Refining Fruit-Zone Leaf Removal for Red-Fruited Bordeaux Varieties Grown in a Humid Environment. Cain Charles Hickey

Transcription

1 Refining Fruit-Zone Lef Removl for Red-Fruited Bordeux Vrieties Grown in Humid Environment. Cin Chrles Hickey Disserttion sumitted to the fculty of the Virgini Polytechnic Institute nd Stte University in prtil fulfillment of the requirements for the degree of Doctor of Philosophy In Horticulture Tony K. Wolf, Committee Chir Mish T. Kwsniewski, Co-chir Molly K. Kelly Gregory M. Peck Amnd C. Stewrt My 2, 2016 Blcksurg, VA Keywords: cnopy mngement, lef removl, pre-loom, fruit set, rdition, temperture, grpe, nthocynins, phenolics, crotenoids, Cernet Suvignon, Cernet frnc, Petit Verdot

2 Refining fruit-zone lef removl for red-fruited Bordeux vrieties grown in humid environment. Cin Chrles Hickey Current fruit-zone mngement recommendtion in the estern US ims for 1-2 sl shoot lef lyers fter fruit set to limit fungl disese nd sunurn incidence, nd prevent extreme heting of grpes. The gol of this work ws to ssess if fruit-zone lef removl to n uncommonly greter extent, nd/or t n erlier phenologicl stge, would fvorly lter yield components or fruit composition in three populrly grown, red-fruited, Bordeux vrieties Cernet frnc, Petit Verdot, nd Cernet Suvignon. Pre-loom lef removl to vrious extents reduced crop yield y 41-78% when compred to no lef removl cross sesons nd vrieties. Pre-loom lef removl implementtion in consecutive sesons tended to further reduce crop yield components compred to implementtion in the first yer. Pre-loom lef removl tended to reduce cluster compctness nd unch rot incidence when compred to post-fruit set nd no lef removl. Bsl lef removl to the gretest extents inconsistently reduced solule solids nd titrtle cidity cross vrieties nd sesons. Pre-loom nd post-fruit set lef removl to the gretest extent consistently incresed totl grpe phenolics nd nthocynins compred to no lef removl in Cernet Suvignon, ut inconsistently incresed totl grpe phenolics compred to no lef removl in Cernet frnc nd Petit Verdot. Bsl lef removl to the gretest extents tended to increse the synthesis nd degrdtion of crotenoids more consistently thn no lef removl, nd this ws prticulrly true for zexnthin. Petit Verdot nd Cernet frnc wine color nd rom were inconsistently distinguishle etween lef removl tretments, nd color intensity ws rted higher in wines mde with fruit from pre-loom lef removl compred to modest post fruit-set lef removl plots. Witing until fter fruit set to remove fruit-zone leves mintined crop yield nd offered comprle improvements in fruit composition to pre-loom lef removl. Pre-loom lef removl of no more thn four leves is recommended to limit crop yield reduction, nd modestly improve fruit composition. This work showed tht fruit-zone lef removl does not need to e conservtive in the estern US, prticulrly ecuse the climte does not pper to e detrimentl to fruit composition, nd open fruit-zones reduce grpe fungl disese incidence.

3 Acknowledgements I m grteful for Tony s guidnce nd support throughout my tenure s grdute student. I fil t emulting his dediction nd undying work ethic. However, I m hrder worker thn I ever would hve een hd he not een my dvisor. Thnk you for ll you hve done for me over the lst six yers. This is just not possile without your guidnce, support, ptience nd kindness. I will certinly miss your presence down the hllwy s my creer continues, ut trust we will remin collegues nd friends for the rest of our lives. Thnks to my committee for your service nd your desire to help me think criticlly nd develop professionlly. Thnks in prticulr to Mish, who ws lwys more thn willing to help me with nything I sked, prticulrly with the crotenoid portion of this work. Thnks to every single person who hs een employed t the Alson H. Smith, Jr. Agriculturl Reserch nd Experiment Center. You ll welcomed me with open rms nd mde my tenure t the AREC n enjoyle time. I d especilly like to thnk Tremin, who ws one of the very first friends I mde fter moving to Winchester. Thnks for ll of your help throughout my grdute reserch creer. Your dvice ws lwys helpful. I d lso like to thnk Hnnh, Dni, Trevor, nd Rchel. You ll were exemplry helpers. This work is yours just s much s it is mine. Thnks to the Virgini Tech Horticulture Deprtment. The short time I spent in Blcksurg ws mde etter due to ll of the wonderful people in this deprtment. I d like to thnk Roger, the est deprtment hed t Virgini Tech. Roger shows tht hrd work is importnt, ut eing good person is perhps more importnt. I ll miss ll of our converstions out good music. Thnks to Terry Btes for supporting me over the lst few months s I worked through finishing this disserttion. I look forwrd to working together over the next severl yers. I m in det to you ig time. Prefce/ttriution This work ws not possile without the help of those mentioned ove, ut prticulrly the help of Tony Wolf, Mish Kwsniewski, nd Jim Meyers. Tony is co-uthor on ll chpters, s he supervised ll spects, nd extensively edited the work. Mish Kwsniewski is co-uthor on Chpter 1, s he directed the crotenoid extrction nd quntifiction methods. Jim Meyers is co-uthor on Chpter 2, s he developed the erry temperture prediction models. I thnk these gentlemen for their fine contriutions. iii

4 Dediction I dedicte this work to Lis, my ever-loving wife. She hs dedicted her time nd ptience into this process s much s I hve. In fct, I would not hve gone down this creer pth hd it not een for her constnt encourgement nd support. Thnk you for lwys letting me disclose my weighted mind, nd for responding with logicl dvice nd guidnce. Thnks for elieving in me, nd for lwys knowing wht is etter for me thn I do myself. I cnnot wit to e prents together! I love you. iv

5 Tle of contents Section Pge Astrct...ii Acknowledgements, prefce/ttriution...iii Dediction...iv Chpter 1. Extent nd timing of fruit-zone lef nd lterl shoot removl lters yield components, grpe phenolics, nd grpe crotenoids in Cernet frnc nd Petit Verdot...1 Asrct...1 Introduction...2 Mterils nd Methods...7 Results...17 Discussion...87 Literture cited Chpter 2. Extent nd timing of fruit-zone lef nd lterl shoot removl lters yield components nd fruit composition in Cernet Suvignon grpes Asrct Introduction Mterils nd Methods Results Discussion Literture cited v

6 Section Pge Chpter 3. Preliminry ssessment of n hourly mient temperture nd rdition-driven model of grpe erry temperture under vrile sky conditions of the estern US Asrct Introduction Mterils nd Methods Results Discussion Literture cited vi

7 List of tles Tle Pge Chpter 1 Tle 1. Monthly growing degree dy (GDD) nd rinfll sum, nd monthly men nd mximum tempertures for Jun, Jul, Aug, Sep, nd Oct in 2013 nd Tle 2. Lef removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) collected t three periods (EL 31, 33, nd 35), nd cne pruning weights collected during vine dormncy, in Cernet frnc, Tle 3. Lef removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) collected t three periods (EL 31, 33, nd 35), nd cne pruning weights collected during vine dormncy, in Petit Verdot, Tle 4. Simple liner reltionship (R 2 ) etween mient ir nd erry temperture s function of fruit-zone lef removl, cnopy side, nd time of dy in Cernet frnc nd Petit Verdot in 2013 nd Tle 5. Lef removl tretment effect on the simple liner reltionship (R 2 ) etween mient ir temperture nd Cernet frnc erry temperture over ll times of dy, nd the estimted minimum mient ir temperture (T) required for erries to rech, nd mount of time spent t, 30 nd 35 C on est cn west cnopy sides, in the post-verison periods of 2013 nd Tle 6. Averge mient temperture (T) nd photosyntheticlly ctive rdition (PAR) during direct-sun time periods when est nd west-side Cernet frnc erry temperture ws mnully mesured 30 nd 35 C in individul experimentl units in Tle 7. Lef removl tretment effect on the simple liner reltionship (R 2 ) etween mient ir temperture nd Petit Verdot erry temperture over ll dtes nd times of dy, nd the estimted minimum mient ir temperture (T) required for erries to rech, nd mount of time spent t, 30 nd 35 C on est cn west cnopy sides, during the post-verison periods in 2013 nd Tle 8. Averge mient temperture (T) nd photosyntheticlly ctive rdition (PAR) during direct-sun time periods when est nd west-side Petit Verdot erry temperture ws mnully mesured 30 nd 35 C in individul experimentl units in vii

8 Tle Pge Tle 9. Pre-loom nd post-fruit set lef removl effect on crop yield components, cluster compctness, crop lod, nd count nd sl shoot fruitfulness in Cernet frnc, Tle 10. Pre-loom nd post-fruit set lef removl effect on crop yield components, cluster compctness, crop lod, nd count nd sl shoot fruitfulness in Petit Verdot, Tle 11. The effect of pre-loom lef removl on the rtio of % reduction in yield components: % reduction in crop yield in Cernet frnc nd Petit Verdot in 2013 nd Tle 12. The effect of pre-loom lef removl on the rtio of % reduction in components of cluster weight in Cernet frnc nd Petit Verdot in 2013 nd Tle 13. Post-fruit set nd pre-loom lef removl effect on components of cluster compctness in 2013 nd Tle 14. Pre-loom nd post-fruit set lef removl effect on components of Cernet frnc erry weight on 9 Sep 2013, nd components of Petit Verdot erry weight on 19 Sep Tle 15. Post-fruit set nd pre-loom lef removl effect on primry fruit chemistry t hrvest in Cernet frnc nd Petit Verdot, Tle 16. Pre-loom nd post-fruit set lef removl effect on totl erry nthocynins (TBA) nd totl erry phenolics (TBP) in Cernet frnc nd Petit Verdot from Tle 17. The simple liner regression reltionship etween hrvest erry weight nd totl erry phenolics (TBP) nd totl erry nthocynins (TBA) in Cernet frnc nd Petit Verdot from Tle 18. Post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids in Tle 19. Pre-loom nd post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids in Tle 20. Pre-loom nd post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids from the EAST cnopy side in viii

9 Tle Pge Tle 21. Pre-loom nd post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids from the WEST cnopy side in Tle 22. Post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids over Tle 23. Post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids in Tle 24. Pre-loom nd post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids in Tle 25. Pre-loom nd post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids from the EAST cnopy side in Tle 26. Pre-loom nd post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids from the WEST cnopy side in Tle 27. Post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids over Tle 28. Tringle difference test of Cernet frnc nd Petit Verdot wines from the 2012 vintge...85 Tle 29. Consumer preference of Cernet frnc nd Petit Verdot wines from the 2013 vintge...86 Chpter 2 Tle 1. Method used to estimte hourly percent increse in PFS-6 erry temperture compred to the verge of PB-4/8 erry temperture in 2014 nd Tle 2. Sesonl nd monthly growing degree dy (GDD) nd rinfll ccumultion, nd men nd mximum monthly tempertures for Jul, Aug, Sep, nd Oct t the Alson H. Smith, Jr. AREC in Winchester, VA, Tle 3. Pre-loom nd post-fruit set lef/lterl removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) mesured t verison, nd dormnt cne pruning weights, Tle 4. Logged nd mnul-mesured coefficients of determintion for liner reltionships etween mient ir nd erry temperture s function of fruit-zone lef removl, cnopy side, nd time of dy ix

10 Tle Pge Tle 5. Pre-loom nd post-fruit set lef removl effects on crop yield components, crop lod, nd count nd sl shoot fruitfulness from Tle 6. Pre-loom lef removl effect on components of cluster compctness from Tle 7. Botrytis unch rot (BBR) nd unch stem necrosis (BSN) incidence s ffected y pre-loom nd post-fruit set lef removl in Tle 8. Pre-loom nd post-fruit set lef removl effects on juice solule solids concentrtion (SSC), ph, nd titrtle cidity (TA) in Tle 9. Pre-loom nd post-fruit set lef removl effects on totl erry nthocynins (TBA) nd phenolics (TBP) on the est nd west cnopy sides in Chpter 3 Tle 1. Pre-loom nd post-fruit set lef/lterl removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) mesured t verison in Cernet Suvignon over Tle 2. The reltionship, difference, nd stndrd devition etween predicted nd mesured erry temperture in pre-loom lef removl plots Tle 3. The reltionship, difference, nd stndrd devition etween predicted erry temperture in the pre-loom lef removl plots nd mesured erry temperture in post-fruit set lef removl plots x

11 List of figures Figure Pge Chpter 1 Figure 1. Hourly verge photosyntheticlly ctive rdition (PAR) nd ir temperture, monthly verge PAR nd ir temperture during nd , nd percent times spent t 25% increments of the sesonl (1 Jun-8 Oct) nd post-verison (15 Aug-9/19 Oct 2013/2014) mient PAR rnges during nd periods in 2013 nd Figure 2. The liner ir-erry temperture reltionship for est-exposed erries in the AM, nd for est- nd west-exposed erries t NOON in Cernet frnc nd Petit Verdot in Figure 3. Averge mient photosyntheticlly ctive rdition (PAR) nd temperture on erry temperture mesurement dtes in 2013 nd Figure Cernet frnc erry temperture on EAST nd WEST cnopy sides in AM, NOON, nd PM s ffected y lef removl...31 Figure Cernet frnc erry temperture on EAST nd WEST cnopy sides in AM, NOON, nd PM s ffected y lef removl...32 Figure Petit Verdot erry temperture on EAST nd WEST cnopy sides in AM, NOON, nd PM s ffected y lef removl...38 Figure Petit Verdot erry temperture on EAST nd WEST cnopy sides in AM, NOON, nd PM s ffected y lef removl...39 Figure 8. The effect of pre-loom lef removl implementtion in two consecutive yers on the percent reduction of crop yield components when compred to no lef removl in Cernet frnc nd Petit Verdot...48 Figure 9. Pre-loom nd post-fruit set lef removl effect on Cernet frnc erry weight over the course of 2013 nd Figure 10. Pre-loom nd post-fruit set lef removl effect on Petit Verdot erry weight over the course of 2013 nd Figure 11. Post-fruit set lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Cernet frnc grpe erries over the course of xi

12 Figure Pge Figure 12. Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Cernet frnc grpe erries over the course of Figure 13. Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Cernet frnc grpe erries collected from the EAST cnopy side over the course of Figure 14. Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Cernet frnc grpe erries collected from the WEST cnopy side over the course of Figure 15. Post-fruit set lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Petit Verdot grpe erries over the course of Figure 16. Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Petit Verdot grpe erries over the course of Figure 17. Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Petit Verdot grpe erries collected from the EAST cnopy side over the course of Figure 18. Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide, zexnthin, lutein, nd β-crotene in Petit Verdot grpe erries collected from the WEST cnopy side over the course of Figure 19. Cernet frnc nd Petit Verdot grpe zexnthin: lutein rtios t different phenologicl stges s ffected y no lef removl, nd post-fruit set removl of leves to medium nd high extents...83 Figure 20. The chnge in Cernet frnc nd Petit Verdot grpe crotenoids from pre-verison to hrvest s ffected y no lef removl, nd post-fruit set lef removl to medium nd high extents from Chpter 2 Figure 1. Chronology of pre-loom lef removl implementtion nd reimplementtion in experimentl lock 1 over the course of the entire study, Figure 2. Percent of time spent t 25% increments of photosyntheticlly ctive rdition (PAR) during the post-verison period from in 2014 nd 2015, from in 2014 nd 2015, nd from in 2014 nd xii

13 Figure Pge Figure 3. Diurnl ptterns of mient ir temperture, mient nd fruit-zone PAR, nd erry temperture s ffected y pre-loom removl of eight, four, nd no fruit-zone leves Figure 4. Logged mient PAR nd mnully mesured erry temperture in the AM, NOON, nd PM on severl dtes over the growing sesons Figure 5. Logged mient PAR nd mnully mesured erry temperture in the AM, NOON, nd PM on severl dtes over the 2014 nd 2015 growing sesons Figure 6. Diurnl mient PAR nd percent PAR trnsmission to the fruit-zone over , nd pre-loom lef removl tretment effect on logged erry temperture hours ove 30 nd 35 C in the post-verison period of 2013, 2014, nd Figure 7. The effect of implementtion of pre-loom removl of eight nd four leves over , nd over 2014 nd 2015, on percent reduction in erry numer per cluster, cluster weight, nd crop yield, when compred to no lef removl Figure 8. Pre-loom nd post-fruit set lef removl effect on erry weight over the course of the 2013, 2014, nd 2015 sesons Figure 9. The liner reltionship etween hrvest erry weight nd totl erry nthocynins nd phenolics over , nd s ffected y pre-loom lef removl in 2013, 2014, nd Figure 10. The liner reltionship etween hrvest erry weight nd totl erry nthocynins nd phenolics over , nd s ffected y post-fruit set lef removl of six nd no leves in 2014 nd Chpter 3 Figure 1. The liner reltionship etween mient ir temperture nd erry temperture over Figure 2. The liner reltionship etween mient ir temperture nd erry temperture in ech of the three yers of dt collection, Figure 3. The liner reltionship etween mient ir temperture nd est nd west cnopy side erry temperture in ech of the three yers of dt collection, xiii

14 Figure Pge Figure 4. The liner reltionship etween mient ir temperture nd erry temperture, ccounting for pre-loom lef removl tretments over Figure 5. The liner reltionship etween mient ir temperture nd erry temperture in ech pre-loom lef removl tretment over Figure 6. The liner reltionship etween mient ir temperture nd erry temperture in ech pre-loom lef removl tretment, nd in ech of the three yers of dt collection, Figure 7. The liner reltionship etween mient ir temperture nd est nd west cnopy side erry temperture in ech of the lef removl tretment plots over Figure 8. The liner reltionship etween mient photosyntheticlly ctive rdition (PAR) nd erry temperture over Figure 9. Diurnl mient PAR nd percent PAR trnsmission to the fruit-zone over , nd pre-loom lef removl tretment effect on logged erry temperture hours ove 30 nd 35 C in the post-verison period of 2013, 2014, nd Figure 10. The liner reltionship etween mient PAR nd est nd west cnopy side erry temperture in ech of the two yers of dt collection, Figure 11. The liner reltionship etween mient PAR nd erry temperture, ccounting for pre-loom lef removl tretments over Figure 12. The liner reltionship etween mient PAR nd erry temperture in ech pre-loom lef removl tretment over Figure 13. The liner reltionship etween mient PAR nd erry temperture in ech pre-loom lef removl tretment, nd in ech of the two yers of dt collection, Figure 14. The liner reltionship etween mient PAR nd est nd west cnopy side erry temperture in ech of the lef removl tretment plots over Figure 15. The liner reltionship etween fruit-zone photosyntheticlly ctive rdition (PAR) nd erry temperture over Figure 16. The liner reltionship etween fruit-zone PAR nd erry temperture in ech of the three yers of dt collection, xiv

15 Figure Pge Figure 17. The liner reltionship etween fruit-zone PAR nd est nd west cnopy side erry temperture in ech of the three yers of dt collection, Figure 18. The liner reltionship etween fruit-zone PAR nd erry temperture, ccounting for pre-loom lef removl tretments over Figure 19. The liner reltionship etween fruit-zone PAR nd erry temperture in ech pre-loom lef removl tretment over Figure 20. The liner reltionship etween fruit-zone PAR nd erry temperture in ech pre-loom lef removl tretment, nd in ech of the three yers of dt collection, Figure 21. The liner reltionship etween fruit-zone PAR nd est nd west cnopy side erry temperture in ech of the lef removl tretment plots over Figure 22. The liner reltionship etween fruit-zone RH nd erry temperture over Figure 23. The liner reltionship etween fruit-zone RH nd mient ir temperture over Figure 24. The diurnl trend of fruit-zone RH, nd mient ir temperture over Figure 25. The diurnl reltionship (R 2 ) etween fruit-zone RH nd mient ir temperture, fruit-zone RH nd erry temperture, nd mient ir temperture nd erry temperture over Figure 26. The liner reltionship etween mient ir temperture nd erry temperture, ccounting for green nd red erry color, over Figure 27. The liner reltionship etween mient PAR nd erry temperture, ccounting for green nd red erry color, over Figure 28. The diurnl cnopy side-specific reltionship (R 2 ) etween mient ir nd erry temperture, s ffected y pre-loom removl of eight, four, nd no leves over Figure 29. The diurnl cnopy side-specific reltionship (R 2 ) etween mient PAR nd erry temperture, s ffected y pre-loom removl of eight, four, nd no leves over xv

16 Figure Pge Figure 30. The diurnl cnopy side-specific reltionship (R 2 ) etween fruit-zone PAR nd erry temperture, s ffected y pre-loom removl of eight, four, nd no leves over Figure 31. Diurnl ptterns of mient ir temperture, mient nd fruit-zone PAR, nd erry temperture s ffected y pre-loom removl of eight, four, nd no fruit-zone leves Figure 32. The men mient ir temperture t ech hour ngle dte tht mnul erry temperture ws mesured for purposes of comprison to predicted erry temperture Figure 33. The men mient PAR t ech hour ngle dte tht mnul erry temperture ws mesured for purposes of comprison to predicted erry temperture Figure 34. The liner reltionship etween predicted erry temperture nd mnully mesured erry temperture on the est cnopy side of pre-loom lef removl plots Figure 35. The liner reltionship etween predicted erry temperture nd mnully mesured erry temperture on the west cnopy side of pre-loom lef removl plots Figure 36. The liner reltionship etween predicted erry temperture on the est cnopy side of pre-loom lef removl plots nd mnully mesured erry temperture on the est cnopy side of post-fruit set lef removl plots Figure 37. The liner reltionship etween predicted erry temperture on the west cnopy side of pre-loom lef removl plots nd mnully mesured erry temperture on the west cnopy side of post-fruit set lef removl plots xvi

17 Extent nd timing of fruit-zone lef nd lterl shoot removl lters yield components, grpe phenolics, nd grpe crotenoids in Cernet frnc nd Petit Verdot. Astrct Bckground nd ims: The rtionle for fruit-zone lef removl in humid environment hs een driven more y disese mngement thn y documented chnges in fruit composition. Though common prctice for severl decdes, lef removl recommendtions remin generl nd re not vriety-, timing-, or mgnitude-specific. We evluted if the timing or mgnitude of fruit-zone lef nd lterl removl would lter fruit composition nd crop yield components of two regionlly populr red-fruited vrieties. Methods nd results: Two seprte experiments in djcent Cernet frnc nd Petit Verdot vineyrds evluted the effects of three post-fruit set lef/lterl shoot removl tretments [no removl (NO), removl from opposite the sl primry cluster nd the node directly ove (MED), nd removl from the node directly ove the distl primry cluster down to the cordon (HIGH)] nd one pre-loom (P-B) lef/lterl shoot removl tretment - removl from the six primry sl nodes. Post-fruit set lef removl hd mrginl, inconsistent effects on crop yield nd components. In Cernet frnc, P-B reduced crop yield y n verge of 50%, explined y reductions in cluster weight (39%), erry numer per cluster (33%), cluster numer (8%), nd erry weight (6%) compred to NO. In Petit Verdot, P-B reduced crop yield y n verge of 53%, explined y reductions in cluster weight (37%), cluster numer per vine (32%), erry weight (25%), nd erry numer per cluster (18%) compred to NO. Re-implementtion of P-B over two consecutive sesons cused further reduction in these yield components. Aggressive lef removl (HIGH nd P-B) tended to reduce solule solids in Petit Verdot ut not in Cernet frnc. HIGH tended to reduce titrtle cidity (TA) in oth vrieties, wheres P-B tended to reduce TA only in Cernet frnc. P-B more consistently incresed totl erry phenolics in Petit 1

18 Verdot thn in Cernet frnc, ut lef removl did not increse totl erry nthocynins. When compred to NO nd MED, HIGH nd P-B tended to increse crotenoid ccumultion to greter extent in the pre-verison period, nd increse crotenoid degrdtion to greter extent in the post-verison period; this ws prticulrly consistent for zexnthin. The color nd rom of wines from different lef removl tretments were distinguished from one nother, leit infrequently. Color intensity ws rted higher in wines mde with fruit from P-B plots compred to wines mde with fruit from MED plots. Conclusions: Pre-loom lef removl reduced crop yield, nd differentilly ffected crop yield components etween vrieties. Pre-loom lef removl did not ffect grpe nthocynins nd inconsistently improved totl grpe phenolics. Lef removl of severl sl leves tended to increse crotenoid synthesis nd degrdtion compred to removing fewer leves, nd this ws more consistent in Petit Verdot compred to Cernet frnc. Lef removl hs potentil to increse the color intensity of young red wines nd chnge rom nd stringency, ut preference of these ttriutes ws not determined. Significnce of the study: Removing fewer leves efore loom, or more leves immeditely fter fruit set my e est fruit-zone mngement strtegies to modestly improve fruit composition, sustin n economicl crop yield, nd crete fruit-zone environment ssocited with reduced disese incidence in humid environment. Introduction The reltionship of the cnopy to fruit qulity, whether with respect to sink-source reltions (Brvdo et l. 1985; Dokoozlin nd Kliewer 1995) or indirect/direct sunlight influence (Jcksoon nd Lomrd 1993; Smrt nd Roinson 1991, Meyers nd Vnden-Heuvel 2008), hs een extensively reserched. The impct of sunlight nd temperture on grpe composition hs 2

