cause of inflorescence necrosis in Vitis vinifera L. Vitis 33, Keller M, Mills LJ, Wample RL, Spayd SE (2004) Crop load management in Concord grapes using different pruning techniques. American Journal of Enology and Viticulture 55, Keller M, Tarara JM (2010) Warm spring temperatures induce persistent seasonlong changes in shoot development in grapevines. Annals of Botany 106, Khan ZU, McNeil DL, Samad A (1998) Root pruning of apple trees grown at ultrahigh density affects carbohydrate reserves distribution in vegetative and reproductive growth. New Zealand Journal of Crop and Horticultural Science 26, Kinet JM (1993) Environmental, chemical and genetic control of flowering. Horticultural Reviews 15, Kinet JM, Sachs RM, Bernier G (1985) 'The Physiology of Flowering.' (CRC Press Inc: Florida) King RW (2012) Mobile signlas in day length-regulated flowering: Gibberellins, flowering locus T and sucrose. Russian Journal of Plant Physiology 59, King RW, Evans LT (2003) Gibberellins and flowering of grasses and cereals: Prizing open the lid of the "florigen" black box. Annual Review of Plant Biology 54,

318 Klages KU, Boldingh HL, Cooney JM, MacRae EA (2004) Planteose is a short-term storage carbohydrate in Actinidia leaves. Functional Plant Biology 31, Klages KU, Donnison H, Boldingh HL, MacRae EA (1998) myo-inositol is the major sugar in Actinidia arguta during early fruit development. Australian Journal of Plant Physiology 25, Kliewer WM (1965) Changes in concentration of glucose, fructose, and total soluble solids in flowers and berries of Vitis vinifera. American Journal of Enology and Viticulture 16, Kliewer WM (1966) Sugars and organic acids of Vitis vinifera. Plant Physiology 41, Koblet W (1969) Wanderung von assimilaten in rebtrieben und einfluss der blattflache auf ertrag und qualitat der trauben. Wein-Wiss 24, Koblet W, Candolfi- Vasconcelos MC, Aeschimann E, Howell GS (1993) Influence of defoliation, rootstock and training system on Pinot Noir grapevines. I. Mobilization and reaccumulation of assimilates in woody tissue. Viticultural and enological science 48, Koch K (2004) Sucrose metabolism: Regulatory mechanisms and pivotal roles in sugar sensing and plant development. Current Opinion in Plant Biology 7, Koornneef M, van der Veen JH (1980) Induction and anlysis of gibberellin sensitive mutants in Arabidopsis thaliana (L.) Heynh. Theoretical and Applied Genetics 58, Kotoda N, Hayashi H, et al. (2010) Molecular characterization of FLOWERING LOCUS T-like genes of apple (Malus x domestica Borkh.). Plant and Cell Physiology 51, Kotoda N, Iwanami H, Takahashi S, Abe K (2006) Antisense expression of MdTFL1, a TFL1-like gene, reduces the juvenile phase in apple. Journal of the American Society for Horticultural Science 131, Kotoda N, Wada M (2005) MdTFL1, a TFL1-like gene of apple, retards the transition from the vegetative to reproductive phase in transgenic Arabidopsis. Plant Science 168, Kotoda N, Wada M, Komori S, Kidou S, Abe K, Masuda T, Soejima J (2000) Expression pattern of homologues of floral meristem identity genes LFY and AP1 during flower development in apple. Journal of the American Society for Horticultural Science 125, Koussa T, Cherrad M, Bertrand A, Broquedis M (1998) Comparison of the contents of starch, soluble carbohydrates and abscisic acid of latent buds and internodes during the vegetative cycle of grapevine. Vitis 37, Kozlowski TT (1992) Carbohydrate sources and sinks in woody plants. Botanical Review 58, Kriedemann PE (1968) Photosynthesis in vine leaves as a function of light intensity, temperature and leaf age. Vitis 7, Kriedemann PE, Kliewer WM, Harris JM (1970) Leaf age and photosynthesis in Vitis vinifera L. Vitis 9, Lai R, Woolley DJ, Lawes GS (1990) The effect of inter-fruit competition, type of fruiting lateral and time of anthesis on the fruit-growth of kiwifruit (Actinidia deliciosa). Journal of Horticultural Science 65, Lakso AN, Robinson TL, Pool RM (1989) Canopy microclimate effects on patterns of fruiting and fruit development in apples and grapes. In 'Manipulation of fruiting'. (Ed. CJ Wright) pp (Butterworth & Co Plymouth) Lakso AN, White MD, Tustin DS (2001) Simulation modeling of the effects of short and long-term climatic variations on carbon balance of apple trees. Acta Horticulturae 557, Lakso AN, Wunsche JN, Palmer JW, Grappadelli LC (1999) Measurement and modeling of carbon balance of the apple tree. Hortscience 34, Lavee S (1986) Olive. In 'Handbook of Fruit Set and Development'. (Ed. SP Monselise) pp (CRC Press: Boca Raton) 282