19 een of high interest, prticulrly regrding flvonoids (Bergqvist et l. 2001, Bond nd Sdrs 2014, Downey et l. 2006, Spyd et l. 2002, Trr et l. 2008). In generl, high temperture is ssocited with lower grpe nthocynins, lthough some light is importnt for their synthesis (Mori et l. 2007, Spyd et l. 2002, Trr et l. 2008). Consequently, fruit-zone lef removl hs ecome more conservtive, iming for 1 to 2 fruit-zone lef lyers (Wolf 2008), even though: (1) exposing fruit reduces fungl disese incidence (English et l. 1989, Wolf et l. 1986); (2) nthocynins re not lwys reduced in well-exposed grpes (Bogicevic et l. 2015, Chorti et l. 2010, Kotseridis et l. 2012); nd (3) n overwhelming numer of studies reported fruit exposure to e eneficil, nd shde detrimentl, to grpe nd wine qulity (Bledsoe et l. 1988, Cronneu 1985, Lee et l. 2007, Hunter et l. 1991, Jckson nd Lomrd 1993, Reynolds et l. 1996, Ryon et l. 2008, Smrt et l. 1985, Smith et l. 1988, Smrt nd Roinson 1991). Severl studies reported tht pre-loom lef removl incresed grpe nthocynins nd phenolics. Mechnisms of metolite improvements were: n incresed skin: pulp rtio (Poni et l. 2006), thicker erry skins (Plliotti et l. 2011, Pstore et l. 2013, Poni et l. 2008, Trdguil et l. 2010) or improved sink-source dynmics, s y ringing yields into lnce with the vegettive growth (Digo et l. 2012, Intrieri et l. 2008, Gtti et l. 2012, Poni et l. 2006). However, the generl positive effect of sun exposure on grpe phenolics could not e ruled out (Gtti et l. 2012, Pstore et l. 2013, Trdguil et l. 2012). Erly lef removl my, thus, offset the detrimentl temperture effects ssocited with fruit exposure vi the induction of other mechnisms tht re eneficil to grpe nthocynins, such s those mentioned ove. A recent keyword serch of grpe flvonoid in We of Science (Thomson Reuters) registered 1447 records wheres serching for grpe norisoprenoid nd grpe crotenoid 3

20 registered 65 nd 176, respectively. This is strong cse tht reltively more reserch hs een conducted on grpe flvonoids reltive to norisoprenoids nd crotenoids. Thus, given the sensory impct of flvonoids reltively more studies hve een conducted on chnging wine color, mouthfeel, nd stringency potentil compred to rom potentil. Further investigtion is required to improve our understnding of how cnopy mngement prctices impct precursors to wine rom compounds, such s crotenoids (Lee et l. 2007). Compred to other fruits, crotenoids re found in smll concentrtions in grpes (Goodwin 1980), ut, similr to other fruits, they re present in skins t levels two to three times greter thn found in the pulp (Rzungles et l. 1988). Crotenoid concentrtion is high erly in grpe development nd declines therefter, exhiiting shrp decline t verison (Rzungles et l. 1998). Since crotenoids re specificlly locted in the plstids, their degrdtion s grpe skins chnge from green to red corresponds to the disintegrtion of chloroplsts tht re not trnsformed into chromoplsts (Okomi et l. 1975). The proposed mechnism of crotenoid degrdtion (nd norisoprenoid formtion) in plnts hs three steps, involving oxidtive clevge, enzymtic trnsformtions, nd cid-ctlyzed conversions. More recently, crotenoid clevge dioxygenses (CCD) were found to e responsile for the production of pocrotenoids (i.e. C13-norisoprenoids) in plnts (Auldridge et l. 2006). Regrdless of the mechnism, crotenoid degrdtion results in C13-norisoprenoids, which hve powerful rom properties (Bumes et l. 2002, Mendes-Pinto 2009, Winterhlter nd Rousefff 2002). C13-norisoprenoids hve vrious sensory descriptors such s violet, woody, nd rsperry (β-ionone) nd cooked pple, florl, nd quince (β-dmscenone) (Winterhlter nd Rouseff 2002), nd cn increse fruity roms while simultneously muting herceous tones (Escudero et l. 2007, Pineu et l. 2007). C13-norisoprenoids hve low olfctory perception thresholds (Mendes-Pinto 2009), nd 4

21 thus re perceived in young, mono-vrietl red wines (Ferreir et l. 2000, Pineu et l. 2007), nd re importnt odornts of Cernet Suvignon, Cernet frnc, nd Merlot (Fn et l. 2010). Since light is primrily responsile for crotenoid iosynthesis nd regultion (Rzungles et l. 1998, Winterhlter nd Rouseff 2002), crotenoid contriution to wine rom cn e ffected y viticulture prctices tht chnge the fruit-zone light environment, thus ltering the grpe crotenoid profile (Bureu et l nd 1998, Mendes-Pinto 2009, Oliveir et l. 2004, Rzungles et l nd 1998). Crotenoids cn quntittively nd qulittively influence wine rom due to the synthesis nd degrdtion ptterns of crotenoid suclsses. Quntittively: Fruit exposure enhnces oth pre-verison ccumultion nd post-verison rekdown of crotenoids (Frin et l. 2010, Rzungles et l. 1998), resulting in greter norisoprenoid concentrtions t hrvest (Bureu et l nd 2000). Qulittively: Xnthophylls exhiit different ptterns thn crotenes throughout erry mturtion (Rzungles et l. 1996, Olivier et l. 2004, Crupi et l. 2010). Xnthophylls re suject to inter-conversion to/from epoxified, or de-epoxified, forms s consequence of light intensity (Ymmto 1979). De-epoxified xnthophylls, such s zexnthin, re chrcteristic of high light intensity (Bumes et l. 2002, Bureu et l. 1998). Thus, grpes differentilly exposed to light cn hve different proportions of xnthophylls (Bureu et l nd 1998) nd norisoprenoids, the ltter hving specific crotenoid precursors (Crupi et l. 2010). The comined effects of light nd temperture, which could not e seprted in field setting, were responsile for incresing pre-verison crotenoids in sun-exposed compred to shded erries (Bureu et l. 1998). Crotenoid concentrtions were greter in grpes from vineyrd with greter rdition nd from wrmer seson compred to from vineyrd with lower rdition nd cooler seson, respectively (Frin et l. 2010). Though erry temperture 5

22 my e n importnt determinnt of grpe crotenoids, temperture dt ws not elorted upon when 5-6 C difference in grpe temperture ws reported (Bureu et l. 1998; Rzungles et l. 1998). Only one (Kwsniewski et l. 2010) grpe crotenoid study hs chrcterized the fruit-zone environment using dvnced techniques, such s enhnced point qudrt nlysis (EPQA) (Myers nd Vnden-Heuvel 2008), even though fruit-zone lef lyer numer ws positively relted to C13-norisoprenoids ecuse it ws good proxy for the comined effects of rdition nd temperture, oth of which were positively relted to severl C13-norisoprenoids (Lee et l. 2007). Thus, more thorough chrcteriztion of fruit-zone rdition nd erry temperture is required to etter understnd their effects on crotenoids, similr to wht hs een done for nthocynins (Bergqvist et l. 2001; Spyd et l. 2002; Trr et l. 2008). Most studies evluting the effects of fruit exposure on crotenoids hve involved lef pulling or shding of clusters fter fruit set (Bureu et l nd 1998, Kwsniewski et l. 2010, Rzungles et l. 1998). It is, therefore, unknown if removing leves efore loom could further increse crotenoid synthesis efore verison, thus dding improvement in wine rom potentil to the long list of other enefits of pre-loom lef removl (Plliotti et l. 2011, Poni et l. 2008, Trdguil et l. 2010). Little is known out the effects of unconventionlly erly lef removl on yield nd fruit composition in Cernet frnc nd Petit Verdot, two vrieties widely plnted in the estern U.S. Thus, it ws sought to evlute if the timing nd/or mgnitude of fruit-zone lef removl would impct crop yield, fruit composition t hrvest, nd crotenoid synthesis nd degrdtion ptterns. It ws hypothesized tht pre-loom lef removl would reduce crop yield, nd tht the removl of reltively more sl leves t the pre-loom nd post-fruit set stges would 6

23 fvorly lter fruit composition nd wine qulity y incresing flvonoids nd the synthesis nd degrdtion of crotenoids. Mterils nd Methods Tretments nd experimentl design: Two seprte, completely rndomized designs, comprising of five-vine-pnel experimentl units, ech replicted six times, were set up in djcent locks in commercil vineyrd in Shenndoh County, Virgini. The vineyrd soil ws n Endv silt lom. Cernet frnc (clone 214) nd Petit Verdot (clone 1) vines, oth grfted onto 3309 rootstocks, were plnted in 2001 t 3.0-m (row) x 1.8-m (vine) spcing. Vines were trined to i-lterl cordons ~0.9 m ove ground, nd were plnted in rows running generlly northest/southwest. Creeping red fescue (Festuc rur L.) ws estlished under-trellis in the Cernet frnc vines in Three post-fruit set (modified EL stge 31, Dry nd Coome 2004) lef/lterl shoot removl tretments were evluted in 2012: no removl (NO), removl from opposite the sl primry cluster nd the node directly ove (MED), nd removl from the node directly ove the distl primry cluster down to the cordon (HIGH). An dditionl lef/lterl shoot removl tretment ws implemented in 2013 nd repeted in 2014: removl from the six primry sl leves t the modified EL stge 18/19 (P-B). Frequent re-visits to experimentl units were mde to mintin cnopy porosity in the lef removl tretments through hrvest. Min plot tretments were re-rndomized ech yer. However, in 2014, the P- B nd NO plots tht were used in 2013 were mintined seprtely from the min plots. Tretments in these seprte plots were re-implemented to evlute ny crryover effects on responses noted elow. Dt were collected over the 2012, 2013, nd 2014 growing sesons. Missing oservtions in individul vines or whole experimentl units were uncommon, ut occurred in 7

24 some cses, such s significnt wildlife depredtion in As elorted upon elow, mesurements were tken in ech experimentl unit on multiple dys throughout the growing seson (enhnced point qudrt nlysis, erry weight, nd erry temperture), t hrvest (cluster compctness, crop yield components, primry chemistry), during the dormnt seson (dormnt cne pruning weights), nd during the pre-loom stge (vine fruitfulness). Generl vine mngement: Cordons were spur-pruned ech winter. Shoot density ws not djusted in 2012, however it ws djusted to ~18-20 shoots per m of cordon in 2013, nd djusted to ~15-16 shoots per m of cordon in Shoots were trined verticlly upright with the id of ctch wires. Cnopies were shoot-hedged efore shoots extended more thn ~0.9 m ove the top ctch wire. Fertilizer ws not pplied during the study, nd no nutritionl deficiency symptoms were evident. Disese mngement ws stndrd for the region. Meteorology: Temperture, rinfll, nd photosyntheticlly ctive rdition (PAR) were logged on 1-min intervls in 2013 nd 2014 using HOBO dt logger (model H21-002, Onset Computer Corp., Bourne, MA) locted ~23 m from the experimentl site. Growing degree dys were generted using se temperture of 10 C. Sesonl dt re presented from 1 Jun through 28 Oct. A severe hil storm in lte Jun 2014 cused severe lef nd erry scrring, erry desicction nd scission, nd depressed crop yield in 2014 reltive to other yers. Enhnced point qudrt nlysis nd dormnt cne pruning weights: Point qudrt nlysis (PQA) dt were collected t EL stges 31, 33, nd 35 in 2012, nd t EL stges 27, 33, nd 35 in 2013 nd A thin metl rod ws inserted into the fruiting zone long the trnsverse xis of the cnopy, using tpe mesure to guide insertions, s descried in Smrt nd Roinson (1991); this ws repeted ~25 times (~ every 0.35 m) in ech experimentl unit. Fruit-zone photosynthetic photon flux density (PPFD) ws mesured etween hrs in one cordon 8

25 of ech vine using n AccuPAR ceptometer (Model PAR-80, Decgon Devices, Inc., Pullmn, WA). Mesurements were tken y inserting the ceptometer inside fruit-zones, prllel to, nd directly ove, the cordon, nd orienting the light interception side of the ceptometer in three different directions (45 est, verticl, 45 west); those redings were then verged. Between ech fruit-zone insertion of the ceptometer, the mximum mient PPFD ws recorded outside of the cnopy. This permitted PPFD to e expressed s percentge of mient rdition for use with enhnced point qudrt nlyses (EPQA version 1.6.2) (Meyers nd Vnden-Heuvel 2008). EPQA softwre ws used to generte lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA). Cne pruning weights were collected y vine using field scle during the dormnt periods of 2012, 2013, nd Cne pruning weights were dditionlly collected in re-implemented P-B nd NO plots in 2014 nd 2015 to evlute if pre-loom lef removl hd crryover effects on vine cpcity. Berry temperture: Berry temperture ws mesured on six dtes etween Jul nd Sep in 2013 nd on eight dtes etween Jul nd Oct in Berry temperture ws mesured in every experimentl unit t three different times of dy on ech collection dte: morning (~ hrs), round solr noon (~ hrs), nd lte fternoon (~ hrs). Berry temperture ws mesured y inserting mini hypodermic thermocouple (model HYP1/2, Omeg Eng., Stmford, CT) eneth the skins of erries nd recorded using hndheld digitl thermometer (model HH 25, Omeg Eng., Stmford, CT). In every experimentl unit, t ech time of dy, on ech collection dte, nd on oth est nd west cnopy sides, the temperture of three erries on clusters exterior fce, positioned t the top, center, nd ottom of clusters, were mesured on two clusters orne on opposite cordons of one vine for totl of six temperture mesurements. In 2013, the time spn ws recorded over temperture mesurement t ech time 9

26 of dy (AM, NOON, PM), nd in 2014, over temperture mesurement in ech experimentl unit; the gol ws to investigte the reltionship of erry temperture with mient temperture nd rdition (recorded with the HOBO wether sttion). Time spent 30 nd 35 C ws estimted using the minimum estimted mient ir temperture necessry to rech these erry tempertures in the post-verison period in 2013 nd This mient ir temperture ws derived from the simple liner reltionship of mient ir temperture nd erry temperture over ll dtes of mnul erry temperture mesurement, which ws developed for ech specific lef removl tretment. Crop yield components, cluster compctness, nd vine fruitfulness: Crop yield weight per vine ws mesured with field scle t hrvest on 27 Sep 2012, 3 Oct 2013, nd 14 Oct 2014 for Cernet frnc, nd 4 Oct 2012, 9 Oct Oct 2014 for Petit Verdot. Cluster weight ws determined from the quotient of crop yield weight nd cluster numer. Berry weight t hrvest ws determined from 120-erry smples in 2012, 2013 nd 150-erry smples in Berry numer per cluster ws determined from the quotient of verge cluster weight nd verge erry weight. Crop lod ws otined y dividing crop yield y pruning weight on per vine sis. Clusters not representtive of tretments were not used for clculting crop yield or verge cluster weight, ut were counted when summing cluster numer per vine. Pre-hrvest erry weight ws determined from 120- or 240- composite erry smples rndomly nd eqully collected from est nd west cnopy sides in ech experimentl unit on severl pre-hrvest dtes in 2013 nd Petit Verdot nd Cernet frnc erry tissue nlysis ws performed on 30 rndomly selected, thwed erries from frozen 120- composite erry smples collected from ech experimentl unit in Seeds were counted nd seed nd skin weights were mesured to the thousndth of grm. To evlute pre-loom lef removl impct over time, crop yield 10

27 components nd erry weight t hrvest in 2014 were collected in P-B nd NO plots originlly implemented in 2013, nd re-implemented in These re-implemented plots were not errysmpled with equl frequency s current-seson plots, explining why components of yield were greter in the re-implemented plots. For this reson, re-implemented plots were not sttisticlly compred to current-seson plots in Cluster compctness ws indexed y determining the rtio of totl erry numer to min rchis length in 2013 nd Ten clusters from ech experimentl unit tht were +/- 25% of the tretment s verge hrvested cluster weight were rndomly selected from hrvest ins to determine cluster compctness. Vine fruitfulness ws ssessed t the modified EL stge y dividing the numer of inflorescences y the numer of shoots on oth sl (cordonoriginting) nd count (one-yer old spur-originting) shoots. In 2014, vine fruitfulness ws mesured on ll vines in ech NO nd P-B experimentl unit from In 2015, vine fruitfulness ws ssessed on three vines in ech NO nd P-B experimentl unit tht hd een implemented in oth 2013 nd In spring 2015, the wek nd sprse shoot growth ws suspected to e due to cold injury; this my hve confounded ccurte vine fruitfulness ssessment. Primry fruit chemistry: Juice from 60-erry smples collected immeditely prior to hrvest from ech experimentl unit ws nlyzed for solule solids concentrtion ( Brix), ph, nd titrtle cidity (TA). Juice ws otined y hnd-pressing fresh erries in plstic g nd then centrifuging for five min t ~ 3500 rpm. Solule solids were mesured with digitl refrctometer (Pocket PAL-1, ATAGO USA, Inc., Bellevue, WA). Juice ph ws mesured, nd TA ws determined y titrtion to n endpoint of ph 8.2, using n 848 Titrino Plus uto-titrtor (Metrohm USA, Riverview, FL) nd 0.1 N NOH se. Primry fruit chemistry ws evluted 11

28 on ll current-seson tretments, s well s in the NO nd P-B plots originlly implemented in 2013, nd re-implemented in Estimted totl grpe phenolics nd nthocynins: Asorption spectroscopy ws used to estimte totl grpe phenolics nd nthocynins from rndomly collected erry smples. Composite erry smples of erries were collected eqully from oth cnopy sides in every experimentl unit prior to hrvest nd frozen t -20 to -80 C until tests commenced. Composite erry smples from est- nd west- cnopy sides were collected into seprte gs on 11 Aug, 10 Sep, nd 22 Oct in Berry smples were thwed, homogenized with Mgic Bullet (Homelnd Housewres LLC, Los Angeles, CA), nd hd 30 ml of M KCl uffer (ph 1.0) nd 0.4 M sodium cette uffer (ph 4.5) dded to seprte 1.0 g liquots of the erry homogente. The homogente-uffer smples were shken for 10 min nd then centrifuged for 5 min t ~ 3500 rpm. The superntnt ws pipetted into 10 mm pth length Hellm qurtz cuvette (Thermo Fisher Scientific Inc., Pittsurgh, PA), nd the sornce t 520 nd 700 nm ws mesured in duplicte with Genesys 8 ThermoSpectronic spectrophotometer (Cmridge, UK). The smple contining the erry homogente nd M KCl uffer (ph 1.0) ws further diluted (2 prts smple: 1 prt M KCl uffer rtio) nd its sornce ws red in duplicte t 280 nm. Crotenoid nlysis: Extrction of crotenoids ws done using methods descried in Wrolstd et l. (2005) nd Kwsniewski et l (2010), using the sme rndomly collected erry smples used for totl phenolics nd nthocynins nlysis over the period. In rief, thwed erry smples were homogenized for 2 min with Mgic Bullet. Ten ml of 50/50 methnol (MeOH)/tetrhydrofurn (THF) with 0.1% BHT solution ws dded to 10 g liquot of the erry homogente contined in Nlgene Teflon fluorinted ethylene propylene Ok Ridge Centrifuge 12

29 Tue (ThermoScientific, Rochester, NY), gitted y hnd, centrifuged t 5,000 rpm for 5 min (Allegr 25R centrifuge, Beckmn Coulter, Indinpolis, IN) then the superntnt collected; this step ws repeted y dding nother 10 ml of 50/50 MeOH/THF with 0.1% utylted hydroxyltoluene (BHT) solution to the precipitte. The superntnts were comined in 100ml seprtory funnel with 20 ml of petroleum ether with 0.2% BHT nd 10 ml of 20% queous NCl, nd gitted for ~1 min. After phse seprtion for ~ 5 min, the orgnic phse ws collected nd dried under vcuum for 45 min t 60 C nd 3,500 rpm using Centrivp Concentrtor (model , Lconko, Knss City, MO), fitted with cold trp (model , Lconko, Knss City, MO Lconko) nd diphrgm pump (model 2018B-01, Dry Fst, Skokie, IL), re-dissolved in 2 ml ethnol, nd syringe filtered through 0.45 µm polyvinylidene fluoride (PVDF) memrne Millex filter (Merck Millipore Ltd., Tullgreen, Crrigtwohill Co., Cork, IRL) into 2 ml, mer glss HPLC vil with PTFE lid nd silicone septum (Supelco Anlyticl, Bellefonte, PA). Sponifiction ws not used during crotenoid extrction due to the conditions of this process resulting in stxnthin oxidtion nd stereomuttion of lutein nd zextnhin (Goodwin 1976). UPLC- MS nlysis of crotenoids ws conducted using Wters UPLC BEH C-18 column (100 x 2.1 mm, 1.7 µm prticle size) (Wters, Milford, MA) held t 40 C on Wters H-Clss UPLC equipped with Diode rry detector (DAD) nd Xevo Q-Tof mss spectrometer. Moile phse A consisted of 0.1% formic cid, 5% cetonitrile nd 94.9% wter, moile phse B consisted of 0.1% formic cid in wter. Smples were held t 10 C with n injection volume of 2.5 μl. Initil grdient condition ws 50:50 (v/v) A nd B for 0 to 1 minutes with 2 min pre-run 1 to 3 min 20:80, 3 to 4.5 min 10:90, 4.5 to 7.5 min hold t 10:90, 7.5 to 8 min 0:100 nd held t 0:100 until 15 min. Flow ws set to 0.5ml/min. The DAD ws set to collect nm, with 2 13

30 nm resolution. The mximum wvelength oserved for ech crotenoid used for quntifiction. The MS ws set to collect in ESCi+ (switching etween ESI+ nd APcI+ modes) s poor ioniztion of lte eluting compounds occurred with ESI+ only. ESI+ prmeters were cone gs flow of 100 L/hour, desolvtion 800 L/hour, desolvtion temp 450 C, source 125 C, nd kv. The coron voltge for APCi ws set to 20 μa. Leucine Enkephlin ws used s lockmss compound. β-crotene, lutein, nd zexnthin were identified nd quntified with respect to uthentic stndrds nd other xnthophylls were identified nd quntified y comprison of DAD nd MS spectr s well s elution order to previous reports (Crupi et l. 2010). Winemking: Cernet frnc nd Petit Verdot wine lots were mde from equl weights of fruit etween ll tretments. Wine lots were mde in duplicte from composite of grpe clusters from ll (2012), or severl (2013) experimentl units of the sme lef removl tretment. Fruit ws processed using de-stemmer/crusher (Wottle Type 2; Wottle Mschinen & WeinPressenu, Austri) in 2012 nd Lgun 1R de-stemmer nd 650 model crusher (Prospero, Plesntville, NY) in Crusher wheels were set to rek erry skins s uniformly s possile without reking seeds. Tretment musts remined in the sme 55L, high density polyethylene cylindricl tnks with covers (Nlgene , Nlgene Nunc, Rochester, NY) from processing through fermenttion. Potssium metisulfite (PMBS) ws dded t rte of 40 ppm to ll lots efore post-crush cold sok period tht lsted ~ four dys t 4.0 C. Trtric cid ws dded t rte of 1.5 g/l to ll lots fter cold sok. Musts were inoculted with Llvin BM 4x4 yest (Lllemnd, Inc. Montrel, Cnd) t rte of ~ 0.24 g/l using stndrd protocol for yest re-hydrtion (Scott Ls) nd dding the yest re-hydrtion supplement GoFerm (Lllemnd, Inc. Montrel, Cnd) t rte of ~0.3 g/l. 14

31 Musts were mended with equl prts Fermid K t ~1/3 sugr depletion (35 g/lot in 2012; 28 g/lot in 2013). Grpe skin cps of ll lots were punched down two to three times dily, nd fermenttion tempertures were monitored dily with ul thermometer. Solule solids were mesured once dily with hydrometer. When solule solids 1.0 Brix, Clini-Test Regent Tlets (Byer AG, Leverkusen, Germny) were used to determine dryness. Wines were siphoned off the primry lees nd nlyzed for percent lcohol nd residul sugr y the Virgini Tech Enology Anlyticl Services L. Wines were stored in glss croys with equl hedspce t 4.0 C. After confirming the sence of sulfur-like odors nd differences etween duplicte lots, sme-tretment wines were comined into one glss croy nd susequently ottled in 750 ml glss ottles with synthetic cork closures (Nomcorc, Zeulon, NC). Bottled wines were held t 4.0 C until sensory nlysis. Wine sensory nlysis: Tringle difference tests were conducted on the 2012 Cernet frnc nd Petit Verdot wines on 16 nd 17 April 2013 t Virgini Tech Food Science Deprtment s sensory nlysis l using rndomized complete lock design. Pnelists were enrolled in the Wines nd Vines clss t Virgini Tech, where they lerned out sensory nlysis, ut hd not previously received forml sensory trining. Approx ml of wine ws served t room temperture in cler interntionl stndrds orgniztion (ISO) wine glsses. Ech pnelist ws given set of three cler wine glsses; two of the three glsses contined wine mde from prticulr tretment nd one glss contined wine mde from different tretment. Smple presenttion order ws rndomized ut rotted with equl frequency. In ech session, pnelists were sked to identify the one wine tht hd different rom, color, nd flvor. Pnelists clensed their pltes with unslted crckers nd filtered wter. There were eight totl sessions, ech with pnelists. Dt were collected with SIMS softwre (Berkeley Heights, NJ). 15

32 Consumer preference tests of Cernet frnc nd Petit Verdot wines were conducted on the 2013 vintge on 5 nd 6 My 2015, t Virgini Tech Food Science Deprtment s sensory nlysis l using rndomized complete lock design. Cernet frnc tretments included MED, HIGH, nd P-B, nd Petit Verdot tretments included NONE, MED, HIGH, nd P-B. Pnelists were weekly red wine consumers. A lnced complete lock design ws implemented, such tht ech prticipnt evluted wine smples from ll tretments. Smples were leled with rndom, three-digit codes, nd served mondiclly. Approx ml of ech wine ws served t room temperture in cler ISO wine glsses. Consumers clensed their pltes with unslted crckers nd filtered wter. Prticipnts nswered questions on 9-point hedonic scle (1 = dislike extremely, 9 = like extremely) (Perym nd Pilgrim 1957) for ppernce, red color, rom, overll flvor, stringency, mouthfeel, length of finish, nd overll impression. Additionlly, fruity rom nd flvor, vegettive rom nd flvor, nd intensity of red color, stringency, mouthfeel, nd length of finish were evluted on 5-point Just Aout Right (JAR) scle (1 = not nerly enough 3 = just out right; 5 = much too much). Pnelists performed side-y-side rnking of tretment wines t the end of the hedonic tests; the higher the verge rnking, the more preferred the wine. Dt were collected with SIMS softwre (Berkeley Heights, NJ). Sttistics: One-wy ANOVA ws used to evlute the fixed effects of lef removl tretment (α 0.05) using JMP Pro 11 (SAS, Cry NC). Significnt difference of lest squre mens (α 0.05) were determined with Tukey s HSD for most responses. Student s T-test ws used to determine seprtion of lest squre mens (α 0.05) for the following: first round of EPQA mesurements tken in 2013 nd 2014, pre-loom crry-over effects from 2013 to 2014, vine fruitfulness in 2015, nd cnopy side effect on totl phenolics nd nthocynins in Mnul 16