319 Lavee S, May P (1997) Dormancy of grapevine buds - facts and speculation. Australian Journal of Grape and Wine Research 3, Lavee S, Regev U, Samish RM (1967) The determination of induction and differentiation in grape vines. Vitis 6, Lebon E, Pellegrino A, Tardieu F, Lecoeur J (2004a) Shoot development in grapevine (Vitis vinifera) is affected by the modular branching pattern of the stem and intra- and inter-shoot trophic competition. Annals of Botany 93, Lebon G, Brun O, Magne C, Clement C (2005) Photosynthesis of the grapevine (Vitis vinifera) inflorescence. Tree Physiology 25, Lebon G, Duchene E, Brun O, Magne C, Clement C (2004b) Flower abscission and inflorescence carbohydrates in sensitive and non-sensitive cultivars of grapevine. Sexual Plant Reproduction 17, Lebon G, Wojnarowiez G, Holzapfel B, Fontaine F, Vaillant-Gaveau N, Clement C (2008) Sugars and flowering in the grapevine (Vitis vinifera L.). Journal of Experimental Botany 59, Lejeune P, Bernier G, Kinet JM (1991) Sucrose levels in leaf exudate as a function of floral induction in the long day plant Sinapis alba. Plant Physiology and Biochemistry 29, Lejeune P, Bernier G, Requier MC, Kinet JM (1993) Sucrose increase during floral induction in the phloem sap collected at the apical part of the shoot of the long-day plant Sinapis alba L. Planta 190, Lemmetyinen J, Hassinen M, Elo A, Porali I, Keinonen K, Makela H, Sopanen T (2004) Functional characterization of SEPALLATA3 and AGAMOUS orthologues in silver birch. Physiologia Plantarum 121, Lenahan OM, Whiting MD, Elfving DC (2006) Gibberellic acid inhibits floral bud induction and improves 'Bing' sweet cherry fruit quality. Hortscience 41, Lionakis S (1983) Regression coefficients for computing leaf area in kiwifruit plants (Actinidia chinensis Planch.) grown under different daylength and temperature conditions. Fruits 38, Lionakis SM, Schwabe WW (1984) Some effects of daylength, temperature and exogenous growth regulator application on the growth of Actinidia chinensis Planch. Annals of Botany 54, Loescher WH, McCamant T, Keller JD (1990) Carbohydrate reserves, translocation, and storage in woody plant-roots. Hortscience 25, Looney N, Jackson D (1999) Stonefruit. In 'Temperate and Subtropical Fruit Production'. (Eds DI Jackson and NE Looney) pp (CAB International Wallingford) Lorenz D, Eichorn K, Bleiholder H, Klose R, Meier U, Weber E (1994) Phanologische entwicklungsstadien der weinrebe (Vitis vinifera L. ssp. vinifera). - Codierung und beschreibung nach der erweiterten BBCH-sckaa-. Viticultural and enological science 49, Lunn JE, Feil R, et al. (2006) Sugar-induced increases in trehalose 6-phosphate are correlated with redox activation of ADPglucose pyrophosphorylase and higher rates of starch synthesis in Arabidopsis thaliana. Biochemical Journal 397, Lyrene PM, Williamson JG (1997) High-density blueberry plantings in Florida. Acta Horticulturae 446, MacRae EA, Quick WP, Benker C, Stitt M (1992) Carbohydrate metabolism during postharvest ripening in kiwifruit. Planta 188, Makela P, McLaughlin JE, Boyer JS (2005) Imaging and quantifying carbohydrate transport to the developing ovaries of maize. Annals of Botany 96, Manson PJ, Snelgar WP, Stowell BM (1994) Proximity to shelterbelts influences flowering and yield of kiwifruit (Actinidia deliciosa). Journal of Horticultural Science 69, Matsoukas IG, Massiah AJ, Thomas B (2012) Florigenic and antiflorigenic signaling in plants. Plant and Cell Physiology 53,

320 Matthews MA, Anderson MM (1989) Reproductive development in grape (Vitis vinifera L.): Responses to seasonal water deficits. American Journal of Enology and Viticulture 40, Maurel K, Leite GB, Bonhomme M, Guilliot A, Rageau R, Petel G, Sakr S (2004) Trophic control of bud break in peach (Prunus persica) trees: a possible role of hexoses. Tree Physiology 24, May P, Antcliff AJ (1963) The effect of shading on fruitfulness and yield in the Sultana. Journal of Horticultural Science 38, May PB (2000) From bud to berry, with special reference to inflorescence and bunch morphology in Vitis vinifera L. Australian Journal of Grape and Wine Research 6, May PB (2004) 'Flowering and fruitset in grapevines.' (Lythrum Press: Adelaide) McArtney SJ, Ferree DC (1999a) Root and cane pruning affect vegetative development, fruiting, and dry-matter accumulation of grapevines. Hortscience 34, McArtney SJ, Ferree DC (1999b) Shading effects on dry matter partitioning, remobilization of stored reserves and early season vegetative development of grapevines in the year after treatment. Journal of the American Society for Horticultural Science 124, McKenzie DW, Rae AN (1978) Economics of high density apple production in New Zealand. Acta Horticulturae 65, McLaughlin JE, Boyer JS (2004) Glucose localization in maize ovaries when kernel number decreases at low water potential and sucrose is fed to the stems. Annals of Botany 94, McPherson HG, Hall AJ, Stanley CJ (1994) Seasonal and regional variation in budbreak and flowering of kiwifruit vines in New Zealand. New Zealand Journal of Crop and Horticultural Science 22, McPherson HG, Richardson AC, Snelgar WP, Currie MB (2001) Effects of hydrogen cyanamide on budbreak and flowerings in kiwifruit (Actinidia deliciosa 'Hayward'). New Zealand Journal of Crop and Horticultural Science 29, McPherson HG, Stanley CJ, Warrington IJ (1995) The response of bud break and flowering to cool winter temperatures in kiwifruit (Actinidia deliciosa). Journal of Horticultural Science 70, Moncur MW, Rattigan K, Mackenzie DH, McIntyre GN (1989) Base temperature for budbreak and leaf appearance of grapevines. American Journal of Enology and Viticulture 40, Monselise SP, Goldschmidt EE (1982) Alternate bearing in fruit trees. Horticultural Reviews 4, Mooney HA, Gartner BL (1991) Reserve economy of vines. In 'The Biology of Vines'. (Eds FE Putz and HA Mooney) pp (Press Syndicate of the University of Cambridge: Cambridge) Mooney HA, Winner WE (1991) Partitioning response of plants to stress. In 'Response of Plants to Multiple Stresses'. (Eds HA Mooney, WE Winner and EJ Pell) pp (Academic Press Inc.: San Diego) Moore JN, Brown MV, Bordelon BP (1993) Yield and fruit size of 'Bluecrop' highbush blueberries at three plant spacings. Hortscience 28, Morley-Bunker MJ (1999a) Avocados. In 'Temperate and Subtropical Fruit Production'. (Eds D Jackson and N Looney). (CABI Publishing: Oxford) Morley-Bunker MJ (1999b) Olives. In 'Temperate and Subtropical Fruit Production'. (Eds D Jackson and N Looney). (CABI Publishing: Oxford) Morley-Bunker MJ, McNeil DL (1999) Macadamias. In 'Temperate and Subtropical Fruit Production'. (Eds D Jackson and N Looney) pp (CABI Publishing: Oxford) Morton JF (1987) Passionfruit. In 'Fruits of warm climates'. (Ed. JF Morton). (Morton JF: Miami) 284

321 Mullins MG, Bouquet A, Williams LE (1992) 'Biology of the grapevine.' (Cambridge University Press: Cambridge.) Mullins MG, Rajasekaran K (1981) Fruiting cuttings: Revised method for producing test plants of grapevine cultivars. American Journal of Enology and Viticulture 32, Munoz-Fambuena N, Mesejo C, Gonzalez-Mas MC, Iglesias DJ, Primo-Millo E, Agusti M (2012) Gibberellic acid reduces flowering intensity in sweet orange (Citrus sinensis (L.) Osbeck) by repressing CiFT gene expression. Journal of Plant Growth Regulation 31, Mutters RG, Hall AE (1992) Reproductive responses of cowpea to high-temperature during different night periods. Crop Science 32, Nardozza S, Boldingh HL, et al. (2013) Metabolic analysis of kiwifruit (Actinidia deliciosa) berries from extreme genotypes reveals hallmarks for fruit starch metabolism. Journal of Experimental Botany 64, Nemhauser JL, Feldman LJ, Zambryski PC (2000) Auxin and ETTIN in gynoecium morphogenesis. Development 127, Ninot A, Rovira M, Aleta N (2006) Mechanical pruning in walnut hedgerow orchard. Acta Horticulturae 705, Nuzzo V, Biasi R, Dichio B, Montanaro G, Xiloyannis C, Lanzieri A (1999) Influence of different seasonal light availability on flower bud quality in cv Tirynthos (Prunus armeniaca L.). Acta Horticulturae 488, Okuse I, Ryugo K (1981) Compositional changes in the developing 'Hayward' kiwifruit in California. Journal American Society of Horticultural Science 106, Oliveira CM, Priestley CA (1988) Carbohydrate reserves in deciduous fruit trees. Horticultural Reviews 10, Oukabli A, Bartolin S, Viti R (2003) Anatomical and morphological study of apple (Malus x domestica Borkh.) flower buds growing under inadequate winter chilling. Journal of Horticultural Science & Biotechnology 78, Palliotti A, Cartechini