33 erry temperture mesurements were nlyzed seprtely y time of dy nd cnopy side on ech dte. Totl estimted grpe phenolics nd nthocynins t hrvest in 2014 were nlyzed using model contining lef removl tretment nd cnopy side s fixed effects; their interction ws lso evluted. Within ech dte, tretment effect on grpe crotenoids ws nlyzed using one-wy ANOVA, nd, in 2014, cnopy sides were nlyzed seprtely; men seprtion ws determined with Tukey s HSD (α 0.05). Within ech yer, model ws used to determine the fixed effects of, nd interction etween, tretment nd dte on grpe crotenoids; men seprtion ws determined with Tukey s HSD (α 0.05). Across yers, model ws used to determine the fixed effect of tretment, smple dte, nd yer, s well s their interctions, on grpe crotenoids; men seprtion ws determined with Tukey s HSD (α 0.05). One-wy ANOVA ws used to nlyze the difference in zexnthin:lutein rtios, nd degrdtion in crotenoids from pre-verison to hrvest; men seprtion etween lef removl tretments ws determined with Tukey s HSD (α 0.05). Discrimintion significnce in the tringle difference test ws determined using n lph one-tiled nlysis (α 0.05) using JMP Pro 11 (SAS, Cry, NC). Men seprtion of sensory ttriutes nd rnking in the consumer preference test ws determined using Duncn s Multiple Rnge Test (α 0.05) using SAS 9.3 (SAS, Cry, NC). Results Sesonl meteorology: In generl, it ws wrmer nd drier from 1 Jun through 28 Oct in 2013 compred to the sme period in 2014 (Tle 1). There ws 5% greter totl GDD ccumultion in 2013 (1604) when compred to 2014 (1530), minly due to greter GDD ccumultion in Jul nd Aug 2013 compred to these sme months in Men temperture in Jul 2013 ws 1.5 C greter thn Jul 2014 nd mximum temperture in Aug nd Oct 2013 ws 2.8 nd 7.2 C greter thn in Aug nd Oct 2014, respectively; men nd mximum tempertures in Jun nd Sep were similr etween yers. There ws 67% less totl rinfll in 2013 (153 mm) when compred 17

34 to 2014 (465 mm), minly due to greter rinfll in Jun, Jul, nd Oct 2014 compred to these sme months in Amient photosyntheticlly ctive rdition (PAR) nd ir temperture ptterns y hour nd month re further elorted upon in the following su-sections. Tle 1. Monthly growing degree dy (GDD) nd rinfll sum, nd monthly men nd mximum tempertures for Jun, Jul, Aug, Sep, nd Oct in 2013 nd Yer Metric Jun Jul Aug Sep Oct T men ( C) T mx ( C) GDD sum Rinfll sum (mm) T men ( C) T mx ( C) GDD Rinfll sum (mm) Growing degree dys clculted using se 10 C. Cernet frnc hrvested on 3-Oct-2013 nd 14-Oct-2014 nd Petit Verdot hrvested on 9-Oct-2013 nd 24-Oct Hourly PAR nd temperture: Considering solr noon over the course of the seson (1 Jun through 28 Oct) to e ~1330, verge mient PAR ws lwys greter during the morning compred to fternoon (Fig. 1 A, B). In 2013, mient PAR rnged µmol m -2 s -1 from , nd µmol m -2 s -1 from (Fig. 1 A). In 2014, verge PAR mient PAR rnged µmol m -2 s -1 from , nd µmol m -2 s -1 from (Fig. 1 B). In oth sesons, the stndrd devition of PAR ws greter from ( µmol m -2 s -1 ) compred to ( µmol m -2 s -1 ) (Fig. 1 A, B). This suggested greter vriility in sky conditions, or more drstic chnges in solr ngle, efore solr noon compred to fter solr noon. However, when expressed s percentge of mient PAR, stndrd devition ws lower from (52-55%) compred to (72-75% nd 75%) (Fig. 1 A, B). Hourly mient ir temperture ws generlly greter in 2013 thn in 2014, nd ws gretest t ( C) (Fig. 1 A, B). Air temperture ws greter when verged 18

35 over ( C) compred to ( C). Stndrd devition in ir temperture ws virtully the sme efore ( ) nd fter ( ) solr noon in oth sesons, ut ws greter over the course of 2013 (5.2) compred to 2014 (4.5). Monthly PAR nd temperture from nd : Monthly PAR ws greter in the morning ( ) compred to fternoon ( ) in oth sesons (Fig. 1 C, D). Amient PAR ws µmol m -2 s -1 greter in the morning compred to the fternoon over Jun- Oct in 2013 (Fig. 1 C). Amient PAR ws µmol m -2 s -1 greter thn in the morning compred to the fternoon over Jun-Oct in 2014 (Fig. 1 D). Excepting Oct, monthly stndrd devition of mient PAR ws lower in the morning compred to the fternoon. Monthly ir temperture ws greter in the fternoon compred to the morning in oth sesons (Fig. 1 C, D). Afternoon ir temperture ws C greter in the fternoon thn in the morning over Jun- Oct in 2013 (Fig. 1 C). Afternoon ir temperture ws C greter in the fternoon thn in the morning over Jun-Oct in 2014 (Fig. 1 D). Though morning ir temperture ws virtully the sme from Aug-Oct 2013 nd 2014, fternoon ir temperture from Aug-Oct 2013 ws greter thn in Excepting Oct, monthly stndrd devition of ir temperture ws typiclly greter in the morning compred to the fternoon. Time spent t ech 25% mient PAR rnge from nd : The difference in time spent t the 0-50% mient PAR rnge etween morning nd fternoon ws greter during the whole seson nd post-verison periods of 2014 compred to 2013 (Fig. 1 E, F). More time ws spent under cloudy conditions ( 25% PAR) in the fternoon compred to morning, nd more time ws spent under cloudy conditions during the 2014 compred to 2013 post-verison period. By contrst, more time ws spent under sunnier conditions (25-75% PAR rnges) in the morning compred to the fternoon in oth sesons. More time ws spent under 19

36 sunnier conditions in the 2013 compred to 2014 post-verison period, nd more time ws spent under sunny conditions in the morning compred to the fternoon during post-verison in Compred to other PAR rnges, there ws little difference in time spent t the % PAR rnge etween morning nd fternoon. 20

37 Amient PAR ( mol m -2 s -1 ) 3500 Amient PAR Amient temperture A Amient PAR Amient temperture B Amient ir temperture ( C) Hour C Hour D Amient PAR ( mol m -2 s -1 ) Amient PAR from Amient PAR from Amient temperture f rom Amient temperture f rom Amient PAR from Amient PAR from Amient temperture f rom Amient temperture f rom Amient ir temperture ( C) % time spent t PAR rnge preceding x-xis lel Jun Jul Aug Sep Oct Month 70 PAR f rom whole seson PAR f rom whole seson PAR f rom post-verison PAR f rom post-verison PAR ( mol m -2 s -1 ) E Jun Jul Aug Sep Oct Month PAR f rom whole seson PAR f rom whole seson PAR f rom post-verison PAR f rom post-verison PAR ( mol m -2 s -1 ) Fig. 1 Hourly verge mient photosyntheticlly ctive rdition (PAR) nd ir temperture (A, B), monthly verge mient PAR nd ir temperture during nd periods (C, D), nd percent time spent t 25% increments of the sesonl (1 Jun - 28 Oct) nd post-version (15 Aug - 9/19 Oct 2013/2014) mient PAR rnges during nd periods (E, F) in 2013 (A, C, E) nd 2014 (B, D, F). Dt logged every minute from 1 Jun - 28 Oct. F % time spent t PAR rnge preceding x-xis lel 21

38 Enhnced point qudrt nlysis nd dormnt cne pruning weights: Cernet frnc nd Petit Verdot. There ws no difference in fruit-zone lef lyer numer (LLN) or cluster exposure flux vilility (CEFA) efore leves were removed (EL 31) from Cernet frnc vines in 2012 (Tle 2). After leves were removed (EL 33), CEFA followed logicl increse s function of the degree of lef removl: HIGH > MED > NO; while LLN showed corresponding decrese. This reltive fruit-zone porosity ws mintined through verison (EL 35) in Similrly, when LLN nd CEFA were mesured t EL 27, results showed tht P-B resulted in more open fruit-zones erly in the 2013 nd 2014 sesons. Though LLN nd CEFA redings were not tken on MED nd HIGH pnels t EL 27, it is ssumed tht their vlues were similr to NO, given tht no leves hd een removed from those plots t tht stge. As phenology progressed in 2013 nd 2014, fruit-zone porosity ws ffected y lef removl tretment in logicl fshion: NO > MED > HIGH = P-B (LLN), nd P-B = HIGH > MED > NO (CEFA). Besides hving comprtively lower CEFA t EL 27 nd EL 35 in 2013, LLN nd CEFA vlues in Petit Verdot prlleled those of Cernet frnc (Tle 3). This fruit-zone dt vlidted tht the forthcoming response dt ws function of oth the extent nd timing of lef removl. Dormnt cne pruning weight ws not ffected y ny lef removl tretment in either vriety in ny yer (Tles 2 nd 3). In 2014, the re-implementtion of pre-loom lef removl in Cernet frnc plots originlly imposed in 2013 did not ffect cne pruning weight compred to no lef removl plots originlly implemented in 2013 (dt not shown). Similrly, pruning weight ws not ffected y re-implementtion of these tretments in Petit Verdot in 2014 (dt not shown). 22

39 Tle 2. Lef removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) collected t three periods (EL 31, 33, nd 35), nd cne pruning weights collected during vine dormncy, in Cernet frnc, EL 31 EL 33 EL 35 Pruning Tretment weight LLN CEFA LLN CEFA LLN CEFA (kg / m row) NO c c 0.32 MED HIGH c c Significnce ns ns < < < < ns 2013 EL 27 EL 33 EL 35 Pruning Tretment weight LLN CEFA LLN CEFA LLN CEFA (kg / m row) NO c c 0.38 MED n/ n/ HIGH n/ n/ 0.09 c c P-B c c Significnce < < < < < < ns 2014 EL 27 EL 33 EL 35 Pruning Tretment weight LLN CEFA LLN CEFA LLN CEFA (kg / m row) NO c c 0.32 MED n/ n/ HIGH n/ n/ 0.03 c c P-B c c Significnce < < < < < < ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Student s T-test (only for EL 27 in 2013 nd 2014) or Tukey s HSD. 23

40 Tle 3. Lef removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) collected t three periods (EL 31, 33, nd 35), nd cne pruning weights collected during vine dormncy, in Petit Verdot, EL 31 EL 33 EL 35 Pruning Tretment weight LLN CEFA LLN CEFA LLN CEFA (kg / m row) NO c c 0.33 MED HIGH c c Significnce ns ns < < < < ns 2013 EL 27 EL 33 EL 35 Pruning Tretment weight LLN CEFA LLN CEFA LLN CEFA (kg / m row) NO c c 0.35 MED n/ n/ HIGH n/ n/ 0.07 c c P-B c c Significnce < < < < < < ns 2014 EL 27 EL 33 EL 35 Pruning Tretment weight LLN CEFA LLN CEFA LLN CEFA (kg / m row) NO c c 0.29 MED n/ n/ HIGH n/ n/ 0.01 c c P-B c c Significnce < < < < < < ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Student s T-test (only for EL 27 in 2013 nd 2014) or Tukey s HSD. 24

41 Berry temperture, nd mient temperture nd rdition on erry temperture mesurement dtes: Cernet frnc nd Petit Verdot. Berry temperture ws lrgely driven y mient ir temperture, ut ws confounded y fruit-zone lef removl nd cnopy side (Tle 4). The liner ir-erry temperture reltionship ws similr etween lef removl tretments t times of dy when the sun ws not directed on the fruit-zone. Due to the generl north-south orienttion of the vineyrd rows, this ws in the fternoon (PM) for the est cnopy side nd in the morning (AM) for the west cnopy side. When the sun ws cst on the est cnopy side in the AM, lef removl tended to reduce the ility to predict erry temperture with ir temperture when compred to no lef removl; this ws prticulrly true for HIGH in The ir-erry temperture reltionship ws most similr etween lef removl tretments round solr noon (NOON), regrdless of cnopy side. There ws greter difference in the ility to predict erry temperture with ir temperture (R 2 ) etween NO nd HIGH/P-B when the sun ws cst on the west cnopy side in the PM in 2013 compred to It ws likely tht erry temperture ws mesured on sunnier fternoons in 2013 compred to 2014, which follows the generl sunnier/wrmer wether ptterns of 2013 compred to 2014 (see ove meteorology). 25

42 Tle 4. Simple liner reltionship (R 2 ) etween mient ir nd erry temperture s function of fruit-zone lef removl, cnopy side, nd time of dy in Cernet frnc nd Petit Verdot in 2013 nd EAST Cernet frnc WEST Tretment AM NOON PM AM NOON PM NO MED HIGH P-B EAST Petit Verdot WEST Tretment AM NOON PM AM NOON PM NO MED HIGH P-B EAST 2014 c Cernet frnc WEST Tretment AM NOON PM AM NOON PM NO MED HIGH P-B EAST Petit Verdot WEST Tretment AM NOON PM AM NOON PM NO MED HIGH P-B NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Dt collected on 15 Jul, 29 Jul, 12 Aug, 26 Aug, 10 Sep, nd 23 Sep in 2013; erry temperture verged y ech experimentl unit nd mient temperture verged y ech time of dy; AM = ~ , NOON = ; PM = c Dt collected on 8 Jul, 21 Jul, 5 Aug, 19 Aug, 27 Aug, 8 Sep, 23 Sep, nd 7 Oct in 2014; erry nd mient temperture verged y ech experimentl unit; AM = ~ , NOON = ; PM =

43 The reduced ir-est erry temperture reltionship in lef removl tretment plots ppered to e due to greter nd more frequent rdint heting of exposed compred to shded grpes in the AM of The frequency of rdint heting of exposed grpes ws demonstrted y the reltive greter mount of MED, P-B, nd (prticulrly) HIGH dt points ove the ir-erry temperture trend line (Fig. 2 A, B). The mgnitude of rdint heting of grpes ws demonstrted y the distnce of these dt points from the ir-erry temperture trend line; gin prticulrly in HIGH. The lck of tretment effect on the ir-erry temperture reltionship round NOON ws ttriuted to the sun eing locked y ove-hed cnopies in the verticlly-shoot positioned vineyrd. As result, dt points from oth cnopy sides nd ll lef removl tretments were close to the ir-erry temperture trend line (Fig. 2 C, D). 27

44 Est-side erry temperture ( C) NO (y = x) MED (y = x) HIGH (y = x) P-B (y = x) Liner regression y = x R 2 = A NO (y = x) MED (y = x) HIGH (y = x) P-B (y = x) Liner regression y = x R 2 = B Berry temperture ( C) NO (y = x) MED (y = x) HIGH (y = x) P-B (y = x) Liner regression y = x R 2 = C NO (y = x) MED (y = x) HIGH (y = x) P-B (y = x) Liner regression y = x R 2 = D Air temperture ( C) Air temperture ( C) Fig. 2 The liner ir-erry temperture reltionship for est-exposed erries in the AM (A,B), nd for est- nd west-exposed erries t NOON (C,D) in Cernet frnc (A,C) nd Petit Verdot (B,D) in NO, MED, nd HIGH = no lef removl, nd lef removl to medium nd high extents, respectively; P-B = pre-loom lef removl. Ech dt point represents six erry temperture mesurements; n = 45 in A nd B; n = 264 for C nd D. The similrity in the ir-west erry temperture reltionship etween lef removl nd no lef removl plots ppered to e due to more vrile cloud coverge in the PM periods on mnul erry temperture mesurement dtes in There ws greter cross-dte (in 2014) nd within-dte (in 2013 nd 2014) stndrd devition of PAR in the PM compred to AM, suggestive of greter vriility in sky conditions in the PM (Fig. 3 A, B). Cloudy skies would ttenute rdint heting of grpes nd thus limit differences in the ir-erry temperture reltionship etween lef removl tretments. Averge mient PAR ws in fct greter in the AM period thn the PM period in 2014 (Fig. 3 B), the yer tht ir-erry temperture trends were generted (Tle 4). There ws not s gret of difference in mient PAR etween AM nd 28

45 PM periods in 2013, however (Fig. 3 A). Amient temperture nd PAR were greter during ll time periods of erry temperture mesurement in 2013 compred to 2014, nd the within-dte stndrd devition of ir temperture tended to e greter in the AM compred to NOON or PM. Amient PAR ( mol m -2 s -1 ) (500) A (274) (346) (1.2) 5.3 (1.1) 5.0 (0.8) Amient PAR Amient temperture 615 (349) 382 (246) (0.9) (0.8) 502 (333) (0.8) B Amient ir temperture ( C) 0 AM NOON PM AM NOON PM 0 Fig. 3 Averge mient photosyntheticlly ctive rdition (PAR) nd temperture on erry temperture mesurement dtes in 2013 (A) nd 2014 (B). Ctegory mens (under rs) generted cross erry temperture mesurement dtes. The cross-dte stndrd devition (outside prentheses) ws verged cross dtes; the within-dte stndrd devition (in prentheses) ws verged within dte, then verged. Stndrd devition generted 1 Jun - 28 Oct. Mnul erry temperture mesurement in Cernet frnc. When verged cross ll tretments, times of dy, nd mesurement dtes, Cernet frnc erry temperture ws similr etween cnopy sides (Figs. 4 nd 5). In 2013, est side erry temperture ws 27.1 C (Fig. 4 A, B, C) nd west side erry temperture ws 27.1 C (Fig. 4 D, E, F). In 2014, est side erry temperture ws 25.6 C (Fig. 5 A, B, C) nd west side erry temperture ws 25.5 C (Fig. 5 D, E, F). Est nd west side erry temperture rnged from C cross ll tretments in 2013 (Fig. 4 A-F), nd rnged from C cross ll tretment in 2014 (Fig. 5 D, E, F). West side erry temperture ws greter in the PM compred to est side erry temperture in the AM (Figs. 4 nd 5 A, F). Greter differences in erry temperture were oserved etween 29

46 lef removl tretments when the sun ws ngled on the fruit-zone. In the AM, est side erry temperture rnged C cross tretments in 2013, nd rnged C cross tretments in 2014 (Fig. 4 nd 5 A). In the PM, west side erry temperture rnged C cross tretments in 2013, nd rnged C cross tretments in 2014 (Fig. 4 nd 5 F), Though mient PAR levels were greter t NOON compred to AM or PM, lef removl tretment never ffected erry temperture t NOON. Amient PAR levels were t lest one of the top two recorded in ech time of dy qudrnt when lef removl sttisticlly incresed erry temperture. In the AM on 29 Jul 2013, ll lef removl tretments incresed est side erry temperture (p > F = ) when compred to NO (Fig. 4 A). In the PM on 15 Jul 2013, only P-B incresed west-side erry temperture (p > F = ) when compred to NO (Fig. 4 F). In 2014, HIGH incresed est-side erry temperture when compred to NO in the AM on 8 Jul (p > F = ), 27 Aug (p > F = ), nd 23 Sep (p > F = ) (Fig. 4 A), nd in PM of 5 Aug (p > F = ) (Fig. 5 C). Additionlly, HIGH incresed west-side erry temperture (p > F = ) when compred to NO in PM of 5 Aug 2014 (Fig. 5 F). 30

47 Amient ir temperture ( C) 35 A B NO erry T MED erry T HIGH erry T P-B erry T Air T PAR C Amient PAR ( mol m -2 s -1 ) Amient ir temperture ( C) 35 D E NO erry T MED erry T HIGH erry T P-B erry T Air T PAR F Amient PAR ( mol m -2 s -1 ) 15 Jul 29 Jul 12 Aug 26 Aug 10 Sep 23 Sep 15 Jul 29 Jul 12 Aug 26 Aug 10 Sep 23 Sep 15 Jul 29 Jul 12 Aug 26 Aug 10 Sep 23 Sep Fig Cernet frnc erry temperture on EAST (A, B, nd C) nd WEST (D, E, nd F) cnopy sides in AM (A nd D), NOON (B nd E), nd PM (C nd F) s ffected y lef removl (NO = no removl; MED nd HIGH = post-fruit set removl to medium nd high extents; P-B = preloom removl). Dt points re n verge of six mesurements in ech experimentl unit; n = 6. Tretment mens within dte not shring letter re different ( using Tukey's HSD. Error rs re +/- stndrd error. 31

48 Amient ir temperture ( C) 40 A B NO erry T MED erry T HIGH erry T P-B erry T Air T PAR C Amient PAR ( mol m -2 s -1 ) Amient ir temperture ( C) 40 D E NO erry T MED erry T HIGH erry T P-B erry T Air T PAR F 8-Jul 21-Jul 5-Aug 19-Aug 27-Aug 23-Sep 7-Oct Amient PAR ( mol m -2 s -1 ) Jul 21-Jul 5-Aug 19-Aug 27-Aug 8-Sep 23-Sep 7-Oct 8-Jul 21-Jul 5-Aug 19-Aug 27-Aug 8-Sep 23-Sep 8-Jul 21-Jul 5-Aug 19-Aug 27-Aug 23-Sep 7-Oct Fig Cernet frnc erry temperture on EAST (A, B, nd C) nd WEST (D, E, nd F) cnopy sides in AM (A nd D), NOON (B nd E), nd PM (C nd F) s ffected y lef removl (NO = no removl; MED nd HIGH = post-fruit set removl to medium nd high extents; P-B = preloom lef removl). Dt points re n verge of six mesurements in ech experimentl; unit; n = 6. Tretment mens within dte not shring letter re different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. 32

49 Commonly cited criticl grpe temperture thresholds for nthocynin ccumultion (30 or 35 C) were oserved in Cernet frnc in 2013 nd 2014, when 108 nd 132 erry temperture mesurements (n = 6) were tken on ech cnopy side of ech experimentl unit, respectively. In 2013, erry temperture 30 C ws oserved 37 nd 39 times in HIGH, 33 nd 34 times in MED, 26 nd 33 times in NONE, nd 40 nd 37 times in P-B on the est nd west cnopy sides, respectively. Berry temperture 35 C ws oserved one time in HIGH on the est cnopy side in the PM, nd 11 times in HIGH, six times in MED, one time in NONE, nd four times in P-B on the west cnopy side in the PM in In 2014, erry temperture 30 C ws oserved 32 nd 35 times in HIGH, 19 nd 28 times in MED, 13 nd 28 times in NONE, nd 27 nd 35 times in P-B on the est nd west cnopy sides, respectively. Berry temperture 35 C ws oserved two times in HIGH nd NONE, nd three times in MED nd P-B only on the west cnopy side in the PM in (NOTE: dt from ove prgrph not shown). Cernet frnc erries spent more estimted time 30 C in 2013 compred to 2014, nd on the west compred to est cnopy side (Tle 5). Typiclly, more ggressive lef removl required lower mient ir temperture for erries to rech 30 nd 35 C. Consequently, erries spent more time ove these tempertures in lef removl plots compred to no lef removl plots. In 2013, erries on oth cnopy sides spent more time 30 C in HIGH plots compred to ll other tretment plots, including P-B. In 2014, est-exposed erries ttined gretest durtion 30 C in HIGH plots, nd west-exposed erries ttined gretest durtion 30 C in oth HIGH nd P-B plots. Berry temperture 35 C ws estimted to occur only in west-exposed erries, nd for no more thn totl of 1.5 hours in ny lef removl tretment. 33

50 Tle 5. Lef removl tretment effect on the simple liner reltionship (R 2 ) etween mient ir temperture nd Cernet frnc erry temperture over ll times of dy, nd the estimted minimum mient ir temperture (T) required for erries to rech, nd mount of time spent t, 30 nd 35 C on est cn west cnopy sides, in the postverison periods of 2013 nd EAST C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B WEST C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B EAST-2014 c 30 C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B WEST-2014 c 30 C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Dt collected on 15 Jul, 29 Jul, 12 Aug, 26 Aug, 10 Sep, nd 23 Sep in 2013; erry temperture verged y ech experimentl unit nd mient temperture verged y ech time of dy; AM = ~ , NOON = ; PM = c Dt collected on 8 Jul, 21 Jul, 5 Aug, 19 Aug, 27 Aug, 8 Sep, 23 Sep, nd 7 Oct in 2014; erry nd mient temperture verged y ech experimentl unit. 34

51 Reltively lower mient PAR ws required to het erry tempertures 30 C in lef removl plots compred to no lef removl plots during direct-sun time periods, especilly on the est side in the AM (Tle 6). This trend ws not evident during NOON mesurement periods (dt not shown). Reltively lower mient temperture nd higher PAR resulted in erry tempertures 30 C on the est-side in the AM, while reltively higher mient tempertures nd lower PAR resulted in erry tempertures 30 C on the west-side in the PM. Reltively lower mient PAR ws required to het west-side erries 35 C in HIGH nd P-B plots compred to NONE nd MED plots during PM periods. Tle 6. Averge mient temperture (T) nd photosyntheticlly ctive rdition (PAR) during direct-sun time periods when est nd west-side Cernet frnc erry temperture ws mnully mesured 30 nd 35 C in individul experimentl units in C Tretment Amient T ( C) EAST side in AM Amient PAR (µmol m -2 s -1 ) Amient T ( C) WEST side in PM Amient PAR (µmol m -2 s -1 ) NO MED HIGH P-B C c EAST side in AM WEST side in PM Tretment Amient T Amient PAR Amient T Amient PAR ( C) (µmol m -2 s -1 )) ( C) (µmol m -2 s -1 ) NO n/ n/ MED n/ n/ HIGH n/ n/ P-B n/ n/ NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Amient conditions n verge of 2 nd 18 (NONE). 4 nd 20 (MED), 7 nd 21 (HIGH), nd 5 nd 20 (P-B) mesurements on est nd west cnopy sides respectively c No tempertures 35 C were mesured on the est side in the AM; mient conditions n verge of 2 (NONE). 3 (MED). 2 (HIGH), nd 3 (P-B) mesurements on the west cnopy side. 35