A, Nasini L, Silvestroni O, Mattioli S, Neri D (2004) Seasonal carbon balance of 'Sangiovese' grapevines in two different central Italy environments. Acta Horticulturae 652, Palliotti A, Cartechini A, Silvestroni O, Mattioli S, Petoumenou D, Berrios JG (2007) Long-term effects of seeded cover-crop on vegetative characteristics, yield and grape and wine composition of 'Grechetto' grapevines in central Italy. Acta Horticulturae 754, Palliotti A, Silvestroni O, Petoumenou D (2010) Seasonal patterns of growth rate and morphophysiological features in green organs of Cabernet Sauvignon grapevines. American Journal of Enology and Viticulture 61, Palmer JW (2007) Apples and kiwifruit, can we learn from each other? Acta Horticulturae 753, Palmer JW, Dryden G (2006) Fruit mineral removal rates from New Zealand apple (Malus domestica) orchards in the Nelson region. New Zealand Journal of Crop and Horticultural Science 34, Palmer JW, Prive JP, Tustin DS (2003) Temperature. In 'Apples: Botany, Production and Uses'. (Eds DC Ferree and IJ Warrington) pp (CABI Publishing: Wallingford, UK) Palmer JW, Wunsche JN, Meland M, Hann A (2002) Annual dry-matter production by three apple cultivars at four within-row spacings in New Zealand. Journal of Horticultural Science & Biotechnology 77, Perilleux C, Bernier G, Kinet JM (1991) Reproductive development in Lolium temulentum L. - Spike morphogenesis and grain set limitations. Journal of Experimental Botany 42, Petrie PR, Clingeleffer PR (2005) Effects of temperature and light (before and after budburst) on inflorescence morphology and flower number of Chardonnay grapevines (Vitis vinifera L.). Australian Journal of Grape and Wine Research 11,

322 Petrie PR, Trought MCT, Howell GS (2000) Growth and dry matter partitioning of Pinot Noir (Vitis vinifera L.) in relation to leaf area and crop load. Australian Journal of Grape and Wine Research 6, Petrie PR, Trought MCT, Howell GS, Buchan GD (2003) The effect of leaf removal and canopy height on whole-vine gas exchange and fruit development of Vitis vinifera L. Sauvignon Blanc. Functional Plant Biology 30, Petrie PR, Trought MCT, Howell GS, Buchan GD, Palmer JW (2009) Whole-canopy gas exchange and light interception of vertically trained Vitis vinifera L. under direct and diffuse light. American Journal of Enology and Viticulture 60, Piller GJ, Greaves AJ, Meekings JS (1998) Sensitivity of floral shoot growth, fruit set and early fruit size in Actinidia deliciosa to local carbon supply. Annals of Botany 81, Piller GJ, Meekings JS (1997) The acquisition and utilization of carbon in early spring by kiwifruit shoots. Annals of Botany 79, Pillitteri LJ, Lovatt CJ, Walling LL (2004a) Isolation and characterization of a TERMINAL FLOWER homolog and its correlation with juvenility in citrus. Plant Physiology 135, Pillitteri LJ, Lovatt CJ, Walling LL (2004b) Isolation and characterization of LEAFY and APETALA1 homologues from Citrus sinensis L. Osbeck 'Washington'. Journal of the American Society for Horticultural Science 129, Polito VS, Grant JA (1984) Initiation and development of pistillate flowers in Actinidia chinensis. Scientia Horticulturae 22, Poni S, Bernizzoni F, Civardi S (2007) The issue of canopy efficiency in the grapevine: Assessment and approaches for its improvement. Acta Horticulturae 754, Poni S, Casalini L, Bernizzoni F, Civardi S, Intrieri C (2006) Effects of early defoliation on shoot photosynthesis, yield components, and grape composition. American Journal of Enology and Viticulture 57, Poni S, Intrieri C (2001) Grapevine photosynthesis: effects linked to light radiation and leaf age. Advance Horticultural Science 15, Potuschak T, Doerner P (2001) Cell cycle controls: Genome-wide analysis in Arabidopsis. Current Opinion in Plant Biology 4, Pratt C (1971) Reproductive anatomy in cultivated grapes-a review. American Journal of Enology and Viticulture 22, Pratt C (1979) Shoot and bud development during the prebloom period of Vitis. Vitis 18, 1-5. Priestley CA, Catlin PB, Olsson EA (1976) The distribution of 14C-labelled assimilates in young apple trees as influenced by doses of supplementary nitrogen. I. Total 14C radioactivity in extracts. Annals Botany 40, Putterill J, Laurie R, Macknight R (2004) It's time to flower: the genetic control of flowering time. Bioessays 26, Ramos A, Marques L, Lopes R, Ribeiro E, Martins JMS (2008) The "yield potential' - a new methodology to relate orchard productivity and fruit quality. Acta Horticulturae 800, Redalen G (1976) Pollination and fruit set in raspberries. Acta Horticulturae 60, Reid JB, Sorensen I, Petrie RA (1993) Root demography in kiwifruit (Actinidia deliciosa). Plant, Cell and Environment 16, Reinoso H, Luna V, Pharis RP, Bottini R (2002) Dormancy in peach (Prunus persica) flower buds. V. Anatomy of bud development in relation to phenological stage. Canadian Journal of Botany-Revue Canadienne De Botanique 80, Richardson AC, Boldingh HL, McAtee PA, Gunaseelan K, Luo ZW, Atkinson RG, David KM, Burdon JN, Schaffer RJ (2011) Fruit development of the diploid kiwifruit, Actinidia chinensis 'Hort16A'. BMC Plant Biology 11. Richardson AC, Marsh KB, et al. (2004) High growing temperatures reduce fruit carbohydrate and vitamin C in kiwifruit. Plant Cell and Environment 27,

323 Richardson AC, McAneney KJ, Dawson TE (1997) Carbohydrate dynamics in kiwifruit. Journal of Horticultural Science 72, Richardson AC, Walton EF, Boldingh HL, Meekings JS (2007) Seasonal carbohydrate changes in dormant kiwifruit buds. Acta Horticulturae 753, Richardson AC, Walton EF, Meekings JS, Boldingh HL (2010) Carbohydrate changes in kiwifruit buds during the onset and release from dormancy. Scientia Horticulturae 124, Rinne PLH, Kaikuranta PM, van der Schoot C (2001) The shoot apical meristem restores its symplasmic organization during chilling-induced release from dormancy. Plant Journal 26, Rogiers SY, Clarke SJ (2013) Nocturnal and daytime stomatal conductance respond to root-zone temperatures in 'Shiraz' grapevines. Annals of Botany 111, Rogiers SY, Greer DH, Hutton RJ, Landsberg JJ (2009) Does night-time transpiration contribute to anisohydric behaviour in a Vitis vinifera cultivar? Journal of Experimental Botany 60, Rogiers SY, Smith JP, Holzapfel BP, Hardie WJ (2011) Soil temperature moderates grapevine carbohydrate reserves after bud break and conditions fruit set responses to photoassimilatory stress. Functional Plant Biology 38, Roitsch T, Ehness R, Goetz M, Hause B, Hofmann M, Sinha AK (2000) Regulation and function of extracellular invertase from higher plants in relation to assimilate partitioning, stress responses and sugar signalling. Australian Journal of Plant Physiology 27, Rovira M, Ninot A, Aleta N (2001) Pistillate flower abortion in walnut (J. regia L.). Acta Horticulturae 544, Rowe N, Speck T (2005) Plant growth forms: An ecological and evolutionary perspective. New Phytologist 166, Rudich J, Zamski E, Regev Y (1977) Genotypic variation for sensitivity to high temperature in the tomato: Pollination and fruit set. Botanical Gazette 138, Sala A, Woodruff DR, Meinzer C (2012) Carbon dynamics in trees: feast or famine? Tree Physiology 32, Sale PR (2003) Good returns possible from passionfruit but the risks are high. The Orchardist 76, Sale PR, Lyford PB (1990) Cultural, management and harvesting practices for kiwifruit in New Zealand. In 'Kiwifruit: Science and managment'. (Eds IJ Warrington and GC Weston). (Ray Richards in association with the New Zealand Society for Horticultural Science,: Auckland) Salinero MC, Vela P, Sainz MJ (2009) Phenological growth stages of kiwifruit (Actinidia deliciosa 'Hayward'). Scientia Horticulturae 121, Samach A, Smith HM (2013) Constraints to obtaining consistent annual yields in perennials. II: Environment and fruit load affect induction of flowering. Plant Science 207, Sanchez LS, Dokoozlian NK (2005) Bud microclimate and fruitfulness in Vitis vinifera L. American Journal of Enology and Viticulture 56, Satoh-Nagasawa N, Nagasawa N, Malcomber S, Sakai H, Jackson D (2006) A trehalose metabolic enzyme controls inflorescence architecture in maize. Nature 441, Scholefield PB, Neales TP, May P (1978) Carbon balance of the sultana vine (Vitis vinifera L.) and the effects of autumn defoliation by harvest pruning. Australian Journal of Plant Physiology 5, Sedgley M (1988) Flowering of deciduous perennial fruit crops. Horticultural Reviews, Seleznyova AN, Greer DH (2001) Effects of temperature and leaf position on leaf area expansion of kiwifruit (Actinida deliciosa) shoots: Development of a modelling framework. Annals of Botany 88,