52 Mnul erry temperture mesurement in Petit Verdot. When verged cross ll tretments, times of dy, nd mesurement dtes, Petit Verdot erry temperture ws similr etween cnopy sides (Figs. 6 nd 7). In 2013, est side erry temperture ws 26.9 C (Fig. 6 A, B, C) nd west side erry temperture ws 27.0 C (Fig. 6 D, E, F). In 2014, est nd west side erry temperture ws 25.1 C (Fig. 7 A-F). Est nd west side erry temperture rnged from C cross ll tretments in 2013 (Fig. 6 A-F). Est nd west side erry temperture rnged from C cross ll tretments in 2014 (Fig. 7 A-F). West side erry temperture ws greter in the PM compred to est-side erry temperture in the AM. Greter differences in erry temperture were oserved etween lef removl tretments when the sun ws ngled on the fruit-zone. In the AM, est side erry temperture rnged C cross tretments in 2013, nd rnged C cross tretments in 2014 (Fig. 6 nd 7 A). In the PM, west side erry temperture rnged C cross tretments in 2013, nd rnged C cross tretments in 2014 (Fig. 6 nd 7 F). Though mient PAR levels were greter t NOON compred to AM or PM, lef removl tretment never ffected erry temperture t NOON. Lef removl sttisticlly incresed Petit Verdot erry temperture when PAR levels were reltively high, s on the following dtes. In the PM of 15 Jul 2013, HIGH nd P-B incresed est side erry temperture (p > F = ), nd ll lef removl tretments incresed west side erry temperture (p > F = ), when compred to NO (Fig. 6 C, F). In the PM of 26 Aug 2013, HIGH incresed est side erry temperture compred to NO (p> F = ), nd oth HIGH nd P-B incresed west side erry temperture (p > F = ), when compred to NO nd MED (Fig. 6 C, F). In the PM of 23 Sep 2013, HIGH incresed west side erry temperture (p > F = ) compred to NO nd MED (Fig. 6 F). In the AM on 23 Sep 2014, oth HIGH nd P-B hd greter est side 36

53 erry temperture (p > F = ) when compred to NO nd MED (Fig. 7 A), nd, in the PM on 5 Aug 2014, HIGH hd greter est side erry temperture (p > F =0.0082) when compred to NO nd MED (Fig. 7 C). 37

54 Amient ir temperture ( C) 35 A B NO erry T MED erry T HIGH erry T P-B erry T Air T PAR C Amient PAR ( mol m -2 s -1 ) Amient ir temperture ( C) D E NO erry T MED erry T HIGH erry T P-B erry T Air T PAR F Amient PAR ( mol m -2 s -1 ) 15 Jul 29 Jul 12 Aug 26 Aug 10 Sep 23 Sep 15 Jul 29 Jul 12 Aug 26 Aug 10 Sep 23 Sep 15 Jul 29 Jul 12 Aug 26 Aug 10 Sep 23 Sep Fig Petit Verdot erry temperture on EAST (A, B, nd C) nd WEST (D, E, nd F) cnopy sides in AM (A nd D), NOON (B nd E), nd PM (C nd F) s ffected y lef removl (NO = no removl; MED nd HIGH = post-fruit set removl to medium nd high extents; P-B = pre-loom lef removl). Dt points re n verge of six mesurements in ech experimentl unit; n = 6. Tretment mens within dte not shring letter re different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. 38

55 Amient ir temperture ( C) 40 A B NO erry T MED erry T HIGH erry T P-B erry T Air T PAR C Amient PAR ( mol m-2s -1 ) Amient ir temperture ( C) 40 D E NO erry T MED erry T HGIH erry T P-B erry T Air T PAR F Amient PAR ( mol m -2 s -1 ) Jul 21-Jul 5-Aug 19-Aug 27-Aug 8-Sep 23-Sep 7-Oct 8-Jul 21-Jul 5-Aug 19-Aug 27-Aug 8-Sep 23-Sep 8-Jul 21-Jul 5-Aug 19-Aug 27-Aug 23-Sep 7-Oct Fig Petit Verdot erry temperture on EAST (A, B, nd C) nd WEST (D, E, nd F) cnopy sides in AM (A nd D), NOON (B nd E), nd PM (C nd F) s ffected y lef removl (NO = no removl; MED nd HIGH = post-fruit set removl to medium nd high extents; P-B = pre-loom removl). Dt points re n verge of six mesurements in ech experimentl unit; n = 6. Tretment mens within dte not shring letter re different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. 39

56 Commonly cited criticl grpe temperture thresholds for nthocynin ccumultion (30 or 35 C) were oserved in Petit Verdot in 2013 nd 2014, when 108 nd 132 erry temperture mesurements (n = 6) were tken on ech cnopy side of ech experimentl unit, respectively. In 2013, erry temperture 30 C ws oserved 33 nd 34 times in HIGH, 30 nd 29 times in MED, 24 nd 27 times in NONE, nd 33 nd 33 times in P-B on the est nd west cnopy sides, respectively. Berry temperture 35 C ws oserved one time in HIGH on the west cnopy side t NOON, nd 12 times in HIGH, seven times in MED, three times in NONE, nd eight times in P-B on the west cnopy side in the PM in In 2014, erry temperture 30 C ws oserved 17 nd 28 times in HIGH, 9 nd 24 times in MED, 6 nd 22 times in NONE, nd 14 nd 26 times in P-B on the est nd west cnopy sides, respectively. Berry temperture 35 C ws oserved two times in HIGH only on the west cnopy side in the PM in 2014 (NOTE: dt from ove prgrph not shown). Petit Verdot erries spent more estimted time 30 C in 2013 compred to 2014, nd on the west compred to est cnopy side (Tle 7). As leves were removed more ggressively, lower mient ir temperture ws required for erries to rech 30 nd 35 C. Consequently, erries spent more time ove these tempertures in lef removl plots compred to no lef removl plots. Besides est-exposed erries in 2014, in which cse there ws little difference in time tht erries spent 30 C etween tretments, the time tht erries spent 30 C ws in the following order of hierrchy: HIGH > P-B > MED > NO. Berry temperture 35 C ws estimted to rrely occurred in Petit Verdot. 40

57 Tle 7. Lef removl tretment effect on the simple liner reltionship (R 2 ) etween mient ir temperture nd Petit Verdot erry temperture over ll dtes nd times of dy, nd the estimted minimum mient ir temperture (T) required for erries to rech, nd mount of time spent t, 30 nd 35 C on est cn west cnopy sides, during the post-verison periods in 2013 nd EAST C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B WEST C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B EAST-2014 c 30 C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B WEST-2014 c 30 C 35 C Tretment R 2 Minimum ir T Time Minimum ir T Time ( C) (hrs.) ( C) (hrs.) NO MED HIGH P-B NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Dt collected on 15 Jul, 29 Jul, 12 Aug, 26 Aug, 10 Sep, nd 23 Sep in 2013; erry temperture verged y ech experimentl unit nd mient temperture verged y ech time of dy; AM = ~ , NOON = ; PM = c Dt collected on 8 Jul, 21 Jul, 5 Aug, 19 Aug, 27 Aug, 8 Sep, 23 Sep, nd 7 Oct in 2014; erry nd mient temperture verged y ech experimentl unit. 41

58 The only time tht Petit Verdot erry temperture ws mesured 30 C in NO plots ws on the west cnopy side in the PM (Tle 8). The only time tht erry temperture ws mesured 35 C ws when it ws mesured in HIGH plots on the west cnopy side in the PM. Reltively lower mient temperture nd higher PAR resulted in erry tempertures 30 C on the est-side in the AM, while reltively higher mient tempertures nd lower PAR resulted in erry tempertures 30 C on the west-side in the PM. Tle 8. Averge mient temperture (T) nd photosyntheticlly ctive rdition (PAR) during direct-sun time periods when est nd west-side Petit Verdot erry temperture ws mnully mesured 30 nd 35 C in individul experimentl units in C Tretment Amient T ( C) EAST side in AM Amient PAR (µmol m -2 s -1 ) Amient T ( C) WEST side in PM Amient PAR (µmol m -2 s -1 ) NO n/ n/ MED HIGH P-B C c EAST side in AM WEST side in PM Tretment Amient T Amient PAR Amient T Amient PAR ( C) (µmol m -2 s -1 ) ( C) (µmol m -2 s -1 ) NO n/ n/ n/ n/ MED n/ n/ n/ n/ HIGH n/ n/ P-B n/ n/ n/ n/ NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. No tempertures 30 C were mesured on the est side in the AM in NO plots; mient conditions n verge of 0 nd 18 (NONE). 1 nd 16 (MED), 3 nd 19 (HIGH), nd 2 nd 17 (P-B) mesurements on est nd west cnopy sides respectively c No tempertures 35 C were mesured on the est side in the AM nor on the west-side in the PM in NO, MED, nd P-B plots; mient conditions in HIGH n verge of 2 mesurements on the west cnopy side. 42

59 Crop yield components nd vine fruitfulness: Cernet frnc. Post-fruit set lef removl hd no effect on components of yield in Cernet frnc in 2012 (Tle 9). When verged over nd compred to NO, P-B reduced cluster numer per vine y 10% (significnt only in 2014), erry weight y 6%, erry numer per cluster y 33%, cluster weight y 39%, nd crop yield y 50%. Due to lower crop yield without concomitnt decrese in dormnt cne pruning weight (see Tle 3), P-B reduced crop lod y n verge ( ) of 37% when compred to NO. The only time tht post-fruit set lef removl tretment ffected crop yield components over ws when MED reduced erry numer per cluster y 20%, cluster weight y 23%, nd crop yield y 26% when compred to NO in When pre-loom lef removl tretments originlly implemented in 2013 were re-implemented in the sme plots in 2014 (P-B 13) nd compred to no lef removl (NO 13), P-B 13 reduced erry weight y 13%, erry numer per cluster y 42%, cluster weight y 50%, cluster numer per vine y 21%, nd crop yield y 61%. Count shoot fruitfulness ws reduced y 0.18 cluster/shoot in 2015 due to repeted pre-loom lef removl implementtion over (P-B 13), nd sl shoot fruitfulness ws reduced y 0.09 cluster/shoot due to pre-loom lef removl implementtion in the previous seson (P-B 14). 43

60 Tle 9. Pre-loom nd post-fruit set lef removl effect on crop yield components, cluster compctness, crop lod, nd count nd sl shoot fruitfulness in Cernet frnc, Tretment Crop yield (kg / vine) Cluster # / vine Cluster weight (g) Berry # /cluster Berry weight (g) Crop lod Fruitfulness (count / sl) 2012 NO n/ MED n/ HIGH n/ Significnce c ns ns ns ns ns ns n/ 2013 NO n/ MED 2.14 c c n/ HIGH n/ P-B 1.61 c c n/ Significnce c < ns < < ns n/ 2014 NO / 0.44 MED n/ HIGH n/ P-B / 0.36 Significnce c < < < ns Crryover effects on 2014 crop yield / 2015 vine fruitfulness NO ' / 0.26 PB- ' / 0.26 Significnce c < < < < < / ns NO '14 n/ n/ n/ n/ n/ n/ 1.67 / 0.30 P-B '14 n/ n/ n/ n/ n/ n/ 1.50 / 0.21 Significnce c n/ n/ n/ n/ n/ n/ ns / NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls; NO 13, P-B 13 = no nd pre-loom lef removl originlly implemented in 2013 re-implemented in 2014, respectively; NO 14 nd P-B 14 = no nd pre-loom lef removl originlly implemented in Cluster numer per shoot; count = one-yer old spur-originting shoot, sl = cordon-originting shoot. Fruitfulness ssessed in yer presented, ut effects ttriuted to previous seson s lef removl. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Student s T-test (only 2014 crryover effects nd 2015 vine fruitfulness) or Tukey s HSD. 44

61 Petit Verdot. Post-fruit set lef removl to the medium extent (MED) reduced crop yield compred to no lef removl (NO) in 2012 (Tle 10). When verged over nd compred to the verge of ll other tretments, P-B reduced cluster numer per vine y 29% (significnt only in 2014), erry weight y 25%, erry numer per cluster y 18% (significnt only in 2013), cluster weight y 37%, nd crop yield y 53%. Post-fruit set lef removl did not ffect ny component of yield in 2013, ut in 2014, HIGH reduced cluster numer per vine y 25% when compred NO. Pre-loom lef removl (P-B) reduced crop lod only when compred to MED (66%) nd NO (72%) in When pre-loom lef removl tretments originlly implemented in 2013 were re-implemented in the sme plots in 2014 (P-B 13) nd compred to no lef removl (NO 13), P-B 13 reduced erry weight y 24%, erry numer per cluster y 36%, cluster weight y 50%, cluster numer per vine y 35%, nd crop yield y 66%. Count shoot fruitfulness in 2014 ws incresed y P-B when compred to NO, leit only y 0.13 clusters/shoot. Count shoot fruitfulness ws reduced y 0.18 clusters/shoot in 2015 due to repeted pre-loom lef removl implementtion over (P-B 13), nd y 0.25 cluster/shoot due to pre-loom lef removl implementtion in the previous seson (P-B 14). 45

62 Tle 10. Pre-loom nd post-fruit set lef removl effect on crop yield components, cluster compctness, crop lod, nd count nd sl shoot fruitfulness in Petit Verdot, Tretment Crop yield (kg / vine) Cluster # / vine Cluster weight (g) Berry # /cluster Berry weight (g) Crop lod Fruitfulness (count / sl) 2012 NO n/ MED n/ HIGH n/ Significnce c ns ns ns ns ns n/ 2013 NO n/ MED n/ HIGH n/ P-B n/ Significnce c < ns < < n/ 2014 NO / 1.24 MED n/ HIGH c n/ P-B c c 2.51 / 1.24 Significnce c < ns < / ns Crryover effects on 2014 crop yield / 2015 vine fruitfulness NO ' / 0.95 PB- ' / 0.93 Significnce c < < < < / ns NO '14 n/ n/ n/ n/ n/ n/ 2.26 / 0.98 P-B '14 n/ n/ n/ n/ n/ n/ 2.01 / 0.87 Significnce c n/ n/ n/ n/ n/ n/ / ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls; NO 13, P-B 13 = no nd pre-loom lef removl originlly implemented in 2013 re-implemented in 2014, respectively; NO 14 nd P-B 14 = no nd pre-loom lef removl originlly implemented in Cluster numer per shoot; count = one-yer old spur-originting shoot, sl = cordon-originting shoot. Fruitfulness ssessed in yer presented, ut effects ttriuted to previous seson s lef removl. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Student s T-test (only 2014 crryover effects nd 2015 vine fruitfulness) or Tukey s HSD. 46

63 Pre-loom reduction in crop yield components over time: Cernet frnc nd Petit Verdot. When pre-loom lef removl yield components were expressed s percent reduction compred to no lef removl, cluster numer per vine, erry weight, erry numer per cluster, cluster weight, nd crop yield tended to e further reduced in the second yer of two-yer consecutive tretment implementtion (Fig. 8). Berry weight ws the yield component tht ws lest ffected y the second consecutive yer of pre-loom lef removl implementtion. In Cernet frnc (Fig. 8 A), the following yield components were further reduced in the second consecutive yer of pre-loom lef removl implementtion: cluster numer per vine (16%), erry weight (7%), erry numer per cluster (13%), cluster weight (16%), nd crop yield (16%). In Petit Verdot (Fig. 8 B), the following yield components were further reduced in the second consecutive yer of pre-loom lef removl implementtion: cluster numer per vine (18%), erry weight (5%), erry numer per cluster (19%), cluster weight (19%), nd crop yield (25%). 47

64 0 Reduction in crop yield component (%) A Cluster numer per vine Berry weight Berry numer per cluster Cluster weight Crop yield Reduction in crop yield component (%) B Cluster numer per vine Berry weight Berry numer per cluster Cluster weight Crop yield Fig. 8. The effect of pre-loom lef removl implementtion in two consecutive yers on the percent reduction of crop yield components when compred to no lef removl in Cernet frnc (A) nd Petit Verdot (B). Ech dt point is n verge of 30 vines, except erry weight, which is n verge over ech five-vine experimentl unit; n = 6. NOTE: dt not collected in these plots in

65 Pre-loom lef removl effect on the proportionl reduction in crop yield components: Cernet frnc nd Petit Verdot. The proportionl reduction in crop yield components, presented s rtios of crop yield components: crop yield, were studied to investigte which component of yield most explined the generl crop yield depression due to pre-loom lef removl (Tle 11). The proportionl reduction in cluster weight ws similr etween vrieties, nd greter thn the proportionl reduction in other yield components. Thus, reduction in cluster weight ws most responsile for crop yield reduction in oth vrieties. However, the proportionl reduction in erry numer per cluster ws reduced y n verge of 52% more so in Cernet frnc compred to Petit Verdot. By contrst, the proportionl reduction in cluster numer per vine nd erry weight were reduced y n verge of 88% nd 122% more so, respectively, in Petit Verdot compred to Cernet frnc. Thus, erry numer per cluster reduction ws reltively more responsile for yield reduction in Cernet frnc, wheres the reduction in cluster numer per vine nd erry weight were more responsile for yield reduction in Petit Verdot. Tle 11. The effect of pre-loom lef removl on the rtio of % reduction* in yield components: % reduction* in crop yield in Cernet frnc nd Petit Verdot in 2013 nd Yer Cluster # /vine: crop yield Cluster weight: crop yield Berry # / cluster: crop yield Berry weight: crop yield Cernet frnc Petit Verdot *% reduction clculted in comprison to no lef removl (NO). The mens tht pre-loom lef removl reduced cluster weight ws vriety-dependent (Tle 12). The following % reductions in rtios re ll verged over The rtio of erry weight: erry numer per cluster reduction ws 252% greter in Petit Verdot compred to Cernet frnc; the rtio opposite this (erry numer per cluster: erry weight) ws 237% greter 49

66 in Cernet frnc compred to Petit Verdot. Similrly, the rtio of erry weight: cluster weight reduction ws 130% greter in Petit Verdot compred to Cernet frnc, wheres the rtio of erry numer per cluster: cluster weight reduction ws 46% greter in Cernet frnc compred to Petit Verdot. This, gin, illustrted tht pre-loom lef removl hd greter impct on erry numer per cluster in Cernet frnc while nd erry weight ws more sustntilly ffected in Petit Verdot. Tle 12. The effect of pre-loom lef removl on the rtio of % reduction* in components of cluster weight in Cernet frnc nd Petit Verdot in 2013 nd Berry weight: Berry # / cluster: Berry weight: Berry # / cluster: Yer erry # / cluster erry weight cluster weight cluster weight Cernet frnc Petit Verdot *% reduction clculted in comprison to no lef removl (NO) Components of cluster compctness: Cernet frnc nd Petit Verdot. In 2013, cluster compctness ws not ffected y ny lef removl tretment in Cernet frnc (Tle 13). Preloom lef removl (P-B) reduced erry numer per cluster to the gretest extent (32%) when compred to NO. However, pre-loom lef removl resulted in numericlly shorter rchis lengths. The result ws numericl reduction in cluster compctness (18%) compred to ll other tretments. In 2014, P-B significntly reduced Cernet frnc cluster compctness when compred to NO (39%) nd MED (48%) due primrily to similr reduction in erry numer per cluster compred to NO (38%) nd MED (44%). In 2013, P-B reduced Petit Verdot cluster compctness compred to HIGH (20%) due to 15% reduction in erry numer per cluster nd miniml reduction in rchis length (7%). In 2014, ll lef removl tretments reduced cluster 50

67 compctness compred to NO (verge of 22%) due to numericlly longer rchis lengths (10%) nd similr erry numers per cluster. Tle 13. Post-fruit set nd pre-loom lef removl effect on components of cluster compctness in 2013 nd Tretment Berry # per cluster Cernet frnc Rchis length (cm) Cluster compctness c Berry # per cluster Petit Verdot Rchis Cluster length compctness (cm) NO MED 45 c HIGH P-B 38 d c Significnce c < ns ns < ns NO MED HIGH P-B Significnce c < ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Berry numer per length of min rchis; cluster shoulders not counted or mesured. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. Pre-hrvest erry weight: Cernet frnc. In 2013, lef removl inconsistently ffected Cernet frnc erry weight (Fig. 9 A). When compred to NO, P-B reduced erry weight y g on 14 Jul nd 19 Sep, nd oth P-B nd HIGH reduced erry weight on 26 Jul. Berry weight ws more consistently reduced y P-B in 2014 (Fig 9 B). In 2014, pre-loom lef removl (P-B) reduced erry weight y 10-12% during the 11 Aug through 14 Oct period when compred to ll other tretments. Additionlly, P-B reduced erry weight compred to MED on 17 Jul, nd compred to NO nd MED on 29 Jul. Berry weight increse ws 0.04 g less in P-B compred to other lef removl tretments over 29 Jul to 11 Aug. Berry weight ws lower on the west (Fig. 9 51

68 D) compred to est (Fig. 9 C) cnopy side in the post-verison of It ws of interest to evlute if there were cnopy side differences in erry weight due to hil dmge on the west cnopy side in Cnopy side differences in erry weight were gretest in HIGH ( g), moderte in MED, nd lest in NO. On the est cnopy side, P-B reduced erry weight compred to ll tretments on 11 Aug (0.11 g) nd 2 Sep (0.18 g), nd only when compred to MED nd HIGH on 22 Sep (0.17 g), nd MED on 14-Oct (0.18 g) (Fig. 9 C). On the west cnopy side, P-B reduced erry weight compred to ll tretments on 11 Aug (0.10 g) nd 14- Oct (0.16 g), nd compred to NO nd MED on 22-Sep (0.16 g) A B Berry weight (g) Berry weight (g) C 14-Jul 26-Jul 14-Aug 27-Aug 19-Sep 3-Oct 11-Aug 2-Sep 22-Sep 14-Oct NO MED HIGH P-B NO MED HIGH P-B D 17-Jul 29-Jul 11-Aug 2-Sep 22-Sep 14-Oct 11-Aug 2-Sep 22-Sep 14-Oct Fig. 9. Pre-loom (P-B), post-fruit set (MED = medium extent, HIGH = high extent), nd no (NO) lef removl effect on Cernet frnc erry weight over the course of 2013 (A) nd 2014 (B = verge, C = est cnopy side, D = west cnopy side). Ech dt point is n verge of 120 (prehrvest) or 150 (hrvest) erries evenly collected from ll vines in ech experimentl unit; n = 6. Error rs re +/- stndrd error. NO MED HIGH P-B NO MED HIGH P-B 52

69 Petit Verdot. In 2013, P-B reduced erry weight compred to ll other lef removl tretments on every smpled dte (Fig.10 A). The difference in erry weight incresed over time due to reltively slower increse in erry weight in P-B compred to other tretments. Berry weight increse ws g less in P-B compred to other tretments, depending on smple dte. In 2014, P-B reduced erry weight on every smpling dte compred to ll other lef removl tretments (Fig. 10 B). Similr to 2013, erry weight increse ws g less in P-B compred to other tretments, depending on the time period oserved. It ws of interest to evlute if there were cnopy side differences in erry weight due to hil dmge on the west cnopy side in Petit Verdot erry weight ws lower on the west (Fig. 10 D) compred to est (Fig. 10 C) cnopy side, leit to lesser mgnitude thn in Cernet frnc. Pre-loom lef removl (P-B) reduced erry weight on oth cnopy sides y rnge of g, compred to other tretments; however, tht reduction ws less on the west compred to est side cnopy side. 53

70 Berry weight (g) Berry weight (g) A C NO MED HIGH P-B 14-Jul 26-Jul 14-Aug 27-Aug 19-Sep 9-Oct 11-Aug 10-Sep 30-Sep 23-Oct NO MED HIGH P-B B D NO MED HIGH P-B 17-Jul 29-Jul 11-Aug 10-Sep 30-Sep 23-Oct 11-Aug 10-Sep 30-Sep 23-Oct Fig. 10. Pre-loom (P-B), post-fruit set (MED = medium extent, HIGH = high extent), nd no (NO) lef removl effect on Petit Verdot erry weight over the course of 2013 (A) nd 2014 (B = verge, C = est cnopy side, D = west cnopy side). Ech dt point is n verge of 120 (pre-hrvest) or 150 (hrvest) erries evenly collected from ll vines in ech experimentl unit; n = 6. Error rs re +/- stndrd error. NO MED HIGH P-B Components of erry weight: Cernet frnc nd Petit Verdot. Pre-loom lef removl reduced estimted Cernet frnc pulp weight y 8% when compred to HIGH (Tle 14). Pre-loom lef removl reduced Petit Verdot erry weight y 11% when compred to the verge of ll other tretments. While erries weighed less, the weight of erry skins ws not concomitntly reduced y P-B in Petit Verdot. Pre-loom lef removl reduced estimted Petit Verdot pulp weight y 16% when compred to the verge of MED nd HIGH. 54

71 Tle 14. Pre-loom nd post-fruit set lef removl effect on components of Cernet frnc erry weight on 9 Sep 2013, nd components of Petit Verdot erry weight on 19 Sep Estimted Skin: Estimted Berry Individul Skin Tretment Seed # Totl seed pulp Totl seed: erry erry skin: pulp weight seed weight weight / erry weight (g) weight (g) (g) (g) weight rtio weight weight (g) rtio rtio Cernet frnc NO MED HIGH P-B Significnce c ns ns ns ns ns ns ns ns Petit Verdot NO MED o.343 HIGH P-B Significnce c ns ns ns ns ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Clculted y sutrcting the sum of skin nd totl seed weight from erry weight. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 55