324 Sepulveda GR, Kliewer WM (1983) Estimation of leaf area of two grapevine cultivars (Vitis vinifera L.) using laminae linear measurements and fresh weight. American Journal of Enology and Viticulture 34, Shavrukov YN, Dry IB, Thomas MR (2004) Inflorescence and bunch architecture development in Vitis vinifera L. Australian Journal of Grape and Wine Research 10, Shellie KC (2010) Water deficit effect on ratio of seed to berry fresh weight and berry weight uniformity in winegrape cv. Merlot. American Journal of Enology and Viticulture 61, Skene KGM (1972) Effect of ringing on cytokinin activity in shoots of the grape vine. Journal of Experimental Botany 23, Skinner PW, Matthews MA (1989) Reproductive development in grape (Vitis vinifera L.) under phosphorus-limited conditions. Scientia Horticulturae 38, Smagula JM (1993) Effect of boron on lowbush blueberry fruit set and yield. Acta Horticulturae 346, Smart RE (1985) Principles of grapevine canopy microclimate manipulation with implications for yield and quality - a review. American Journal of Enology and Viticulture 36, Smith AM, Stitt M (2007) Coordination of carbon supply and plant growth. Plant, Cell and Environment 30, Smith GS, Buwalda JG (1994) Kiwifruit. In 'CRC Handbook of environmental physiology of fruit crops' pp (CRC Press Incorporated: Boca Raton, Florida) Smith GS, Clark CJ, Boldingh HL (1992) Seasonal accumulation of starch by components of the kiwifruit vine. Annals of Botany 70, Smith JP, Holzapfel BP (2009) Cumulative responses of grapevine (Vitis vinifera L.) to late season perturbation of carbohydrate reserve status. American Journal of Enology and Viticulture 60, Smith RJ, Kliewer WM (1984) Estimation of Thompson seedless grapevine leaf area. American Journal of Enology and Viticulture 35, Snelgar WP, Hall AJ, McPherson HG (2008) Modelling flower production of kiwifruit (Actinidia deliciosa) from winter chilling. New Zealand Journal of Crop and Horticultural Science 36, Snelgar WP, Manson PJ (1992) Determination of the time of flower evocation in kiwifruit vines. New Zealand Journal of Crop and Horticultural Science 20, Snelgar WP, Manson PJ, Hopkirk G (1991) Effect of overhead shading on fruit size and yield potential of kiwifruit (Actinidia deliciosa). Journal of Horticultural Science 66, Snelgar WP, Manson PJ, Martin PJ (1992) Influence of time of shading on flowering and yield of kiwifruit vines. Journal of Horticultural Science 67, Snelgar WP, Manson PJ, McPherson HG (1997) Evaluating winter chilling of kiwifruit using excised canes. Journal of Horticultural Science 72, Snowball AM (1995) The seasonal cycle of leaf, shoot and bud development in kiwifruit. Journal of Horticultural Science 70, Snowball AM (1996) The timing of flower evocation in kiwifruit. Journal of Horticultural Science 71, Snowball AM (1997a) Axillary shoot bud development in selected Actinidia species. New Zealand Journal of Crop and Horticultural Science 25, Snowball AM (1997b) Seasonal cycle of shoot development in selected Actinidia species. New Zealand Journal of Crop and Horticultural Science 25, Soar CJ, Speirs J, Maffei SM, Loveys BR (2004) Gradients in stomatal conductance, xylem sap ABA and bulk leaf ABA along canes of Vitis vinifera cv. Shiraz: molecular and physiological studies investigating their source. Functional Plant Biology 31,

325 Solomakhin A, Blanke MM (2008) Coloured hailnets alter light transmission, spectra and phytochrome, as well as vegetative growth, leaf chlorophyll and photosynthesis and reduce flower induction of apple. Plant Growth Regulation 56, Sommer K, Islam M, Clingeleffer P (2000) Light and temperature effects on shoot fruitfulness in Vitis vinifera L. cv. Sultana: Influence of trellis type and grafting. Australian Journal of Grape and Wine Research 6, Song G, Walworth A, Zhao D, Jiang N, Hancock JF (2013) The Vaccinum corymbosum FLOWERING LOCUS T-like gene (VcFT): a flowering activator reverses photoperiodic and chilling requirements in blueberry. Plant Cell Reports 32, Spiegel-Roy P, Goldschmidt EE (1996) 'Biology of Citrus.' (Cambridge University Press: Cambridge) Srinivasan C, Mullins MG (1980) Effects of temperature and growth regulators on formation of anlagen, tendrils and inflorescences in Vitis vinifera L. Annals of Botany 45, Srinivasan C, Mullins MG (1981) Physiology of flowering in the grapevine - A review. American Journal of Enology and Viticulture 32, Stanton MA, Scheerens JC, Funt RC, Clark JR (2007) Floral competence of primocane-fruiting blackberries Prime-Jan and Prime-Jim grown at three temperature regimens. Hortscience 42, Stephenson AG (1981) Flower and fruit abortion: Proximate causes and ultimate functions. Annual Revew of Ecological Systems 12, Stitt M, Gibon Y, Lunn JE, Piques M (2007) Multilevel genomics analysis of carbon signalling during low carbon availability: Coordinating the supply and utilisation of carbon in a fluctuating environment. Functional Plant Biology 34, Sun K, Hunt K, Hauser BA (2004) Ovule abortion in Arabidopsis triggered by stress. Plant Physiology 135, Swanepoel JJ, Archer E (1988) The ontogeny and development of Vitis vinifera L Cv Chenin Blanc inflorescence in relation to phenological stages. Vitis 27, Takeda F, Strik BC, Peacock D, Clark JR (2002) Cultivar differences and the effect of winter temperature on flower bud development in blackberry. Journal of the American Society for Horticultural Science 127, Thiele GF (1999) Berry fruit. In 'Temperate and Subtropical Fruit Production'. (Eds D Jackson and N Looney) pp (CABI Publishing: Oxford) This P, Lacombe T, Thomas MR (2006) Historical origins and genetic diversity of wine grapes. Trends in Genetics 22, Thomas JE, Barnard C (1937) Fruit bud studies. III. The Sultana: Some relations between shoot growth, chemical composition, fruit bud formation and yield. Journal of the Council for Scientific and Industrial Research 10, Thompson E, Clark JR, Strik BC, Finn CE (2008) Flowering and fruiting morphology of primocane-fruiting blackberries. Acta Horticulturae 777, Thornton I, El-Zeftawi B (1983) 'Culture of irrigated citrus fruit.' (Victorian Government Printing Office: Melbourne) Thorp TG, Barnett AM, Miller SA (2003) Effects of cane size and pruning system on shoot growth, flowering and productivity of 'Hayward' kiwifruit vines. Journal of Horticultural Science & Biotechnology 78, Thorp TG, Jie Z, Lay-Yee M (1990) Horticultural characteristics of seven pistillate and three staminate New Zealand cultivars of kiwifruit. N.Z. Journal of Crop and Horticultural Science 18, Trinder P (1969) Determination of glucose in blood using glucose oxidase with an alternative oxygen accepter. Annals of Clinical Biochemistry 6, Tromp J (2000) Flower-bud formation in pome fruits as affected by fruit thinning. Plant Growth Regulation 31, Tukey HB, Young JO (1942) Gross morphology and histology of developing fruit of the apple. Botanical Gazette 104,

326 Turck F, Fornara F, Coupland G (2008) Regulation and identity of florigen: FLOWERING LOCUS T moves center stage. Annual Review of Plant Biology 59, Turner AD, Wien HC (1994) Dry-matter assimilation and partitioning in pepper cultivars differing in susceptibility to stress-induced bud and flower abscission. Annals of Botany 73, Varkonyi-Gasic E, Moss SM, Voogd C, Wu RM, Lough RH, Wang Y, Hellens RP (2011) Identification and characterization of flowering genes in kiwifruit: sequence conservation and role in kiwifruit flower development. BMC Plant Biology 11, 72. Varkonyi-Gasic E, Moss SMA, Voogd C, Wang T, Putterill J, Hellens RP (2013) Homologs of FT, CEN, and FD respond to developmental and environmental signals affecting growth and flowering in the perennial vine kiwifruit. New Phytologist 198, Vasconcelos MC, Greven M, Winefield CS, Trought MCT, Raw V (2009) The flowering process of Vitis vinifera: A review. American Journal of Enology and Viticulture 60, Walton EF (1995) Occurence of secondary floral shoots in summer on Actinidia eriantha Benth. N.Z. Journal of Crop and Horticultural Science 23, Walton EF, De Jong TM (1990) Growth and compositional changes in kiwifruit berries from three Californian locations. Annals of Botany 66. Walton EF, Fowke PJ (1993) Effect of hydrogen cyanamide on kiwifruit shoot flower number and position. Journal of Horticultural Science 68, Walton EF, Fowke PJ (1995) Estimation of the annual cost of kiwifruit vine growth and maintenance. Annals of Botany 76, Walton EF, Fowke PJ, Weis K, McLeay PL (1997) Shoot axillary bud morphogenesis in kiwifruit (Actinidia deliciosa). Annals of Botany 80, Walton EF, Podivinsky E, Wu RM (2001) Bimodal patterns of floral gene expression over the two seasons that kiwifruit flowers develop. Physiologia Plantarum 111, Walton EF, Richardson AC, Waller JE, Dow BW (2000) Effect of time of cane initiation on subsequent fruitfulness in kiwifruit. New Zealand