72 Primry fruit chemistry: Cernet frnc nd Petit Verdot. Lef removl tretment did not ffect juice solule solids in ny yer in Cernet frnc (Tle 15). In 2013 only, HIGH nd P-B incresed juice ph compred to NO. Juice titrtle cidity (TA) ws reduced y HIGH (17%) in 2012, y HIGH (15%) nd P-B (17%) in 2013, nd only y P-B (11%) in 2014, when compred to NO. Re-implementtion of pre-loom lef removl in 2014 in plots originlly implemented in 2013 reduced juice TA y 0.36 g/l when compred to re-implementtion of no lef removl plots in 2014 (dt not shown). In Petit Verdot, juice solule solids ws reduced y HIGH y 2% in 2013, nd y HIGH (5%) nd P-B (7%) in 2014, when compred to NO (Tle 12). Juice ph ws incresed y HIGH y 3% when compred to NO in 2012, nd reduced y HIGH (3%) nd P-B (6%) compred to MED in Juice TA ws reduced y HIGH y n verge of 15% when compred to NO nd MED in 2012, nd y 10% when compred to NO in Juice TA ws incresed y P-B y 20% when compred to the verge of ll other tretments in Re-implementtion of pre-loom lef removl in 2014 in the sme plots reduced solule solids y 0.6 Brix nd ph y 0.07 units, nd incresed TA y 1.53 g/l when compred to no lef removl plots re-implemented in 2014 (dt not shown). 56

73 Tle 15. Post-fruit set nd pre-loom lef removl effect on primry fruit chemistry t hrvest in Cernet frnc nd Petit Verdot, Tretment Solule solids ( Brix) Cernet frnc Petit Verdot Titrtle Solule ph cidity solids ph (g/l) ( Brix) Titrtle cidity (g/l) NO MED HIGH Significnce ns ns ns < NO c MED c HIGH P-B Significnce ns < ns NO MED HIGH P-B c Significnce ns ns < < < NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. Estimted totl grpe phenolics nd nthocynins: Cernet frnc nd Petit Verdot. Lef removl tretment did not ffect totl erry nthocynins in Cernet frnc (Tle 16). In 2014, P-B incresed totl erry phenolics y 20% compred to NO nd 21% compred to MED; HIGH incresed totl erry phenolics y 12% compred to MED. In 2014, Cernet frnc erry nthocynins were greter on the est compred to west cnopy side, nd vice-vers for erry phenolics. Totl erry phenolics were incresed y P-B y n verge of 18% in 2013 nd 28% in 2014 when compred to the verge of ll other tretments. 57

74 Tle 16. Pre-loom nd post-fruit set lef removl effect on totl erry nthocynins (TBA) nd totl erry phenolics (TBP) in Cernet frnc nd Petit Verdot from Cernet frnc Tretment nd Cnopy side TBA (mg/g erry) TBP (u/g erry) TBA (mg/g erry) TBP (u/g erry) TBA (mg/g erry) TBP (u/g erry) NO c MED c HIGH P-B n/ n/ EAST n/ n/ n/ n/ WEST n/ n/ n/ n/ Significnce c Lef removl (LR) ns ns ns ns ns < Cnopy side (CS) n/ n/ n/ n/ LR*CS n/ n/ n/ n/ ns ns Tretment nd Cnopy side TBA (mg/g erry) Petit Verdot TBP (u/g erry) TBA (mg/g erry) TBP (u/g erry) TBA (mg/g erry) TBP (u/g erry) NO MED HIGH P-B n/ n/ EAST n/ n/ n/ n/ WEST n/ n/ n/ n/ Significnce c Lef removl (LR) ns ns ns ns < Cnopy side (CS) n/ n/ n/ n/ ns ns LR*CS n/ n/ n/ n/ ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. u = sornce units when red t 280 nm wvelength c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 58

75 There ws significnt, negtive liner reltionship etween hrvest erry weight nd totl erry phenolics in two out of three yers in oth vrieties (Tle 17). There ws never significnt liner reltionship etween hrvest erry weight nd totl erry nthocynins in ny yer or vriety. Tle 17. The simple liner regression reltionship etween hrvest erry weight nd totl erry phenolics (TBP) nd totl erry nthocynins (TBA) in Cernet frnc nd Petit Verdot from Cernet frnc TBP TBA Yer R 2 Significnce R 2 Significnce ns ns ns ns Petit Verdot TBP TBA Yer R 2 Significnce R 2 Significnce ns ns ns ns Grpe crotenoids: Cernet frnc. When verged cross ll dtes in 2012, HIGH incresed zexnthin compred to MED nd NO (Tle 18). While lutein 5,6-epoxide ws greter t postfruit set (12 Jul) nd pre-verison (24 Jul) compred to verison (7 Aug), zexnthin ws lower t pre-version compred to post-fruit set nd verison. Lutein nd β-crotene were reltively lower t ech susequent smple dte s the seson progressed. Zexnthin ws incresed y HIGH compred to NO nd MED t pre-verison, nd y HIGH compred to NO t verison (Fig. 11). 59

76 Tle 18. Post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids in Tretment Lutein 5,6- epoxide (µg/g erry) Zexnthin (µg/g erry) Lutein (µg/g erry) β-crotene (µg/g erry) NO MED HIGH Dte 12-Jul Jul Aug c 0.77 c 24-Sep nd nd 0.18 d 0.27 d Significnce c Tretment ns ns ns Dte < < < Tretment*Dte ns ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent. Lutein 5,6-eopoxide nd zexnthin elow detection threshold (nd) t 24 Sep smple. When detected, zexnthin rnged to µg/g erry cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 60

77 Lutein 5,6-epoxide ( g/g erry) A nd NO MED HIGH Zexnthin ( g/g erry) B nd NO MED HIGH Jul 24 Jul 7 Aug 24 Sep Jul 24 Jul 7 Aug 24 Sep Lutein ( g/g erry) C NO MED HIGH -crotene ( g/g erry) D NO MED HIGH Jul 24 Jul 7 Aug 24 Sep 12 Jul 24 Jul 7 Aug 24 Sep Fig. 11 Post-fruit set lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Cernet frnc grpe erries over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 24 Sep. Mens derived from composite 120-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. 61

78 When verged cross ll dtes in 2013, HIGH incresed lutein 5,6-epoxide compred to MED nd NO (Tle 19). Both HIGH nd P-B incresed zexnthin compred to NO nd MED, nd incresed lutein compred to only NO. Lutein 5,6-epoxide ws lower t every susequent smple dte in 2013, while pek zexnthin occurred t pre-verison (26 Jul), which ws greter thn t post-fruit set (14 Jul) nd verison (14 Aug). Both lutein nd β-crotene were greter t ll three pre-hrvest time periods compred to t hrvest. Lutein 5,6-epoxide ws incresed y HIGH when compred to MED t post-fruit set (Fig. 12). Zexnthin ws incresed y HIGH nd P-B when compred to NO nd MED t post-fruit set, nd ws incresed y HIGH when compred to NO nd MED, nd y P-B when compred to NO, t pre-verison. Lutein ws incresed y P-B nd NO when compred to MED t pre-verison nd hrvest, respectively. Tle 19. Pre-loom nd post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids in Tretment Lutein 5,6- epoxide (µg/g erry) Zexnthin (µg/g erry) Lutein (µg/g erry) β-crotene (µg/g erry) NO MED HIGH P-B Dte 14-Jul Jul Aug c Sep nd nd Significnce c Tretment < ns Dte < < < < Tretment*Dte ns ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Lutein 5,6-epoxide nd zexnthin elow detection threshold (nd) in 30 Sep smple. When detected, zexnthin rnged to µg/g erry cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). 62

79 Lutein 5,6-epoxide ( g/g erry) A nd NO MED HIGH P-B Zexnthin ( g/g erry) B c c nd NO MED HIGH P-B Jul 26 Jul 14 Aug 30 Sep Jul 26 Jul 14 Aug 30 Sep Lutein ( g/g erry) C NO MED HIGH P-B -crotene ( g/g erry) D NO MED HIGH P-B Jul 26 Jul 14 Aug 30 Sep 14 Jul 26 Jul 14 Aug 30 Sep Fig. 12 Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Cernet frnc grpe erries over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 30 Sep. Mens derived from composite 120-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. When grpes were smpled from the est cnopy side nd verged over ll dtes in 2014, HIGH incresed lutein 5,6-epoxide when compred to NO, nd oth HIGH nd P-B incresed zexnthin when compred to NO nd MED (Tle 20). Wheres lutein 5,6-epoxide ws greter t post-fruit set (16 Jul) compred to pre-verison (29 Jul) nd verison (11 Aug), zexnthin ws greter t post-fruit set nd pre-verison compred to verison. Lutein ws greter t verison compred to pre-verison nd hrvest (14 Oct), nd β-crotene ws greter t ll pre-hrvest smple dtes compred to t hrvest. Zexnthin ws incresed y HIGH nd P- B when compred to NO nd MED t pre-verison, nd y P-B when compred to ll other 63

80 tretments t verison (Fig. 13). Lutein ws incresed y P-B when compred to NO nd MED t pre-verison. Tle 20. Pre-loom nd post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids from the EAST cnopy side in Tretment Lutein 5,6- epoxide (µg/g erry) Zexnthin (µg/g erry) Lutein (µg/g erry) β-crotene (µg/g erry) NO MED HIGH P-B Dte 16-Jul Jul Aug Oct nd nd 0.40 c 0.59 Significnce c Tretment < ns ns Dte < < < < Tretment*Dte ns ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Lutein 5,6-epoxide nd zexnthin elow detection threshold (nd) in 14 Oct smple. When detected, lutein 5,6-epoxide rnged to µg/g erry, nd zexnthin rnged to µg/g erry, cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). 64

81 Lutein 5,6-epoxide ( g/g erry) A nd NO MED HIGH P-B Zexnthin ( g/g erry) B c c nd NO MED HIGH P-B Jul 29 Jul 11 Aug 14 Oct Jul 29 Jul 11 Aug 14 Oct Lutein ( g/g erry) C NO MED HIGH P-B -crotene ( g/g erry) D NO MED HIGH P-B Jul 29 Jul 11 Aug 14 Oct Jul 29 Jul 11 Aug 14 Oct Fig. 13 Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Cernet frnc grpe erries collected from the EAST cnopy side over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 14 Oct. Mens from composite 150-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. NOTE: 16 Jul is n verge of est nd west cnopy sides. When grpes were smpled from the west cnopy side nd verged over ll dtes in 2014, oth HIGH nd P-B incresed zexnthin when compred to MED nd NO (Tle 21). Lutein 5,6-epoxide followed similr trends over the seson compred to the est cnopy side nd ws greter t post-fruit set (16 Jul) compred to the susequent two smpled dtes. Zexnthin nd lutein followed similr trends, reching pek levels t pre-verison (29 Jul) compred to prior nd susequent smple dtes. β-crotene ws greter t post-fruit set nd pre-verison compred to verison (11 Aug) nd hrvest (14 Oct). Lutein 5,6-epoxide ws incresed y HIGH when compred to ll other tretments t pre-verison (Fig. 14). Zexnthin ws 65

82 incresed y HIGH when compred to MED nd NO t oth pre-verison nd verison, nd ws incresed y P-B when compred to NO t pre-verison nd when compred to MED nd NO t verison. Tle 21. Pre-loom nd post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids from the WEST cnopy side in Tretment Lutein 5,6- epoxide (µg/g erry) Zexnthin (µg/g erry) Lutein (µg/g erry) β-crotene (µg/g erry) NO MED HIGH P-B Dte 16-Jul Jul Aug Oct nd nd 0.42 c 0.63 Significnce c Tretment ns < ns ns Dte < < < < Tretment*Dte ns ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Lutein 5,6-epoxide nd zexnthin elow detection threshold (nd) in 14 Oct smple. When detected, lutein 5,6-epoxide rnged to µg/g erry, nd zexnthin rnged to µg/g erry, cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). 66

83 Lutein 5,6-epoxide ( g/g erry) A nd 16 Jul 29 Jul 11 Aug 14 Oct NO MED HIGH P-B Zexnthin ( g/g erry) B c c nd 16 Jul 29 Jul 11 Aug 14 Oct NO MED HIGH P-B Lutein ( g/g erry) C NO MED HIGH P-B -crotene ( g/g erry) D NO MED HIGH P-B Jul 29 Jul 11 Aug 14 Oct Jul 29 Jul 11 Aug 14 Oct Fig. 14 Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Cernet frnc grpe erries collected from the WEST cnopy side over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 14 Oct. Mens from composite 150-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. NOTE: 16 Jul is n verge of est nd west cnopy sides. When verged over the course of , HIGH incresed lutein 5,6-epoxide nd zexnthin compred to NO nd MED, nd incresed lutein when compred to only MED (Tle 22). Lutein 5,6-epoxide nd β-crotene were reltively lower t ech susequent smple dte compred to previous smple dtes. Zexnthin nd lutein followed similr ptterns levels of oth crotenoids were greter t post-fruit set nd pre-version compred to lter smple dtes. Lutein 5,6 epoxide ws greter in 2013 nd 2014 compred to 2012 while zexnthin ws greter in 2014 compred to 2012 nd When compred to 2014, β-crotene ws greter in 2012 nd lower in

84 The interction etween tretment nd yer for lutein 5,6-epoxide showed tht it ws comprtively greter in HIGH compred to other tretments in 2013, nd there ws reltively less difference etween tretments in 2012 nd The interction etween normlized smple dte (NSD) nd yer for lutein 5,6-epoxide reveled tht its reltionship chnged cross smple dtes over 2012 to 2014 such tht it ws similr t pre-verison nd post-fruit set in 2012, reltively different etween ll NSDs t 2013, ut similr t pre-verison nd verison in The interction etween tretment nd yer for zexnthin showed tht the mgnitude of difference in its levels chnged etween HIGH compred to MED nd NO over the yers: zexnthin levels in HIGH were reltively similr to NO nd MED in 2012, while zexnthin levels in HIGH were reltively greter thn NO nd Med in 2013 nd The interction etween NSD nd yer for zexnthin showed tht it ws lower t pre-verison compred to post-fruit set nd verison in 2012, nd, in 2013 nd 2014, it ws reltively greter t preverison compred to post-fruit set nd verison. The interction etween NSD nd yer for lutein reveled tht its reltionship etween smple dtes chnged over : it ws greter t post-fruit set thn t pre-verison nd verison in 2012, ut greter t pre-verison nd verison compred to post-fruit set in The interction etween NSD nd yer for β- crotene showed tht its reltionship etween severl NSDs chnged over the course of , ut nmely tht it ws different etween ll NSDs in 2012, reltively similr etween postfruit set, pre-verison, nd verison, ll which were greter thn t hrvest, in 2013, nd reltively similr etween verison nd hrvest, which were lesser thn t post-fruit set nd preverison, in

85 Tle 22. Post-fruit set lef removl tretment effect on Cernet frnc grpe crotenoids over Tretment Lutein 5,6- epoxide (µg/g erry) Zexnthin (µg/g erry) Lutein (µg/g erry) β-crotene (µg/g erry) NO MED HIGH NSD Post-fruit set Pre-verison Verison c c Hrvest nd nd 0.30 c 0.42 d Yer c c Significnce d Tretment < ns NSD < < < Yer < < ns < Tretment*NSD ns ns ns ns Tretment*Yer ns ns NSD*Yer < < < < Tretment*NSD*Yer ns ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; NOTE: P-B left out of nlysis ecuse it ws not implemented in NSD = normlized smple dte in order to evlute the effect of smple dte order cross yers. Lutein 5,6-eopoxide nd zexnthin were elow detection threshold (nd) in severl hrvest smples. When detected, lutein 5,6-epoxide rnged to µg/g erry cross tretments in 2014, nd zexnthin rnged to µg/g erry cross tretments over c 2014 dt ws verged cross cnopy sides. d Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. Petit Verdot. When verged cross ll dtes in 2012, MED incresed lutein 5,6-epoxide compred to NO nd HIGH, nd MED nd HIGH incresed zexnthin compred to NO (Tle 23). Excepting lutein 5,6-epoxide, crotenoids were greter t pre-hrvest compred to t hrvest (1 Oct). Pek zexnthin ws oserved t post-fruit set (12 Jul) nd verison (20 Aug) while pek β-crotene ws oserved t post-fruit set. The interction ws the reversed reltionship etween zexnthin in MED nd HIGH from pre-verison (24 Jul) to verison, or 69

86 the generl decrese in the mgnitude of difference in zexnthin etween tretments t hrvest when compred to ll previous periods (Fig. 15). Lutein 5,6-epoxide ws incresed y MED nd HIGH when compred to NO t pre-verison (Fig. 15). Zexnthin ws incresed y MED nd HIGH when compred to NO t post-fruit-set, nd incresed y HIGH when compred to NO t pre-verison, nd y MED when compred to NO t verison. Lutein ws greter in NO compred to HIGH t hrvest. Tle 23. Post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids in Lutein 5,6- Tretment Zexnthin Lutein β-crotene epoxide (µg/g erry) (µg/g erry) (µg/g erry) (µg/g erry) NO MED HIGH Dte 12-Jul Jul Aug Oct nd c c Significnce c Tretment < ns ns Dte ns < < < Tretment*Dte ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent. Lutein 5,6-eopoxide elow detection threshold (nd) in 1 Oct smple c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 70

87 Lutein 5,6-epoxide ( g/g erry) A nd NO MED HIGH Zexnthin ( g/g erry) B NO MED HIGH Jul 24 Jul 20 Aug 1 Oct Jul 24 Jul 20 Aug 1 Oct C NO MED HIGH 5 D NO MED HIGH Lutein ( g/g erry) crotene ( g/g erry) Jul 24 Jul 20 Aug 1 Oct 0 12 Jul 24 Jul 20 Aug 1 Oct Fig. 15 Post-fruit set lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Petit Verdot grpe erries over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 1 Oct. Mens derived from composite 120-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. When verged cross ll dtes in 2013, HIGH incresed lutein 5,6-epoxide compred to MED, nd oth HIGH nd P-B incresed zexnthin nd lutein compred to NO nd MED (Tle 24). Lutein 5,6-epoxide ws greter t post-fruit set (14 Jul) nd pre-verison (26 Jul) compred to t version (27 Aug), nd zexnthin ws greter t ll three pre-hrvest dtes compred to t hrvest (9 Oct). Pek lutein nd β-crotene ws oserved t post-fruit set, t which time oth of these crotenoids were greter compred to ll prior nd susequent smple dtes. The interction of tretments nd dte on lutein ws either tht the mgnitude of difference in lutein chnged etween HIGH nd P-B from post-fruit set to pre-verison, or the 71

88 generl decrese in the mgnitude of difference in lutein etween tretments over the course of the 2013 seson (Fig. 16). Lutein 5,6-epoxide ws incresed y HIGH when compred to NO t post-fruit set, nd, t pre-verison, ws greter in HIGH when compred to MED nd NO nd greter in P-B when compred to MED (Fig. 16). Zexnthin ws incresed y HIGH nd P-B when compred to NO nd MED t pre-verison, nd y only HIGH when compred to NO nd MED t verison. Lutein ws incresed y oth HIGH nd P-B when compred to NO nd MED t pre-verison. β-crotene ws greter in NO compred to HIGH t hrvest. Tle 24. Pre-loom nd post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids in Lutein 5,6- Tretment Zexnthin Lutein β-crotene epoxide (µg/g erry) (µg/g erry) (µg/g erry) (µg/g erry) NO MED HIGH P-B Dte 14-Jul Jul Aug c 0.68 c 9-Oct nd d 0.64 c Significnce c Tretment < < ns Dte < < < < Tretment*Dte ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. Lutein 5,6-eopoxide elow detection threshold (nd) in 9 Oct smple. When detected, vlues rnged to µg/g erry cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 72

89 Lutein 5,6-epoxide ( g/g erry) A c c nd NO MED HIGH P-B Zexnthin ( g/g erry) B NO MED HIGH P-B Jul 26 Jul 27 Aug 9 Oct Jul 26 Jul 27 Aug 9 Oct Lutein ( g/g erry) C NO MED HIGH P-B -crotene ( g/g erry) D NO MED HIGH P-B Jul 26 Jul 27 Aug 9 Oct 14 Jul 26 Jul 27 Aug 9 Oct Fig. 16 Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Petit Verdot grpe erries over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 9 Oct. Mens derived from composite 120-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. When grpes were smpled from the est cnopy side nd verged over ll dtes in 2014, HIGH incresed lutein 5,6-epoxide compred to NO nd MED, nd P-B increse lutein 5,6-epoxide compred to NO (Tle 25). While oth HIGH nd P-B incresed zexnthin, only HIGH incresed lutein, when compred to NO nd MED. Pek lutein 5,6-epoxide nd lutein ws oserved t pre-verison (11 Aug), t which time these crotenoids were greter compred to ll prior nd susequent smple dtes. Zexnthin ws greter t post-fruit set (16 Jul) nd verison (10 Sep) compred to hrvest (22 Oct), nd β-crotene ws greter t pre-verison nd verison compred to hrvest. 73

90 There ws n interction etween tretment nd dte for every crotenoid, s well s for totl crotenoids nd djusted totl crotenoids. The interction for lutein 5,6-epoxide nd lutein ws the reversed reltionship etween these crotenoids in HIGH nd P-B from pre-verison to verison (Fig. 17). Further, there ws generl decrese in the mgnitude of difference in lutein oserved etween tretments, prticulrly from verison to hrvest. The interction for zexnthin ppered to e the chnge in its reltionship etween HIGH nd P-B over the course of the seson, or the decrese in the mgnitude of difference oserved etween P-B nd HIGH in comprison to MED nd NO, prticulrly from verison to hrvest. The interction for β- crotene ws the chnge in the reltionship etween P-B nd HIGH from pre-verison to verison, or etween P-B nd NO from verison to hrvest. Lutein 5,6-epoxide ws greter in HIGH when compred to ll tretments t post-fruit set, nd when compred to MED nd NO t pre-verison (Fig. 19). Pre-loom lef removl (P-B) incresed lutein 5,6-epoxide when compred to NO nd MED t verison. Both HIGH nd P-B incresed zexnthin compred to NO nd MED t every pre-hrvest smple dte. While HIGH hd greter lutein concentrtion when compred to ll other tretments t pre-verison, P-B hd greter lutein concentrtion when compred to ll other tretments t verison. β-crotene ws incresed y only P-B when compred to NO t verison. 74

91 Tle 25. Pre-loom nd post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids from the EAST cnopy side in Lutein 5,6- Tretment Zexnthin Lutein β-crotene epoxide (µg/g erry) (µg/g erry) (µg/g erry) (µg/g erry) NO c MED c HIGH P-B Dte 16-Jul Aug Sep c c Oct nd c 1.59 d 1.81 Significnce c Tretment < < Dte < < < Tretment*Dte NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. 16 Jul is n verge cross oth cnopy sides, ll other dtes were smpled from the est cnopy side only. Lutein 5,6-eopoxide elow detection threshold (nd) in 22 Oct smple. When detected, vlues rnged to µg/g erry cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 75

92 Lutein 5,6-epoxide ( g/g erry) A nd NO MED HIGH P-B Zexnthin ( g/g erry) B NO MED HIGH P-B Lutein ( g/g erry) C 16 Jul 11 Aug 10 Sep 22 Oct NO MED HIGH P-B -crotene ( g/g erry) D 16 Jul 11 Aug 10 Sep 22 Oct NO MED HIGH P-B Jul 11 Aug 10 Sep 22 Oct 0 16 Jul 11 Aug 10 Sep 22 Oct Fig. 17 Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Petit Verdot grpe erries collected from the EAST cnopy side over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 22 Oct. Mens from composite 150-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. NOTE: 16 Jul is n verge of est nd west cnopy sides. 76

93 When grpes were smpled from the west cnopy side nd verged over ll dtes in 2014, HIGH nd P-B incresed zexnthin when compred to NO nd MED, nd only HIGH incresed lutein when compred to MED (Tle 26). Excepting lutein 5,6-epoxide, crotenoids were greter t pre-verison (11 Aug) when compred to post-fruit set (16 Jul) nd hrvest (22 Oct). The interction for zexnthin ppered to e the chnge in its reltionship etween HIGH nd P-B etween post-fruit set nd verison, or the generl decrese in its mgnitude of difference etween tretments, prticulrly from verison to hrvest (Fig. 18). The interction for lutein ws the chnge in its reltionship etween P-B nd ll other tretments from preversion to version. This interction my lso hve een due to the generl decrese in the mgnitude of difference in lutein etween HIGH nd MED/NO from pre-verison to verison, or the decrese in its mgnitude of difference oserved etween P-B nd other tretments from verison to hrvest. Lutein 5,6-epoxide ws incresed y HIGH when compred to ll other tretments t post-fruit set, nd incresed y P-B when compred to NO nd MED t verison (Fig. 18). Zexnthin ws incresed y HIGH nd P-B compred to NO nd MED t post-fruit set, nd, t pre-verison, y HIGH when compred to NO nd MED nd y P-B when compred to MED. At verison, zexnthin ws lso incresed y HIGH when compred to ll other tretments, nd y P-B when compred to NO nd MED. Lutein ws incresed y HIGH when compred to P-B t pre-verison, nd y P-B when compred to ll other tretments t verison. β-crotene ws incresed y P-B when compred to ll other tretments t verison. 77

94 Tle 26. Pre-loom nd post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids from the WEST cnopy side in Lutein 5,6- Tretment Zexnthin Lutein β-crotene epoxide (µg/g erry) (µg/g erry) (µg/g erry) (µg/g erry) NO MED HIGH P-B Dte 16-Jul Aug Sep c Oct nd c 1.68 d 1.98 Significnce c Tretment ns < ns Dte < < < < Tretment*Dte ns < < NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. 16 Jul is n verge cross oth cnopy sides, ll other dtes were smpled from the est cnopy side only. Lutein 5,6-eopoxide elow detection threshold (nd) in 22 Oct smple. When detected, vlues rnged to µg/g erry cross tretments. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. 78