Journal of Crop and Horticultural Science 28, Wample RL, Bary A (1992) Harvest date as a factor in carbohydrate storage and cold hardiness of Cabernet Sauvignon grapevines. Journal of the American Society for Horticultural Science 117, Warrington IJ, Dunstone RL, Green LM (1977) Temperature effects at three development stages on the yield of the wheat ear. Australian Journal of Agricultural Research 28, Watson M, Gould KS (1994) Development of flat and fan-shaped fruit in Actinidia chinensis var chinensis and Actinidia deliciosa. Annals of Botany 74, Watt AM, Dunn GM, May PB, Crawford SA, Barlow EWR (2008) Development of inflorescence primordia in Vitis vinifera L. cv. Chardonnay from hot and cool climates. Australian Journal of Grape and Wine Research 14, Weaver RJ, McCune SB (1960) Effects of overcropping Alicante Boushet grapevines in relation to carbohydrate nutrition and development of the vine. Proceedings of the American Soceity of Horticultural Science 75, Weyand KM, Schultz HR (2006) Long-term dynamics of nitrogen and carbohydrate reserves in woody parts of minimally and severely pruned riesling vines in a cool climate. American Journal of Enology and Viticulture 57, Wilkie JD, Sedgley M, Olesen T (2008) Regulation of floral initiation in horticultural trees. Journal of Experimental Botany 59, Williams DW, Andris HL, Beede RH, Luvisi DA, Norton MVK, Williams LE (1985a) Validation of a model for the growth and development of the Thompson Seedless grapevine. II. Phenology. American Journal of Enology and Viticulture 36, Williams DW, Williams LE, Barnett WW, Kelley KM, McKenry MV (1985b) Validation of a model for the growth and development of the Thompson Seedless grapevine. I. 290

327 Vegetative growth and fruit yield. American Journal of Enology and Viticulture 36, Williams L, Martinson TE (2003) Nondestructive leaf area estimation of 'Niagara' and 'DeChaunac' grapevines. Scientia Horticulturae 98, Williams LE (1996) Grape. In 'Photo-assimilate distribution in plants and crops. Source-sink relationships.'. (Eds E Zamski and A Schaffer). (Marcel Dekker: New York) Williams LE, Matthews MA (1990) Grapevine. In 'Irrigation of agricultural crops'. (Eds BA Stewart and DR Nielsen) pp (American Society of Agronomy: Madison, USA) Williams LE, Retzlaff WA, Yang WG, Biscay PJ, Ebisuda N (2000) Effect of girdling on leaf gas exchange, water status, and non-structural carbohydrates of field-grown Vitis vinifera L. (cv. Flame seedless). American Journal of Enology and Viticulture 51, Wilson RN, Heckman JW, Somerville CR (1992) Gibberellin is required for flowering in Arabidopsis thaliana under short days. Plant Physiology 100, Winkler AJ, Williams WO (1945) Starch and sugars of Vitis vinifera. Plant Physiology 20, Wolstenholme BN (1986) Energy costs of fruiting as a yield-limiting factor with special reference to avocado. Acta Horticulturae 175, Wu BH, Liu HF, Guan L, Fan PG, Li SH (2011) Carbohydrate metabolism in grape cultivars that differ in sucrose accumulation. Vitis 50, Wunsche JN, Greer DH, Laing WA, Palmer JW (2005) Physiological and biochemical leaf and tree responses to crop load in apple. Tree Physiology 25, Yang YS, Hori Y (1979) Studies on the retranslocation of accumulated assimilates in 'Delaware' grapevines. I. Retranslocation of 14 C-assimilates in the following spring after 14 C feeding in summer and autumn. Tohoku Journal of Agricultural Research 30, Yang YS, Hori Y, Ogata R (1980) Studies of retranslocation of accumulated assimilates in 'Delaware' grapevines. II. Retranslocation of assimilate accumulated during the previous growing season. Tohoku Journal of Agricultural Research 31, Zapata C, Deleens E, Chaillou S, Magne C (2004a) Mobilisation and distribution of starch and total N in two grapevine cultivars differing in their susceptibility to shedding. Functional Plant Biology 31, Zapata C, Deleens E, Chaillou S, Magne C (2004b) Partitioning and mobilization of starch and N reserves in grapevine (Vitis vinifera L.). Journal of Plant Physiology 161, Zik M, Irish VF (2003) Flower development: Initiation, differentiation, and diversification. Annual Review of Cell and Developmental Biology 19, Zucconi F (1986) Peach. In 'Handbook of Fruit Set and Development'. (Ed. SP Monselise) pp (CRC Press: Boca Raton) 291

328 292

329 293 Appendix 1 Kiwifruit and grape pest and disease programme 293

330