95 Lutein 5,6-epoxide ( g/g erry) A nd NO MED HIGH P-B Zexnthin ( g/g erry) B c c c c NO MED HIGH P-B Lutein ( g/g erry) C 16 Jul 11 Aug 10 Sep 22 Oct NO MED HIGH P-B -crotene ( g/g erry) D 16 Jul 11 Aug 10 Sep 22 Oct NO MED HIGH P-B Jul 11 Aug 10 Sep 22 Oct 0 16 Jul 11 Aug 10 Sep 22 Oct Fig. 18 Post-fruit set nd pre-loom lef removl effect on lutein 5,6-epoxide (A), zexnthin (B), lutein (C), nd -crotene (D) in Petit Verdot grpe erries collected from the WEST cnopy side over the course of NOTE: lutein 5,6-epoxide nd zexnthin consistently not deteced (nd) in erries t 22 Oct. Mens from composite 150-erry smples; n = 6. Tretments not shring letter within dte re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. NOTE: 16 Jul is n verge of est nd west cnopy sides. When verged over the course of , HIGH incresed the level of ll crotenoids except β-crotene when compred to NO nd MED (Tle 27). Excepting zexnthin, crotenoids were greter t post-fruit set nd pre-verison compred to t verison nd hrvest, nd were greter t version compred to t hrvest. Lutein 5,6-epoxide ws greter in 2013 compred to 2012 nd 2014 wheres zexnthin, lutein, nd β-crotene were lower in 2013 compred to 2012 nd

96 The interction etween tretments nd normlized smple dte (NSD) for zexnthin showed tht there ws reltively no difference in it etween tretments t hrvest compred to ll other smple dtes, prticulrly those t post-fruit set nd pre-verison, when trends were HIGH > MED > NO. There ws similr interction etween tretment nd NSD for lutein, which reveled tht there ws no difference in it etween tretments t oth verison nd hrvest, ut trends were HIGH > MED = NO t oth post-fruit set nd pre-verison. The interction etween tretment nd yer for lutein 5,6-epoxide showed tht levels were greter in MED compred to HIGH nd NO in 2012, ut were greter in HIGH compred to MED nd NO in 2013 nd The interction etween tretment nd yer for zexnthin reveled tht it ws differentilly ffected y MED over 2012 nd 2013: MED hd similr levels to HIGH in 2012, ut similr levels to NO in The interction etween tretment nd yer for lutein showed tht levels were similr etween tretments in 2012, ut reltively greter in HIGH compred to MED nd NO in 2013 nd The oserved interction etween NSD nd yer for lutein 5,6-epoxide reveled tht it ws reltively similr etween smple periods in 2012 compred to 2013, or tht levels were greter t post-fruit set compred to pre-version in 2013, nd vice-vers in in The interction etween nd NSD nd yer for zexnthin reveled tht there ws less difference in it etween smple dtes in 2013 compred to 2012 nd 2014, when the lrgest difference ws etween post-fruit set nd hrvest. The interction etween NSD nd yer for lutein reveled tht there ws less difference in it t verison compred to post-fruit set nd pre-loom in 2012, nd reltively greter difference in it etween these time periods in 2013 nd Another oserved interction for lutein ws tht it ws greter t post-fruit set thn pre-verison in 2012, nd vice-vers in The interction etween NSD nd yer for β-crotene showed tht it ws 80

97 reltively greter t pre-version thn other smpled dtes in 2013 compred to 2012 nd 2014, nd tht it ws greter t post-fruit set nd version compred to pre-version in 2012, ut greter t pre-version compred to post-fruit set nd verison in 2013 nd The three-wy interction etween tretment, NSD, nd yer showed tht the difference in zexnthin etween ll pre-hrvest nd hrvest smples over ws gretest in HIGH, followed y MED, nd then NO. Further, this three-wy interction showed tht there ws reltively little difference in zexnthin etween smpled dtes in 2013 compred to 2012 nd 2014, ut this ws only the cse in MED nd NO tretments. 81

98 Tle 27. Post-fruit set lef removl tretment effect on Petit Verdot grpe crotenoids over Lutein 5,6- Tretment Zexnthin Lutein β-crotene epoxide (µg/g erry) (µg/g erry) (µg/g erry) (µg/g erry) NO c MED HIGH NSD Post-fruit set Pre-verison Verison c Hrvest nd c 1.23 c 1.10 d Yer c c 1.43 c c Significnce d Tretment < < ns NSD < < < < Yer < < < < Tretment*NSD ns < < ns Tretment*Yer < < ns NSD*Yer < < < < Tretment*NSD*Yer ns ns ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; NOTE: P-B left out of nlysis ecuse it ws not implemented in NSD = normlized smple dte in order to evlute the effect of smple dte order cross yers. Lutein 5,6-eopoxide nd zexnthin elow detection threshold (nd) in severl hrvest smples. When detected, Lutein 5,6-epoxide nd zexnthin vlues rnged to µg/g erry cross tretments over c 2014 dt ws verged cross cnopy sides. NOTE: no lutein 5,6-epoxide detected t hrvest in d Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey s HSD. The zexnthin: lutein rtio in Cernet frnc grpes ws greter in HIGH when compred to NO t every smple dte, nd when compred to MED t hrvest (Fig. 19). The zexnthin: lutein rtio in Petit Verdot grpes ws greter in HIGH when compred to NO t every smple dte, nd when compred to MED t pre-verison nd verison. The zexnthin: lutein rtio in Petit Verdot grpes ws greter in MED when compred to NO t hrvest. 82

99 Zexnthin: lutein rtio Post-fruit set Pre-verison Verison A Zexnthin: lutein rtio Post-fruit set Pre-verison Verison Hrvest NO MED HIGH B 0.00 NO MED HIGH Fig. 19. Cernet frnc (A) nd Petit Verdot (B) grpe zexnthin: lutein rtios t different phenologicl stges s ffected y no lef removl (NO), nd post-fruit set lef removl to medium (MED) nd high (HIGH) extents. Dt verged over ; n = 18. Tretment rs not shring letter within phenologiicl stge re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. 83

100 The chnge in Cernet frnc grpe crotenoids from pre-verison to hrvest (ΔC) ws greter in HIGH compred to MED nd NO in 2014 only (Fig. 20). The Petit Verdot ΔC ws greter in HIGH compred to MED in 2013, nd greter in HIGH compred to NO in A NO MED HIGH C ( g/g erry) B NO MED HIGH C (mg/g erry) Fig. 20. The chnge in Cernet frnc (A) nd Petit Verdot (B) grpe crotenoids ( C) from pre-verison to hrvest s ffected y no lef removl (NO), nd post-fruit set removl to medium (MED) nd high (HIGH) extents from Tretment rs not shring letter within yer re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error, n = 6. NOTE: Lutein nd -crotene totled in Cernet frnc; lutein, -crotene, nd zexnthin totled in Petit Verdot. 84

101 Tringle difference test: Cernet frnc nd Petit Verdot. Pnelists were le to determine the difference in Cernet frnc wine color etween HIGH nd NONE, nd the difference in Cernet frnc wine rom etween MED nd NO (Tle 28). Pnelists were le to determine the difference in Petit Verdot wine rom etween MED nd NO, nd MED nd HIGH. Tle 28. Tringle difference test of Cernet frnc nd Petit Verdot wines from the 2012 vintge. Cernet frnc Tretment comprison Attriute Pnelist # Successes Significnce NO vs. HIGH Color NO vs. HIGH Arom 26 9 ns NO vs. HIGH Tste 26 7 ns NO vs. MED Color ns NO vs. MED Arom NO vs. MED Tste 23 7 ns MED vs. HIGH Color ns MED vs. HIGH Arom ns MED vs. HIGH Tste ns Petit Verdot Comprison Attriute Pnelist # Successes Significnce NO vs. HIGH Color ns NO vs. HIGH Arom 21 3 ns NO vs. HIGH Tste 21 8 ns NO vs. MED Color 17 9 ns NO vs. MED Arom NO vs. MED Tste 17 8 ns MED vs. HIGH Color 22 9 ns MED vs. HIGH Arom MED vs. HIGH Tste 22 6 ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level) determined using n Alph one-tiled nlysis. Consumer preference test: Cernet frnc nd Petit Verdot. The intensity of red color nd stringency of P-B wines rnked higher on the just out right scle compred to MED wines in Cernet frnc (Tle 29). The intensity of red color of P-B nd HIGH wines rnked higher on the just out right scle compred to MED wines in Petit Verdot. 85

102 Tle 29. Consumer preference of Cernet frnc nd Petit Verdot wines from the 2013 vintge. Cernet frnc Attriute d NO MED HIGH P-B Significnce e Appernce n/ ns Red color n/ ns Red color intensity n/ Arom n/ ns Fruity rom n/ ns Vegettive rom n/ ns Fruity flvor n/ ns Vegettive flvor n/ ns Overll flvor n/ ns Astringency n/ ns Astringency intensity n/ Mouthfeel n/ ns Mouthfeel intensity n/ ns Length of finish n/ ns Length of finish intensity n/ ns Overll impression n/ ns Side-y-side rnking n/ ns Petit Verdot c Attriute d NO MED HIGH P-B Significnce e Appernce ns Red color ns Red color intensity Arom ns Fruity rom ns Vegettive rom ns Fruity flvor ns Vegettive flvor ns Overll flvor ns Astringency ns Astringency intensity ns Mouthfeel ns Mouthfeel intensity ns Length of finish ns Length of finish intensity ns Overll impression ns Side-y-side rnking ns NO = no lef removl; MED = post-fruit set removl of leves to medium extent; HIGH = post-fruit set removl of leves to high extent; P-B = pre-loom lef removl of six sl leves nd lterls. 84 pnelists prticipted in the Cernet frnc consumer preference test. c 61 pnelists prticipted in the Petit Verdot consumer preference test. dthe following ttriutes rnked on scle of 1-9: ppernce, red color, rom, overll flvor, mouthfeel, length of finish, nd overll impression; the following ttriutes rnked on scle of 1-5: red color intensity, fruity nd vegettive rom, fruit nd vegettive flvor, stringency intensity, mouthfeel intensity, nd length of finish intensity; side-y-side rnking ws n verge of tretment rnking order, with 1 eing the fvorite. e Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (rows) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Duncn;s Multiple Rnge test. 86

103 Discussion Berry temperture, mient rdition, nd mient ir temperture: Grpe erry temperture ws primrily driven y mient ir temperture, s shown efore (Bergvist et l. 2001, Col et l. 2009). The greter estimted time tht erries spent 30 or 35 C in 2013 compred to 2014 ws likely due to the reltively wrmer ir tempertures experienced in 2013, s higher grpe tempertures re chrcteristic of wrmer regions nd sesons (Bergqvist et l. 2001, Frin et l. 2010). Rdition cn confound erry temperture (Col et l. 2009). In the current study, it ws shown to increse erry temperture, s others hve demonstrted (Bergvist et l. 2001, Spyd et l. 2002, Trr et l. 2008). As long s rdition ws not locked y cloud or the grpevine cnopy, higher rdition typiclly incresed grpe temperture ove mient ir temperture, supportive tht direct sunlight rdintly hets fruits more efficiently thn diffuse light (Smrt nd Sinclir 1976). Sunlight nd temperture re two primry sources of vrition in grpe flvonoids cross sesons (Downey et l. 2006), nd temperture nd sunlight lter the synthesis nd degrdtion ptterns of grpe crotenoids nd norisoprenoids (Frin et l. 2010, Lee et l. 2007). Thus, it ws nticipted tht grpes smpled from different sesons nd/or lef removl tretments would hve different concentrtions of these compounds. However, the vrile cloudiness nd moderte climte of the current study limited temperture differences experienced etween shded nd exposed erries, s elorted upon elow. The current study ws conducted in north-south oriented, verticlly-shoot positioned vineyrd in humid growing region. Consequently, vrile cloudiness ws common, nd when solr rdition reched its diurnl pex, it ws locked y the ove-hed cnopy. Further, while mient PAR ws reltively greter in the AM, mient ir temperture ws reltively greter in the PM. Thus, the diurnl period of gretest rdint heting potentil of grpes ws voided, nd 87

104 concomitnt high mient temperture nd rdition conditions were infrequently experienced when the sun ws ngled on the fruit-zone of either cnopy side. These wether ptterns were not conducive to extreme heting of erries, nor to remrkle differences in erry temperture etween cnopy sides. These ptterns were nticipted to contrst those of wrmer/rid regions (Bergvist et l. 2001, Spyd et l. 2002, Trr et l. 2008). Those studies reported tht exposed erry temperture gretly differed etween cnopy sides, nd tht exposed erry temperture ws frequently greter thn tht of shded erries, even t middy (Bergvist et l. 2001), even when lef removl ws not prcticed (Trr et l. 2008), nd often to C (Bergvist et l. 2001, Spyd et l. 2002, Trr et l. 2008). Bergqvist et l. (2001) noted tht differences in erry temperture etween shded nd exposed fruit-zones my e less extreme in cooler regions. The results of the current study re in greement, ut would dd tht temperture difference etween well-exposed nd shded grpes my e less pronounced even in even wrm, humid growing regions due to vrile cloudiness. Vrile sky conditions mke erry temperture difficult to predict, prticulrly in humid regions (Col et l. 2009). In greement with previous work (Bergvist et l. 2001) the results of the current study confirm tht rdint heting of grpes is the primry determinnt of erry temperture difference etween exposed nd shded grpes nd, when often chnging, would confound the ility to forecst erry temperture. Thus, it cnnot e ssumed tht exposed grpes will necessrily experience gretly different rdition nd tempertures compred to shded erries, ut this will e highly dependent on the temporl difference in sky conditions in humid region. The climtic differences etween the current study nd other studies tht wellchrcterized erry temperture (Bergqvist et l. 2001, Spyd et l. 2002, Trr et l. 2008) 88

105 support tht modelling is n undeveloped pproch when trying to compre regionl difference in erry temperture nd its impct on grpe composition (Bond nd Sdrts 2014). Berry temperture ws mesured periodiclly over the sesons. It is unlikely, however, tht single erry temperture events impct fruit composition, especilly in field setting where temperture nd rdition cnnot e mnipulted without strtegic experimentl design (Spyd et l. 2002, Trr et l. 2008). Thus, sttisticl difference in erry temperture ws determined on occsion, ut this ws of questionle iologicl significnce. Compred to shded grpes, the temperture of exposed grpes ws mesured more frequently, nd estimted to spend more time, 30 or 35 C. However, grpe tempertures in the current study were cooler thn others (Bergqvist et l. 2001), nd it ws estimted tht grpes did not spend nerly s much time 30 or 35 C compred to other studies (Pstore et l. 2013, Trr et l. 2008), regrdless of cnopy side or lef removl mgnitude. In the current study, there ws likely less impct on fruit composition, prticulrly nthocynins, etween exposed/shded fruit-zones nd cnopy sides when compred to previous studies (Bergvist et l. 2001, Spyd et l. 2002, Trr et l. 2008). These studies were conducted in regions where fternoon temperture nd rdition re likely concomitntly high, thus fvoring rdint grpe heting to tempertures tht re detrimentl to grpe nthocynin ccumultion, prticulrly on the west or south side cnopy side. In humid regions, it my e erroneous to chrcterize cnopy side y the reltive term hot, especilly in north-south oriented rows nd prticulrly under cloudy conditions. Due to modest erry temperture increses nd decresed disese incidence experienced in open fruit-zones (English et l. 1989, Wolf et l. 1986), cnopy mngement prctices tht provide direct light into the fruit-zone my e etter suited for temperte humid regions, contrsting the cnopy mngement recommendtions in reltively wrmer/rid regions (Bergqvist et l. 2001, Spyd et l. 2002). 89

106 Berries mesured in Petit Verdot nd pre-loom lef removl plots hd reltively lower temperture nd spent less estimted time 30 C when compred to Cernet frnc nd ggressive post-fruit set lef removl, respectively. Both Petit Verdot nd P-B hd smller erries nd looser clusters compred to Cernet frnc nd HIGH, respectively, which cn result in comprtively lower erry tempertures compred to when erries touch ech other in tighter cluster (Keller 2010). Though nectodl, the greter grpe surfce re exposed to ir flow, the greter the evportive cooling. As such, pre-loom lef removl my not only e prudent prctice to reduce sunurn incidence (Pstore et l. 2013), ut lso llevite het lods on fruit when mient wether conditions re conducive to extreme heting. When compred to Cernet frnc, Petit Verdot ws more vigorous nd hd lrger cnopies, nd hd post-verison period tht ws reltively shorter, nd lter-shifted. These oservtions mke cse for growing Petit Verdot, nd other lter-mturing red vrieties, if temperture-induced nthocynin degrdtion is of concern. Yet vrieties my differ in their susceptiility to temperture-induced chnges in grpe nthocynins (Kotseridis et l. 2012, Mori et l. 2007). Components of crop yield: As shown y others, crop yield cn e reduced y lef removl timing nd extent (Intrieri et l. 2008, Kliewer 1970, Pstore et l. 2013, Trdguil et l. 2010). This reduction in crop is primrily due to reduction in cluster weight oserved in pre-loom lef removl plots. Removing leves efore loom reduced crop yield more consistently, nd to greter extent, thn did ny post-fruit set lef removl tretment. This ws nticipted, s lef removl hd lesser impct on crop yield reduction when implemented well fter compred to immeditely fter loom (Cndolfi-Vsconcelos nd Kolet 1990). Pre-loom lef removl reduced crop yield y n verge of 50-53% when compred to no lef removl over of the current study. Crop yield reduction rnged from 30-71% cross severl different 90

107 mgnitudes of pre-loom lef removl, vrieties, nd climtes (Digo et l. 2012, Plliotti et l. 2011, Pstore et l. 2013, Poni et l. 2006, 2008, nd 2010, Gtti et l. 2012, Intrieri et l. 2008, Trdguil et l nd 2012). Crop yield reduction ws primrily function of pre-loom lef removl s impct on cluster weight, which ws reduced y n verge of 37-39% in Cernet frnc nd Petit Verdot in the current study. While pre-loom lef removl did not differentilly impct cluster weight etween these vrieties, erry numer per cluster nd erry weight were differentilly ffected (discussed elow). Others hve reported 20-69% reduction in cluster weight nd 14-64% reduction in erry numer per cluster (Digo et l. 2012, Plliotti et l. 2011, Pstore et l. 2013, Poni et l. 2006, 2008, nd 2010, Gtti et l. 2012, Intrieri et l. 2008, Trdguil et l nd 2012) s function of pre-loom lef removl. These studies lso reported reduction in fruit set y 8-37%, which ppers to e the primry mechnism y which pre-loom lef removl reduced crop yield. Crohydrte supply to developing flowers is necessry from inflorescence initition in the previous summer through fruit set in the current seson for successful erry formtion (Cndolfi-Vsconcelos nd Kolet 1990, Leon et l. 2008). Source tissue limittion efore loom cn result in flower ortion, prticulrly y disturing meiosis, s nthers nd ovules re prticulrly sensitive to crohydrte deprivtion (Sini 1997, Jen nd Lpointe 2001, Leon et l. 2008). Since crohydrte supply is importnt for fruit set, removl of current seson source tissues cn result in flower necrosis nd scission (Cndolfi-Vsconcelos nd Kolet 1990 nd 1991, Cspri nd Lng 1996, Cspri et l. 1998, Coome 1962). It is frequently cool, cloudy, nd riny in the region tht the current study ws conducted in, nd these conditions re unfvorle for pollen viility nd germintion rtes (Kolet 1966). Though loom-time wether ws not recorded t the experimentl vineyrd, the verge temperture in the week 91

108 corresponding to loom ws 20.5 C in 2013 nd 18.8 C in 2014, s logged with wether sttion locted ~ 29 km wy from the Cernet frnc nd Petit Verdot experimentl vineyrd. These tempertures re elow 25.0 C, the lower threshold of the optiml temperture rnge for pollen germintion (Keller 2010). Though the entire reduction in crop yield cnnot e ttriuted to lower loom-time temperture, there were drsticlly lower erry numers per cluster nd other crop yield components in 2014 compred to Fruit set, flower scission, nd vine or inflorescence crohydrte sttus were not mesured in the current study. However, the reduction in erry numer per cluster suggested tht fruit set ws reduced due to the tndem reduction in crohydrte supply nd su-optiml tempertures surrounding loom. Pre-loom lef removl reduced erry numer per cluster y 14-64% in severl studies (Digo et l. 2012, Plliotti et l. 2011, Pstore et l. 2013, Poni et l. 2006, 2008, nd 2010, Gtti et l. 2012, Intrieri et l. 2008, Trdiguil et l nd 2012). While crohydrte sttus ws not mesured, mny of the ove studies cited the reduction in source tissue mount/crohydrte supply efore loom to e the primry cuse of reduced erry numer per cluster (Cndolfi-Vsconcelos nd Kolet 1990, Cspri nd Lng 1996, Coome 1962). The removl of more leves in Treino resulted in greter reduction in fruit set nd erry numer per cluster compred to removing fewer leves in Sngiovese (Poni et l. 2006). While this comprison ws confounded y experimentl setting (Treino in field; Sngiovese in pot), it did support tht crohydrte supply is importnt Pre-loom, or erly-seson, lef removl differentilly impcted erry numer per cluster etween vrieties in the current study s well s other studies (Poni et l nd 2008, Stini et l. 2010, Trdguil et l. 2010). While Trdguil et l. (2010) did not elorte on findings, vigorous shoots in Vignoles were suggested to serve s more competitive sinks 92

109 compred to V. vinifer vrieties, resulting in comprtively fewer Vignoles erries per cluster (Stini nd Howell 2010). In the current study, Petit Verdot (necdotlly) hd reltively greter vigor thn Cernet frnc, ut pre-loom lef removl reduced Cernet frnc erry numer per cluster to comprtively greter extent. Further reltively less lef re re-growth occurred, interpreted s lower vine vigor, in vrieties in which erry numer per cluster ws reduced to reltively greter extent (Poni et l nd 2008). Thus, the inherent vigor of vriety my not necessrily e good proxy for pre-loom lef removl s differentil impct on erry numer per cluster etween vrieties, s suggested in Stini nd Howell (2010). The reltively greter vine vigor nd lef size of Petit Verdot ws perhps dvntgeous in mintining enough remining lef re to limit crohydrte deficit nd susequent fruit set reduction compred to Cernet frnc. Cernet frnc ws grown with under-trellis cover crops, which hve een shown to reduce vine nitrogen sttus (Hickey et l. 2016). Though plnt nutrient nlyses were not conducted in the current study, it hs een shown tht vines grown with compnion cover crops tht did not receive supplementl nitrogen experienced reduced fruit set compred to those tht received supplementl nitrogen (Keller et l. 2001). Flower scission sensitivity to iotic stressors nd fruit set percentge cn e vriety-dependent, with greter inherent fruit set oserved with greter strch nd sucrose concentrtions in inflorescences, nd iotic stress tolernce incresed with strch presence in ovules nd nthers (Leon et l. 2004). Accounting for the stressors of pre-loom lef removl nd su-optiml loom wether, it ws possile tht Cernet frnc hd higher strch nd sucrose concentrtions while Petit Verdot hd more strch reserves in the ovules nd nthers during meiosis. Genetic predisposition for vine vigor, inherent erry numer per cluster, or cultivr-specific distriution of crohydrtes in the reproductive tissues of the flower re ll vriety-dependent fctors tht 93

110 my confound pre-loom lef removl impct on erry numer per cluster. Becuse oth roots nd leves re crohydrte source tissues during flowering nd fruit set, vriety-specific predisposition for root or rootstock for mss, s well s lef re t loom, my ply role in fruit set percent, nd, consequently, erry numer per cluster. Reduced cluster compctness due to pre-loom lef removl ws nticipted given previous reports to this effect (Plliotti et l. 2011, Poni et l nd 2008, Stini nd Howell 2010, Trdguil et l nd 2012). Though erry numer ws reduced to the gretest extent y pre-loom lef removl, rchis length lso tended to e reduced, perhps due to distl portions of clusters eing orted. However, s consequence, pre-loom lef removl often only inconsistently reduced cluster compctness. Looser clusters experience reduced unch rot incidence (Hed et l. 2009, Plliotti et l. 2011, Poni et l. 2006, Stini nd Howell 2010, Trdguil et l nd 2012), which is desirle trit prticulrly in humid regions. Pre-loom lef removl reduces erry weight (Gtti et l. 2012, Plliotti et l. 2011), ut inconsistently (Bogicevic et l. 2015, Intrieri et l. 2008, Pstore et l. 2013, Poni et l. 2008, Trdguil et l. 2012), nd often in vriety-specific mnner (Poni et l nd 2008, Stini nd Howell et l. 2010, Trdguil et l. 2010). Authors did not elorte on erry weight reduction (Gtti et l. 2012, Plliotti et l. 2011) or explined tht erry size ws not reduced y pre-loom lef removl ecuse photosynthetic rte ws incresed in remining leves, or reduced erry numer per cluster resulted in compenstory erry growth (Intrieri et l. 2008). Lterl shoots were mintined in Poni et l. (2008), nd finl lef re equled the control due to lterl growth compenstion in Sngiovese (Poni et l. 2006); in these cses, the sustined lef re deficit required to limit erry growth compenstion my not hve occurred. The results of the current study contrst those of Poni et l. (2006) tht witing until fter 94

111 flowering is etter pproch thn removing leves efore loom if the gol is to limit erry growth compenstion due to reduced erry numer per cluster. Interestingly, trends of the current study nd others (Poni et l. 2006, Trdguil et l. 2010) suggest tht pre-loom lef removl impct on erry numer per cluster my e indirectly relted to its erry size. In these cses, it ws typicl for pre-loom lef removl to hve reltively greter effect on erry numer per cluster or erry weight, ut not oth, when compring vrieties. Berry growth compenstion occurs when erry numer per cluster is reduced (Poni et l. 2008). However, the results of the current study nd others (Gtti et l. 2012, Plliotti et l. 2011, Poni et l. 2006, Trdguil et l. 2010) suggest tht erry growth my not lwys e compenstory when erry numer per cluster is reduced vi erly source tissues limittion. The ttenuted erry weight gin throughout erry growth stges I-III, prticulrly in the Petit Verdot pre-loom lef removl plots, prlleled the low source tissue tretment implemented shorty fter fruit set in Ollt nd Gudillere (1998). Berry growth reduction my hve een due to lower solute content during erry growth stge I, s well s sustined sink strength deficit (Ollt nd Gudillere 1998). Since pericrp cell division occurs during the first two weeks post-flowering (phse I) (Jon nd Bott 1988), pre-loom lef removl my hve reduced cell division. This ppered to e the cse prticulrly in Petit Verdot, given the immedite nd significnt difference in erry weight t the initil smpling. Pre-loom lef removl reduced estimted pulp weight in Petit Verdot nd Cernet frnc. If pulp weight is n estimtion of cell numer, then pre-loom lef removl my hve reduced erry cell numer. Pre-loom lef removl cn differentilly ffect the growth nd presence of erry tissues. Pre-loom lef removl reduced seed numer per erry in Sngiovese (Poni et l. 2006), ut either incresed or did not ffect seed numer per erry in Brer nd Lmrusco slmino 95

112 (Poni et l. 2008). While pre-loom lef removl did not reduce seed numer, totl nd individul seed weight ws numericlly lower in Petit Verdot in the current study; these trends were similr to wht Poni et l. (2006) reported, ut contrsted wht Poni et l. (2008) reported. Pre-loom lef removl cn increse the erry skin: pulp rtio (Poni et l. 2006), nd differentilly promote erry skin growth (Pstore et l. 2013, Poni et l nd 2010) regrdless of erry size (Poni et l. 2008). The suggested mechnism of the promotion of grpe skin growth ws due to rdition nd temperture exposure for extended periods of time (Pstore et l. 2013, Pllioti et l. 2011, Poni et l nd 2010). Estimted pulp weight ws the only erry tissue component tht ws consistently reduced y pre-loom lef removl in the current study. However, ecuse lighter erries likely hve smller surfce re nd hevier skins re likely thicker, pre-loom lef removl my hve incresed erry skin thickness, prticulrly in Petit Verdot; this ws only specultion, nd ws not sttisticlly proven. Pre-loom lef removl reduced current-seson fruitfulness y 6%, nd susequentseson fruitfulness y 21-30% (Stini nd Howell 2010, Trdguil et l. 2012). However, pre-loom lef removl did not lwys impct fruitfulness Intrieri et l. 2008, Gtti et l. 2012, Plliotti et l. 2011, Pstore et l. 2013). In the current study, pre-loom lef removl reduced cluster numer per vine in one of two yers. Vine fruitfulness, (inflorescence count per shoot numer) ws inconsistently impcted y pre-loom lef removl. The disconnect etween vine fruitfulness nd cluster numer per vine ws due to the fct tht fruitfulness dt were collected the yer fter yield components were collected, nd experimentl plots were rerrnged to limit crry-over effects of pre-loom lef removl. There were times during pre-loom lef removl implementtion when whole shoots or prts of shoots were dmged. The result ws fewer cluster numer per vine, prticulrly in Petit Verdot, which hs shoots tht re more tender thn 96

113 those of Cernet frnc. Aggressive fruit-zone lef removl increses rdition penetrtion to sl uds. The incresed rdition to the sl uds cn increse fruitfulness in the following seson (Perez nd Kliewer 1990, Snchez nd Dokoozlin 2005). This my offset the negtive impct of source tissue limittion on ud fertility, prticulrly in spur-pruned vines. Since vine crohydrte sttus during inflorescence initition (i.e. summer efore flowering) impcts vine fruitfulness (Cndolfi-Vsconcelos nd Kolet 1990), ggressive lef removl pprently did not reduce vine crohydrte sttus enough to limit fruitfulness in the current study. Pre-loom lef removl further reduced crop yield when implemented in two consecutive sesons in oth Cernet frnc nd Petit Verdot. Since inflorescence primordi re initited one yer efore flowering, erly-seson lef removl cn reduce fruitfulness nd, thus, crop yield in oth the current nd following seson (Cndolfi-Vsconcelos nd Kolet 1990, Leon et l. 2008). One of the yield components tht ws further reduced the most due to re-implementtion of pre-loom lef removl ws cluster numer per vine. However, pre-loom lef removl did not reduce vine fruitfulness in the experimentl plots used for crry-over nlysis. Thus, cluster numer reduction ws either due to physicl dmge incurred when removing leves on young, tender shoots, or due to whole-cluster ortion. Berry numer per cluster nd cluster weight were the other yield components tht were further reduced, suggesting tht repeted lef preloom lef removl cn continue to reduce erries per cluster nd cluster weights (Gtti et l. 2012, Plliotti et l. 2011) s function of further reduced fruit set (Stini nd Howell 2010). Crohydrte reserves cn e depressed due to source tissue reduction during the previousseson s fruit ripening period (Cndolfi-Vsconcelos nd Kolet 1990), ut lso due to excessive crop lods, s well s poor photosynthesis conditions prior to current-seson flowering. Thus, pruning weight in the current-seson plots my hve een reduced due to source tissue limittion 97

114 efore loom, s well s the continul depletion of storge reserves in permnent vine prts (Plliotti et l. 2011). Since reduced crohydrte vilility is prtilly responsile for fruit set reduction (Cndolfi-Vsconcelos nd Kolet 1990, Cspri nd Lng 1996, Coome 1962), nd prt of the loom nd fruit set period is supported y moilized crohydrtes (Leon et l. 2008), recurring pre-loom lef removl my hsten the depletion of storge reserves, nd, consequently, reduce vine size nd crop yield potentil t n typiclly erly vine ge. Primry fruit chemistry: Previous studies hve frequently shown tht pre-loom lef removl incresed solule solids concentrtion (Digo et l. 2012, Gtti et l. 2012, Intrieri et l. 2008, Plliotti et l. 2011, Pstore et l. 2013, Poni et l. 2006, 2008, nd 2010, Stini nd Howell 2010, Trdguil et l. 2012). Solule solids were incresed y greter lef re: fruit weight rtios (Gtti et l. 2012, Pllioti et l. 2011, Trdguil et l nd 2012), nd restored nd more photosyntheticlly efficient lef re in the post-verison period, resulting hstened erry sugring (Intrieri et l. 2008, Pllioti et l. 2011, Pstore et l. 2013, Poni et l. 2006). Less solute solule solids re required to rech given solule solids concentrtion in reltively smller erries (Ollt nd Gudillere 1998). This my hve een why there ws more consistent reduction in solule solids due to post-fruit set removl of leves to the gretest extent in Petit Verdot in the current study. While neither lef photosynthetic rtes nor lef re ws mesured in the current study, similr lef re mount ws removed etween pre-loom lef removl nd post fruit set lef removl to gretest extent. Though necdotl, Cernet frnc cnopies were sprser nd shoots ppered to grow t slower rte compred to Petit Verdot. Thus, it ws possile tht lef removl did not reduce solule solids in Cernet frnc due to reltively limited vegettive sink competition nd/or incresed exposed lef re. Even though remining leves from prtilly defolited vines re often more efficient cron ssimiltors 98

115 (Cndolfi Vsconcelos nd Kolet 1990, Buttrose 1966, Kliewer et l. 1970, Poni et l. 2006), it ws possile tht lef re in Petit Verdot ws mintined t deficit tht ws gret enough to limit solule solids; this my hve een the cse prticulrly in 2013, the yer tht crop lods were reltively gretest. Erly lef removl cn dely the onset of ripening compred to lter defolition due to lef re deficit nd reduction in sink strength nd ssimilte trnsloction to fruit (Ollt nd Gudillere 1998, Pstore et l. 2013). A low sink strength my hve een especilly responsile for low solule solids concentrtions in Petit Verdot in 2014, when crop lods were out one third less thn crop lods were in Most pre-loom lef removl studies were conducted in drier regions. Thus, it ws ssumed tht post-verison wether ws more conducive to fruit ripening in those regions compred to the frequently cool/cloudy wether in Virgini. Growing degree dy sum, nd verge monthly PAR, were lower in the post- compred to pre-verison periods of 2013 nd It ws cooler, nd more rin fell, in 2014 compred to 2013, nd these ptterns were mintined in the post-verison period. Additionlly, June 2014 hil storm resulted in some lef dmge. These fctors, comined with reltively lower lef re, my hve resulted in reltively greter ttenution in solule solids ccumultion in the ggressive lef removl tretments in the 2014 seson compred to in the 2013 seson. Fruit-zone lef removl often reduces titrtle cidity (TA) which cn e ttriuted to temperture-driven mlic cid respirtion (Jckson nd Lomrd 1993, Kliewer nd Schultz 1964). Aggressive lef removl tended to reduce titrtle cidity in Cernet frnc nd Petit Verdot, leit inconsistently. The reltively greter reduction in titrtle cidity experienced with pre-loom nd post-fruit set removl of leves to the gretest extent in 2013 my hve een due to the generl wrmer temperture in 2013 when compred to Previous studies lso 99

116 found tht pre-loom lef removl inconsistently ffected (Gtti et l. 2012, Intrieri et l. 2008, Plliotti et l. 2011, Poni et l. 2006, Trdguil et l. 2010) TA. Two studies in which preloom lef removl hd no effect on titrtle cidity lso noted the re-growth of lterls from sl nodes (Intrieri et l. 2008, Plliotti et l. 2011); in these cses, fruit-zone shding my hve limited fruit temperture. Similrly, in the current study, the higher vigor Petit Verdot cnopies my hve locked more incident rdition, resulting in less mlic cid respirtion. Pre-loom lef removl s vrile impct on titrtle cidity ws suggested (Poni et l. 2006) to e due to greter cron ccumultion into trtric cid in sun versus shde erries (Kliewer nd Schultz 1964). The increse in titrtle cidity in Petit Verdot pre-loom lef removl plots ws n unexpected response in Trtric cid reduction during ripening is dilution effect trtrte levels don t chnge s erries gin mss (Johnson nd Crroll 1973). Pre-loom lef removl considerly reduced Petit Verdot erry weight in oth 2013 nd Thus, TA levels my not hve een different (2013), nd even incresed (2014) ecuse cids in grpes from the pre-loom lef removl tretment were not diluted s much s they were in the grpes from other tretments. Totl phenolics nd nthocynins: Pre-loom lef removl incresed totl grpe phenolics in Petit Verdot, nd less consistently in Cernet frnc in this study. Open fruit zones hve een shown to increse grpe phenolics (Jckson nd Lomrd 1993, Price et l. 1995, Smrt nd Roinson 1991). However, totl phenolics were only incresed y pre-loom lef removl in the current study, which suggested tht lef removl timing ws importnt for this response. There re severl suclsses of grpe phenolics tht ccumulte t different periods of grpe development, nd re differentilly ffected y temperture nd rdition (Downey et l. 2006). When these suclsses re comined into totl phenolics, it is difficult to determine if specific 100

117 phenolic compounds were incresed from longer exposure to incresed rdition nd erry tempertures. The physicl chnges incurred y erries, nmely smller size nd lower estimted pulp weight, ws the most likely mechnism ehind the increse in totl phenolics oserved in grpes from the pre-loom lef-removl plots. Pre-loom lef removl only incresed totl phenolics in Cernet frnc in the sme yer tht erry weight ws lso reduced. Further, the negtive reltionship tht existed etween erry weight nd totl phenolics ws (1) significnt in every instnce tht pre-loom lef removl incresed totl phenolics in oth vrieties, nd (2) lwys greter in the Petit Verdot, the vriety tht hd lower erry weight. Differentil erry tissue growth could not e ruled s nother fctor tht incresed totl phenolics, s trends suggested tht there my hve een reltively thicker skins in the erries from pre-loom lef removl plots compred to other tretments. In the current study, ggressive fruit-zone lef removl did not ffect totl grpe nthocynins in either Cernet frnc or Petit Verdot, regrdless if implemented efore loom or fter fruit set. Thus, totl nthocynins were not incresed with the lower crop lods, smller erries, nd open fruit-zone microclimtes ssocited with pre-loom lef removl plots. Since nthocynins re one compound clss of the rod suclss of phenolics clled flvonoids, it ws possile tht they were not in high enough concentrtions in skins to e further concentrted y smller erry weight nd, thus, differences were not detected with the quntifiction methods used in the current study. Anthocynins tend to e reduced t grpe tempertures of C (Downey et l. 2006, Spyd et l. 2002, Trr et l. 2008, Ymne et l. 2006). However, grpe nthocynins were incresed in pre-loom compred to verison-time lef removl plots, even when erries spent ~300 hrs. 30 C in oth plots during the post-version period (Pstore et l. 2013). Pre-loom lef removl cn thus increse nthocynins y dditionl mechnisms thn 101

118 the microclimte creted y lef removl, such s the promotion of thicker erry skins (Pllioti et l. 2011, Poni et l. 2008), or greter lef re: fruit weight rtios (Gtti et l. 2012, Intrieri et l. 2008). While nthocynins were decresed when shded erries were rtificilly heted to detrimentl tempertures for reltively shorter periods thn exposed erries (142 hrs. 30 C nd 15 hrs. 35 C) (Trr et l. 2008), shded fruit is cooler thn exposed fruit, nd is typiclly ner mient ir temperture in field setting (Spyd et l. 2002, nd results presented herein). Thus, unless post-verison wether is chrcterized y consistent mient ir tempertures C, shded fruit my not e heted to tempertures tht re detrimentl to nthocynin ccumultion. In the experimentl vineyrd in which the current study ws conducted, mient ir temperture ws never 35 C, nd 30 C for only 39 nd 16 hrs. during the entire post verison periods of 2013 nd 2014, respectively. Given the grpe temperture trends tht were presented erlier in the discussion, it ws unlikely tht grpe nthocynins were limited y tempertures in the current study, even in the most exposed fruit-zones. The generl lef removl prctice in Virgini is to im for one to two fruit-zone lef lyers (Wolf 2008). This is likely ecuse of the documented cses tht nthocynins were reduced with ggressive lef removl. These populr studies were conducted in regions tht re wrmer, drier, nd sunnier compred to the humid growing regions in the estern US (Bergqvist et l. 2001, Spyd et l. 2002, Trr et l. 2008). Tht grpe nthocynins were not reduced fruit-zones with ~ lef lyers ws considered positive response, s fungl disese incidence is generlly improved with more open fruit-zones (English et l. 1989, Wolf et l. 1986). In the current study, ggressive lef removl resulted in fruit-zones chrcterized y moderte est- nd west-side erry tempertures nd incresed rdition penetrtion, eneficil or t lest not detrimentl, comintion for grpe nthocynins (Trr et l. 2008). Therefore, 102

119 sl lef removl in red-fruited vrieties need not e s conservtive s currently prcticed in mny estern US growing regions, nd need not e focused on the est cnopy side, prticulrly if lef removl is executed efore or reltively soon fter fruit set in order to reduce sunurn incidence (Pstore et l. 2013). Grpe crotenoids: Crotenoids tended to e greter in 2014, the reltively cooler yer, compred to 2013, the reltively wrmer yer, nd this ws more consistent in Petit Verdot thn in Cernet frnc. This result contrsts tht incresed crotenoid synthesis occurs in reltively wrmer compred to cooler growing sesons (Frin et l. 2010). It is possile tht crotenoid degrdtion ws slower in the cooler post-verison period of 2014 compred to 2013, s incresed temperture incresed post-verison crotenoid degrdtion (Bureu et l. 1998). Though temperture cn pprently influence crotenoid synthesis nd degrdtion ptterns (Bureu et l. 1998, Frin et l. 2010), seprting the effect of rdition nd temperture is difficult in field setting nd, thus, it ws unknown if temperture plys greter role in crotenoid synthesis or degrdtion, or if there re temperture thresholds tht re prticulrly conducive or detrimentl to crotenoid ccumultion. Seprting the effects of light nd temperture ws eyond the scope of this field study, ut it ws generlly wrmer, nd erries spent more time 30 C in the post-verison period of 2013 compred to Thus, other fctors equl, it is possile tht crotenoid degrdtion increses t tempertures 30 C. Nonetheless, Petit Verdot erry weight ws lower in 2014 compred to Thus, the consistently greter crotenoid concentrtions in Petit Verdot in 2014 my hve simply een function of incresed concentrtion in smller erries. When verged cross ll dtes within seson, crotenoid differences were more consistently different etween dtes compred to etween lef removl tretments, nd were 103

120 typiclly greter t ll pre-hrvest smple dtes compred to t hrvest. The trends of the current study re consistent with others reports tht crotenoids re highest erly in erry formtion nd then progressively decrese s erries develop, with shrp decrese seen t or fter verison (Rzungles et l. 1988). The within-dte increse in crotenoid concentrtions in lef removl compred to no lef removl plots reveled tht sesonl increses oserved in grpe crotenoids from lef removl plots were due to increses in crotenoids during the pre-verison period. Other studies lso reported tht crotenoid increses in well-exposed grpes occurs efore verison (Bureu et l nd ) nd tht differences re ttenuted therefter, or tht crotenoids were even greter when in shded conditions during post-verison period (Bureu et l nd, Rzungles et l. 1998). In the current study, lutein nd β-crotene were inconsistently greter in reltively shded fruit-zones t hrvest, consistent with other reports tht these two mjor grpe crotenoids were greter t mturity when grpes were shded (Bureu et l nd ). Thus, it is possile tht these two mjor crotenoids re prticulrly sensitive to shde in the post-verison period when compred to minor crotenoid clsses. When there were differences in crotenoids etween lef removl tretments t verison, they were typiclly oserved s incresed zexnthin levels in ggressive post-fruit set nd preloom lef removl plots, or s incresed crotenoids in pre-loom lef removl plots. Zexnthin my hve een the crotenoid tht ws most consistently incresed y lef removl t verison ecuse its ptterns differ thn those of the min grpe crotenoids, lutein nd β- crotene (Rzungles et l. 1988, 1998). De-epoxidized xnthophylls, such s zexnthin, re fvored y high light intensity wheres lutein nd β-crotene re degrded y high light intensity (Bureu et l nd, Rzungles et l. 1998). Incresed crotenoid concentrtions t verison ws minly oserved in pre-loom lef removl plots; this ws necdotlly oserved s 104

121 dely in grpe colortion in these plots, prticulrly in As mentioned ove regrding the lower solule solids concentrtions oserved in Petit Verdot plots tht hd the most sl leves removed, the ttenuted grpe colortion my hve een due to concomitnt decrese in source tissues nd sink strength during ripening in pre-loom lef removl plots. As such, crohydrte supply to ripening fruit my hve een limited in these plots. A decrese in crotenoids s grpes mture my e due to competition etween terpenoids (i.e. crotenoids) nd phenolics (i.e. flvonoids) for common precursors, cette nd mevlonte (Rzungles et l. 1988). If so, then there my not hve een s much competition for these precursors in the preloom lef removl plots during verison ecuse colortion ws delyed. Therefore, the synthesis of phenolic compounds such s nthocynins were not in competition for common precursor compounds. Bsl lef removl effect on grpe crotenoids in the current study re consistent with findings of others tht fruit exposure cn enhnce the pre-verison ccumultion, nd postverison rekdown, of crotenoids (Bureu et l nd, Rzungles et l. 1998). These responses were nticipted, s light is necessry for crotenoid iosynthesis, nd the ctivity of phytoene synthse, n enzyme gene responsile for the first committed step in the crotenoid iosynthetic pthwy, is stimulted y light (Hirscherg et l nd those cited within). Further, fruit exposure enhnces pre-verison ccumultion nd post-verison rekdown of grpe crotenoids (Rzungles et l. 1998). The difference in pek nd hrvest crotenoids were correlted with C13-norisoprenoids in wines the greter the mgnitude of degrdtion in grpe crotenoids, the greter the mount of C13-norisoprenoids in wine (Crupi et l. 2010). Thus, in the current study, fruit-zones ssocited with greter fruit-zone rdition penetrtion potentilly 105

122 hd greter C13-norisoprenoid concentrtions in wines, s there tended to e greter degrdtion in crotenoids from pre-verison to hrvest when grpes were most exposed to light. The four crotenoids nlyzed in the current study were, however, differentilly influenced y sl lef removl zexnthin ws most consistently incresed nd degrded y fruit exposure, while lutein 5,6-epoxide nd lutein were ffected y lef removl prctice more consistently thn β-crotene ws. In Syrh study, shde tended to increse epoxified xnthophyll forms (Bureu et l. 1998), which supported tht high light intensity cuses deepoxidtion of violxnthin to zexnthin while shde cuses epoxidtion of zexnthin to violxnthin vi ntherxthin (Bumes et l. 2002). While results of the current study cnnot disprove tht zexnthin nd lutein experience similr ptterns during ripening (Crupi et l. 2010), trends showed tht zexnthin is much more influenced y fruit exposure thn lutein. This my e function of the different physiologicl roles these two crotenoids ply: β-crotene nd lutein re minly involved in light hrvesting, while xnthophylls re involved in photoprotection, nd id in plnt tissue dpttion to different light conditions (Hirscherg 2001). Nonetheless, xnthophylls increse under strong light nd the rtio etween lutein nd the xnthophylls decreses, nd the mrna levels of enzyme genes responsile for the flux of crotenoid synthesis down the xnthophyll side of the iosynthetic pthwy were incresed to greter extent thn the lutein side when plnts were shifted from low light to strong light (Hirscherg et l nd those cited within). The results of the current study support these findings, s the zexnthin: lutein rtios were consistently greter in well-exposed compred to shded Petit Verdot nd Cernet frnc grpes t every smple dte when verged cross ll three yers. 106

123 Studies conducted in more Mediterrnen-type climtes oserved tht lutein nd β- crotene synthesis nd degrdtion were incresed in sun-exposed reltive to shded grpes (Bureu et l nd, Rzungles et l. 1998). The only study known to look t fruit exposure effects on crotenoids in region nlogous to the current study found tht zexnthin ws incresed with sl lef removl compred to shded control, while lutein nd β-crotene were less consistently ffected (Kwsniewski et l. 2010). Thus, in humid regions, it is possile tht the required rdition nd temperture thresholds needed to differentilly impct lutein nd β-crotene re not consistently experienced. Given the physiologicl role of zexnthin s n importnt photoprotector of photosystem II, where it quenches triplet stte chlorophyll nd reduces the likelihood of rective singlet oxygen (Dll Osto et l. 2012), it is likely tht zexnnthin is the most responsive crotenoid to chnges in light conditions. Similr to flvonols eing good phenolic proxy for grpe exposure to light (Downey et l. 2006), zexnthin my e the est crotenoid indictor for grpe exposure to light cross wide rnge of climtes. It hs een shown tht crotenoid concentrtions cn gretly vry cross vrieties (Crupi et l. 2010), nd tht fruit exposure cn differentilly influence crotenoid degrdtion ptterns etween white (Musct of Frontignn) compred to red (Syrh) vriety (Bureu et l nd ). In the current study, incresed grpe crotenoid concentrtions, prticulrly lutein nd β- crotene, were oserved in Petit Verdot when compred to Cernet frnc, nd, in 2014, on the west compred to est side in Cernet frnc. Since crotenoids re found in grpe skins t levels two to three times greter thn in the pulp (Rzungles et l. 1988), the increse in crotenoid concentrtions in oth of these cses ws speculted to e function of the smller erries. The inherent smller erry size in Petit Verdot my hve een why there were reltively 107

124 more tretment*dte interctions oserved in Petit Verdot reltive to Cernet frnc cross ll sesons. The reltively smll difference in crotenoid concentrtion tht ws needed in order to determine tretment effect over time my hve een ttenuted y the dilution effect of reltively greter pulp mounts in Cernet frnc erries. Though concentrtions of crotenoids chnged in Cernet frnc erry skins over the seson, they chnged less consistently compred to Petit Verdot in 2013, the yer tht erry weight ws unffected y lef removl tretment in Cernet frnc. Further, while lutein 5,6-epoxide ws inconsistently detected in oth Petit Verdot nd Cernet frnc erry smples t hrvest, zexnthin ws lso inconsistently detected in Cernet frnc erry smples t hrvest. Agin, this ws puttively due to the dilution effect of the reltively lrger erry size nd reltive pulp mount in Cernet frnc. Since lutein nd β-crotene were in greter concentrtions prticulrly in Petit Verdot, nd β-crotene nd lutein constitute lmost 85% of the totl crotenoids in mture grpes (Mendes-Pinto 2002), the C13-norisoprenoids derived from β-crotene nd lutein my e more directly involved in wine rom thn others (Winterhlter nd Rouseff 2002). From prcticl winemking stndpoint, the reltively smll erry size in Petit Verdot my hve resulted in n incresed concentrtion of β-ionone in wine musts, s this norisoprenoid hs een identified s product of oth lutein nd β-crotene (Isoe et l. 1969; Knswud nd Crouzet 1990; Eugster nd Mrki-Fischer 1991; Mris 1992). Since β-ionone hs sensory descriptors of violet, woody, nd rsperry (Winterhlter nd Rouseff 2002), n increse in its concentrtion my result in detectle increses in fruity nd florl roms in wine. These fetures my e the reson tht regionl winemking trends hve een towrds decresing the mount of Petit Verdot used in Bordeux lends, possily to chieve wine chrcterized less y fruity nd florl notes, nd more ligned with clssic Bordeux chrcteristics such s erthy, lethery, nd tnnic. 108

125 Wine sensory: Cernet frnc wine color ws distinguished etween the most ggressive postfruit set lef removl tretment nd the no lef removl tretment. Further, wines mde from ggressive lef removl tretments received higher intensity rtings for red color in oth vrieties, nd for stringency in Cernet frnc. Wine phenolics, nthocynins, nd color density tend to e incresed with sun exposure (Di Profio et l. 2011, Ristic et l. 2007, Smrt nd Roinson 1991, Stff et l. 1997, Verzer et l. 2016). Accordingly, sl lef removl incresed Cernet frnc wine color nd improved pltility (Stff et l. 2007) nd Cernet frnc nd Merlot wine viscosity nd length (Di Profio et l. 2011). While tnnins were not mesured in the current study, they re the lest susceptile flvonoid to light exposure (Downey et l. 2006). Shding reduced tnnins in grpes nd wine, which lso hd lower stringency rting (Ristic et l. 2007). Thus, exposed Cernet frnc grpes my hve hd greter tnnin concentrtion nd, consequently, greter wine stringency, s this ws sensory ttriute tht hs een positively correlted with wine tnnins (Csss et l. 2013). The contriution of nthocynins to red wine color is complex, nd is function of oth monomeric forms nd copigmenttion (He et l. 2012). Monomeric nthocynins re lrgely responsile for the color of young red wines (He et l. 2012). While monomeric nthocynin sornce t 520 nm is reduced s ph increses from one to five (Crit et l. 2000), the ph rnge ws low, never exceeding more thn 0.09-unit difference etween tretment wines of the current study, regrdless of yer or vriety. Thus, it ws unlikely tht color intensity ws ffected y the reltionship of wine ph nd monomeric nthocynins. Totl grpe nthocynins were not ffected y lef removl tretment, ut pre-loom lef removl incresed grpe phenolics nd reduced erry size in Petit Verdot in The result my hve een net increse in phenolic nd nthocynin concentrtion in pre-loom lef removl wine musts, s smller erries cn 109

126 result in greter color sturtion in wines, likely due to the concentrtion of skin nthocynins (Csss et l. 2015). Thus, red wine color intensity my hve een function of nthocynin copigmenttionn with other phenolic compounds, s copigmenttion cn e responsile for 30-50% of the color of young red wines, nd the principl cofctors in young red wines re flvonoid nd non-flvonoid phenolics (He et l. 2012). Bsl lef removl improved Cernet frnc wine rom (Stff et l. 2007), nd incresed lck pepper nd lck fruit rom nd decresed green chrcteristic in Cernet frnc, possily due to incresed norisoprenoids nd concomitnt reduction in methoxypyrzines, s shown y Di Profio et l. (2011). In the current study, consumers did not prefer the sensory ttriutes of ny Cernet frnc or Petit Verdot tretment wines over nother. However, Cernet frnc nd Petit Verdot wine rom ws distinguished etween lef removl tretments of different mgnitude. While rom compounds were not quntified in grpes or wines, rom compounds cn e incresed in well-exposed fruit-zones, or decresed in shded fruit-zones (Lee et l. 2007, Mris et l. 1992, Rzungles et l. 1998, Ristic et l. 2007, Verzer et l. 2016). Grpe crotenoids re precursors to norisoprenoids, clss of rom compounds tht re importnt odornts in young, mono-vrietl red wines (Ferreir et l. 2000, Pineu et l. 2007), nd prticulrly for Cernet frnc (Fn et l. 2010). Post-fruit set lef removl to the medium extent hd greter zexnthin levels thn no lef removl in the sme seson tht wine rom ws significntly distinguished etween these two tretments. Zexnthin-derived norisoprenoids include 3-hydroxy-7,8-dihydro-β-ionol nd 3-hydroxy-β-ionone (Mthieu et l. 2005, Crupi et l. 2010), nd zexnthin is potentil in vivo precursor of TDN (Kwsniewski et l. 2010). These norisoprenoids my hve resulted in the ility of pnelists to detect rom difference etween these two tretment wines. Though crotenoid ptterns suggested tht there ws greter romtic 110

127 potentil in wines mde with fruit from ggressive lef removl plots in 2013, there were no significnt difference in consumer preference of rom from wines mde from well-exposed nd well-shded tretments. Thus, the concentrtion of mesured (i.e. crotenoids) nd/or unmesured (methoxypyrzines, norisoprenoids) rom impct compounds were not ffected enough y lef removl prctice to result in difference in wine rom preference. It ws speculted to hve een difficult for pnelists to determine wine preference without hving point of reference for comprison, s wines were served wines mondiclly, yet there ws little difference in tretment preferences when wines were rnked side y side. Conclusion: Best fruit-zone mngement prctices were historiclly ggressive to improve fruit composition nd disese control, ut then ecme more conservtive to limit sunurn nd nthocynin reduction. Comprle to pest control progrm, or other culturl prctices, this report confirms tht fruit-zone lef removl should e region-specific prctice. Aggressive lef removl modestly improved fruit composition nd wine color intensity, nd hs potentil to reduce fungl disese incidence nd, if implemented reltively close to fruit set, reduce sunurn. While very ggressive lef removl efore loom cn result in uneconomicl reduction in crop yield, it is prole tht similr fruit composition is ttinle without drstic crop yield reduction if either fewer leves re removed efore loom, or just s mny leves re removed immeditely fter fruit set. Crotenoids incresed nd decresed to greter extent when leves were removed from fruit-zones, nd zexnthin ppered to e the crotenoid tht ws most susceptile to chnges in rdition nd temperture. Future work should focus on removing fewer leves efore loom or immeditely fter fruit set in red nd white vrieties in order to evlute if fruit composition cn e mintined or improved without drstic crop yield reduction cross vrieties used for different enologicl gols. 111

128 Literture cited Auldridge, M.E., D.R. McCrthy, nd H.J. Klee Plnt crotenoid clevge oxygenses nd their pocrotenoid products. Curr. Opin. Plnt. Biol. 9: Bumes, R., J. Wirth, S. Bureu, Y. Gunt, nd A. Rzungles Biogenertion of C13-norisoprenoid compounds: experiments supportive for n po-crotenoid pthwy in grpevines. Anlytic Chimic Act 458: Bledsoe, A.M., W.M. Kliewer, nd J.J. Mrois Effects of timing nd severity of lef removl on yield nd fruit composition of Suvignon lnc grpevines. Am. J. Enol. Vitic. 1: Bond M., nd V.O. Sdrs Review: criticl pprisl of methods to investigte the effect of temperture on grpevine erry composition. Aus. J. Grpe Wine Res. 21: Bogicevic, M., V. Mrs, M. Mugos, V. Kodzulovic, J. Ricevic, S. Sucur, nd O. Fill The effect of erly lef removl nd cluster thinning tretments on erry growth nd grpe composition in cultivrs Vrnc nd Cernet Suvignon. Chem Biol Techn Agric 2: 1-8. Bergqvist, J., N.K. Dokoozlin, nd N. Eisud Sunlight exposure nd temperture effects on erry growth nd composition of Cernet Suvignon nd Grenche in the centrl Sn Joquin Vlley of Cliforni. Am. J. Enol. Vitic. 52: 1-7. Brvdo, B., Y. Hepner, C. Loinger, S. Cohen, nd H. Tcmn Effect of crop level nd crop lod on growth, yield, must nd wine composition, nd qulity of Cernet Suvignon. Am. J. Enol. Vitic. 36: Bureu, S.M., A.J. Rzungles, R.L. Bumes, nd C.L. Byonove Effect of vine or unch shding on the crotenoid composition in Vitis vinifer L. erries I. Syrh grpes. Vitic. Enol. Sci. 53: Bureu, S.M., A.J. Rzungles, R.L. Bumes, nd C.L. Byonove Effect of vine or unch shding on the crotenoid composition in Vitis vinifer L. erries II. Musct of Frontignn grpes. Vitic. Enol. Sci. 53: Bureu, S.M., R.L. Bumes, nd A.J. Rzungles Effects of vine or unch shding on the glycosylted flvor precursors in grpes of Vitis vinifer L. Cv. Syrh. J. Agric. Food Chem. 48: Bureu, S.M., A.J. Rzungles, nd R.L. Bumes The rom of Musct of Frontignn grpes: effect of the light environment of vine or unch on voltiles nd glycoconjugtes. J. Sci. Food Agric. 80: Buttrose, M.S The effect of reducing lef re on the growth of roots, stems, nd erries of Gordo grpe vine. Vitis 5: Crit, L., T. Fossen, nd O.M. Andersen Colour stility of the six common nthocynidin 3-glucosides in queous solutions. Food Chem. 68: Cndolfi-Vsconcelos, M.C., nd W. Kolet Yield, fruit qulity, ud fertility nd strch reserves of the wood s function of lef removl in Vitis vinifer Evidence of compenstion nd stress recovering. Vitis 29: Cndolfi-Vsconcelos, M.C., nd W. Kolet Influences of prtil defolition on gs exchnge prmeters nd chlorophyll content of field-grown grpevines: mechnisms nd limittion of compenstion cpcity. Vitis 30:

129 Cronneu, A Trellising nd cnopy mngement for cool climte viticulture. Proc. First Int. Symp. Cool Climte Viticulture nd Enology, June, Eugene, OR, Oregon Stte Univsersity, Cspri, H.W., nd A. Lng Crohydrte supply limits fruit-set in commercil Suvignon lnc grpes. In Proceedings of the 4 th Interntionl Symposium on Cool Climte Enology nd Viticulture. T. Henick-Kling et l. (Eds.), pp. II New York Stte Agriculture Experiment Sttion, Genev, NY. Cspri, H., A. Lng, nd P. Alspch Effects of girdling nd lef removl on fruit set nd vegettive growth in grpe. Am. J. Enol. Vitic. 49: Csss, L.F., Lrsen, R.C., Bever, C.W., Mireles, M.S., Keller, M., Riley, W.R., Smithymn, nd R., nd J.F. Hrertson. Sensory impct of extended mcertion nd Regulted Deficit Irrigtion on Wshington stte Cernet Suvignon wines. Am. J. Enol. Vitic. 2013, 64, Csss, L.F., M. Keller, nd J.F. Hrertson Regulted deficit irrigtion lters nthocynins, tnnins nd sensory properties of Cernet Suvignon grpes nd wines. Molecules 20: Chorti, E., S. Guidoni, A. Ferrndino, nd V. Novello Effect of different cluster sunlight exposure levels on ripening nd nthocynin ccumultion in Neiolo grpes. 61: Col, G., O. Fill, nd L. Mrini BerryTone A simultion model for the dily course of grpe erry temperture. Agric. For. Meteorol. 149: Coome, B.G The effect of removing leves, flowers, nd shoot tips on fruit-set in Vitis vinifer L. J. Hortic. Sci. 37: Crupi, P., A. Colett, A., nd A. Antoncci Anlysis of crotenoids in grpes to predict norisoprenoid vrietl rom of wines from Apuli. J. Agric. Food Chem. 8: Dll Osto, L., N.E. Holt, S. Kligotl, M. Fucimn, S. Czznig1, D. Croner, H. A. Frnk, J. Alric, nd R. Bssi Zexnthin plnt photosynthesis y modulting chlorophyll triplet yield in specific light-hrvesting ntenn suunits. J. Biol. Chem. Di Profio, F., A.G. Reynolds, nd A. Ksimos Cnopy mngement nd enzyme impcts on Merlot, Cernet frnc, nd Cernet Suvignon. II. Wine composition nd qulity. Am. J. Enol. Vitic. 62: Digo, M., B. Ayestrn, Z. Gudlupe, S. Poni, nd J. Trdguil Impct of preloom nd fruit set sl lef removl on the flvonol nd nthocynin composition of Temprnillo grpes. Am. J. Enol. Vitic. 63: Dokoozlin, N.K., nd W.M. Kliewer The light environment within grpevine cnopies. II. Influence of lef re density on fruit zone light environment nd some cnopy ssessment prmeters. Am. J. Enol. Vitic. 46: Downey, M.O., N.K. Dokoozlin, nd M.P. Krstic, M.P Culturl prctice nd environmentl impcts on flvonoid composition of grpes nd wine: review of recent reserch. Am. J. Enol. Vitic. 57: Dry, P., nd B. Coome, eds Viticulture 1 Resources, 2 nd Ed. Revised version of grpevine growth stges The modified E-L system. Winetitles, Adelide, Austrli English, J.T., C.S. Thoms, J.J. Mrois, nd W.D. Guler Microclimtes of grpevine cnopies ssocited with lef removl nd control of Botrytis unch rot. Phytopthology 79:

130 Escudero, A., E. Cmpo, L. Frin, J. Ccho, nd V. Ferreir Anlyticl chrcteriztion of the rom of five premium red wines. Insights into the role of odor fmilies nd the concept of fruitiness of wines. J. Agric. Food Chem. 53: Eugster, C.H., nd E. Mrki-Fishcer The chemistry of rose pigments. Angew Chem. Ind. 30: Fn, W., Y. Xu, W. Jing, nd J. Li Identifiction nd quntifiction of impct rom compounds in 4 nonflorl Vitis vinifer vrieties grpes. J. Food Sci. 75: Frin, L., F. Crru, E. Boido, E. Disegn, nd E. Dellcss Crotenoid evolution in Vitis vinifer cv. Tnnt Grpes during Ripening. Am. J. Enol. Vitic. 61: Ferreir, V., R. Lopez, nd J.F. Ccho Quntittive determintion of the odornts of young red wines from different grpe vrieties. J. Sci. Food Agric. 80: Gtti, M., F. Bernizzoni, S. Civrdi, nd S. Poni Effects of cluster thinning nd preflowering lef removl on growth nd grpe composition in cv. Sngiovese. Am. J. Enol. Vitic. 63: Goodwin, T.W In Chemistry nd iochemistry of plnt pigments, 2 nd ed.; Goodwin, T.W. Ed; Acdemic Press: New York; Vol. 1, pp Goodwin, T.W The Biochemistry of Crotenoids, Volume 1: Plnts. Chpmn nd Hll, New York, NY. He, F., N-N. Ling, L. Mu, Q-H. Pn, J. Wng, M.J. Reeves, nd C-Q. Dun. Anthocynins nd their vrition in red wines I. Monomeric nthocynins nd their color expression. Molecules 17: Hed, B Reltionship etween cluster compctness nd unch rot in Vignoles grpes. Plnt Dis. 93: Hickey, C.C., J.S. Stlling, nd T.K. Wolf Under-trellis cover crop nd rootstock lter growth, components of yield, nd fruit composition of Cernet Suvignon. Am. J. Enol. Vitic. 67:. Hirscherg, J Crotenoid iosynthesis in flowering plnts. Curr. Opin. Plnt Biol. 4: Hunter, J., O.T. de Villiers, nd J.E. Wtts The effect of prtil defolition on qulity chrcteristics of Vitis vinifer L. cv. Cernet Suvignon grpes II. Skin colour, skin sugr nd wine qulity. Am. J. Enol. Vitic. 42: Intrieri, C, I. Filippetti, G. Allegro, M. Centinri, nd S. Poni Erly defolition (hnd vs mechnicl) for improved crop control nd grpe composition in Sngiovese (Vitis vinifer L.). Aus. J. Grpe Wine Res. 14: Isoe, S., S.B. Hyeon, nd T. Skn Photo-oxygention of crotenoids. I. The formtion of dihydroctinidiolide nd β-ionone from β-crotene. Tetrhedron Lett. 4: Jckson, D.I., nd P.B. Lomrd Environmentl nd mngement prctices ffecting grpe composition nd wine qulity A Review. Am. J. Enol. Vitic. 44: Jen, D., nd L. Lpointe Limited crohydrte vilility s potentil cuse of fruit ortion in Ruus chmemorus. Physiol. Plnt 112: Johnson, L.A., nd D.E. Crroll Orgnic cid nd sugr contents of scuppernong grpes during ripening. J. Food. Sci 38: Jon, R., nd R. Bott Fruit set nd erly erry development in two grpevine cultivrs. Isrel J. Bot. 307:

131 Knswud, P. nd J.C. Crouzet Mechnism of formtion of voltile compounds y therml degrdtion of crotenoids in queous medium. 1. β-crotene degrdtion. J. Agric. Food Chem. 38: Keller, M., M. Kummer, nd M.c. Vsconcelos Reproductive growth of grpevines in response to nitrogen suplly nd rootstock. Aust. J. Grpe Wine Res. 7: Keller, M The Science of Grpevines: Antomy nd Physiology, 1 st Ed. Acdemic Press. Kliewer, W.M., nd H.B. Schultz Influence of environment on metolism of orgnic cids nd crohydrtes in Vitis vinifer. II. Light. Am. J. Enol. Vitic Kliewer, W.M Effect of time nd severity of defolition on growth nd composition of Thompson seedless grpes. Am. J. Enol. Vitic. 21: Kolet, W Berry set of grpe vines relted to shoot tretments nd climtic fctors. Wein Wiss 21: Kolet, W. M.C. Cndolfi-Vsconcelos, E, Aeschimnn, nd G.S. Howell Influence of defolition, rootstock, nd trining system on Pinot noir grpevines. I. Moilistion nd reccumultionn of ssimiltes in woody tissue. Wein Wiss. 48: Kotseridis, Y., R.L. Bumes, A. Bertrnd, nd G.K. Skouroumounis Quntittive determintion of β-ionone in red wines nd grpes of Bordeux using stle isotope dilution ssy. J. Chromtogr. A. 848: Kotseridis, Y., A. Georgidou, P. Tikos, S. Kllithrk, nd S. Koundours Effects of severity of post-flowering lef removl on erry growth nd composition of three red Vitis vinifer L. cultivrs grown under semirid conditions. J. Agric. Food Chem. 60: Kwsniewski, M.T., J.E. Vnden-Heuvel, B.S. Pn, nd G.L. Scks Timing of cluster light environment mnipultion during grpe development ffects C13 norisoprenoid nd crotenoid concentrtions in Riesling. J. Agric. Food Chem. 58: Lkso, A.N., nd W.M. Kliewer The influence of temperture on mlic cid respirtion in grpe erries. Plnt Physiol. 56: Leon, G., E. Duchene, O. Brun, C. Mgne, nd C. Clement Flower scission nd inflorescence crohydrtes in sensitive nd non-sensitive cultivrs of grpevine. Sex. Reprod. 17: Leon, G. G. Wojnrowiez, B. Holzpfel, F. Fontine, N. Villnt-Gveu, nd C. Clement Sugrs nd flowering in the grpevine (Vitis vinifer L.). J. Exp. Bot. 59: Lee, S-H., M.J. Seo, M. Riu, J.P. Cott, D.E. Block, N.K. Dokoozlin, nd S.E. Eeler Vine microclimte nd norisoprenoid concentrtion in Cernet Suvignon grpes nd wines. Am. J. Enol. Vitic. 58: Mris, J., C.J. Wyk, nd A. Rpp Effect of sunlight nd shde on norisoprenoid levels in mturing Weisser Riesling nd Chenin lnc grpes nd Weisser Riesling wines. S. Afr. J. Enol. Vitic. 13: Mthieu, S., N.Terrier, J. Procureur, F. Bigey, nd Z. Gunt A crotenoid clevge dioxygense from Vitis vinifer L.: functionl chrcteriztion nd expression during grpe erry development in reltion to C13-prouct ccumultion. J. Exp. Bot. 56: Mendes-Pinot, M.M Crotenoid rekdown products the-norisoprenoids-in wine rom. Arch. Biochem. Biophys. 483:

132 Meyers, J.M., nd J.E. Vnden Heuvel Enhncing the precision nd sptil cuity of point qudrt nlysis vi clirted exposure mpping. Am. J. Enol. Vitic. 59: Mori, K., S. Sugy, nd H. Gemm Decresed nthocynin iosynthesis in grpe erries grown under elevted night tempertures. Sci. Hortic. 105: Okomi, G., J. Billot, nd C. Hrtmnn Vrition des Teneurs en Chlorophylles et en Crotenoïdes chez l Cerise u Cours de l Conservtion du Fruit Cueilli à Différents Stdes de l Croissnce et de l Mturtion. Physiol. Veg. 13: Olivier, C., A.C. Ferreir, P. Cost, J. Guerr, nd P. Guedes de Pinho Effect of some viticulturl prmeters on the grpe crotenoid profile. J. Agric. Food Chem. 52: Ollt, N., nd J.P. Gudillere The effect of limiting lef re during stge I of erry growth nd development nd composition of erries of Vitis vinifer L. cv. Cernet Suvignon. Am. J. Enol. Vitic. 49: Plliotti, A., M. Gtti, nd S. Poni Erly lef removl to improve vineyrd efficiency: gs exchnge, source-to-sink lnce, nd reserve storge responses. Am. J. Enol. Vitic. 62: Pstore, C., S. Zenoni, M. Fsoli, M. Pezzotti, G. Bttist Tornielli, nd I. Filipetti Selective defolition ffects plnt growth, fruit trnscriptionl ripening progrm nd flvonoid metolism in grpevine. BMC Plnt Biol 13:1-16. Perez, J., nd W.M. Kliewer Effect of shding on ud necrosis nd ud fruitfulness of Thompson seedless grpevines. Am. J. Enol. Vitic. 41: Perym DR, Pilgrim FJ Hedonic scle method of mesuring food preference. Food Tech. 11: 9 14 Pineu, B., J-C Bre, C.V. Vn Leeuwen, nd D. Duourdieu Which impct for β- dmscenone on red wines rom? J. Agric. Food Chem. 55: Poni, S., L. Cslini, F. Bernizzoni, S. Civrdi, nd C. Intrieri Effects of erly defolition on shoot photosynthesis, yield components, nd grpe composition. Am. J. Enol. Vitic. 57: Poni, S., F. Bernizzoni, S. Civrdi, nd N. Lielli Effects of pre-loom lef removl on growth of erry tissues nd must composition in two red Vitis vinifer L. cultivrs. Aust. J. Grpe Wine Res. 15: Poni, S., nd F. Bernizzoni A three-yer survey on the impct of pre-flowering lef removl on erry growth components nd grpe composition in cv. Brer vines. J. Int. Sci Vigne Vin. 44: Price, S.F., P.J. Breen, M. Vllldo, nd B.Y. Wtson Cluster sun exposure nd quercetin in Pinot noir grpes nd wine. Am. J. Enol. Vitic. 46: Quinln, J.D., nd J.R. Wever Modifiction of pttern of the photosynthte movement within nd etween shoots of Vitis vinifer L. Plnt Physiol. 46: Rzungles, A., C.L. Byonove, R.E. Cordonnier, nd J.C. Spis Grpe Crotenoids: chnges during the mturtion period nd locliztion in mture erries. Am. J. Enol. Vitic. 39: Rzungles, A.J., I. Bic, J.C. Spis, nd C.L. Byonove Prticulr ehvior of epoxy xnthophylls during verison nd mturtion of grpe. J. Agric. Food Chem. 44:

133 Rzungles, A.J., R.L. Bumes, C. Dufour, C.N. Sznper, nd C.L. Byonove Effect of sun exposure on crotenoids nd C13-norisoprenoid glycosides in Syrh erries (Vitis vinifer L.). Sci. Alimtents 18: Reynolds A.G., D.A. Wrdle, nd A.P. Nylor Impct of trining system, vine spcing, nd sl lef removl on Riesling vine performnce, erry composition, cnopy microclimte, nd vineyrd lor requirements. Am. J. Enol. Vitic. 47: Ristic, R., M. Downey, P. Ilnd, K. Bindon, I.L. Frncis, M. Herderich, nd S.P. Roinson Exclusion of sunlight from Shirz grpes lters wine colour, tnnin, nd sensory properties. Aus. J. Gr. Wine Res. 13: Ryon, I., B.S. Pn, D.S. Intrigliolio, A.N. Lkso, nd G.L. Scks Effects of cluster light exposure on 3-isoutyl-2-methoxypyrzine ccumultion nd degrdtion ptterns in red wine grpes (Vitis vinifer L. cv. Cernet Frnc). J. Agric. Food Chem. 56: Stini, P., nd G.S. Howell Effects of erly defolition on yield, fruit composition, nd hrvest seson cluster rot complex of grpevines. HortScience 45: Sini, H.S Effects of wter stress on mle gmetophyte development in plnts. Sex. Plnt Reprod. 10: Snchez, L.A., nd N.K. Dokoozlin Bud microclimte nd fruitfulness in Vitis vinifer L. Am. J. Enol. Vitic. 56: Smrt, R.E., nd T.R. Sinclir Solr heting of grpe erries nd other sphericl fruits. Agric. Meteorol. 17: Smrt, R.E., J.B. Roinson, G.R. Due, nd C.J. Brien Cnopy microclimte mnipultion for the cultivr Shirz. II. Effects on must nd wine composition. Vitis 24: Smrt, R., nd M. Roinson Sunlight into Wine. A Hndook for Winegrpe Cnopy Mngement. Winetitles, Adelide, Austrli. Smith S., I. Codrington, M. Roertson, nd R.E. Smrt Viticulturl nd oenologicl implictions of lef removl for New Zelnd vineyrds. In Proc. Second Int. Cool Climte Viticulture nd Oenology Symp., Aucklnd, NZ. R.E. Smrt, R.J. Thorton, S.B. Rodriguez nd J.E. Young Eds. pp New Zelnd Society of Enology nd Viticulture, Aucklnd. Spyd, S., J.M. Trr, D.L. Mee, nd J.C. Ferguson Seprtion of sunlight nd temperture effects on the composition of Vitis vinifer cv. Merlot erries. Am. J. Enol. Vitic. 53: Stff, S.L., D.C. Percivl, J.A. Sullivn, nd K.H. Fisher Fruit zone lef removl influences vegettive, yield, disese, fruit composition, nd wine sensory ttriutes of Vitis vinifer L. Optim nd Cernet frnc. Cn. J. Plnt Sci. 77: Trr, J., J.M. Lee, S.E. Spyd, nd C.F. Scgel Berry temperture nd solr rdition lter cyltion, proportion nd concentrtion of nthocynins in Merlot grpes. Amer. J. Enol. Vitic. 59: Trdguil, J., F. Mrtinez de Tod, S. Poni, nd M.P. Digo Impct of erly lef removl on yield nd fruit nd wine composition of Vitis vinifer L. Grcino nd Crignn. Am. J. Enol. Vitic. 61: Trdguil, J., J.A. Blnco, S. Poni, nd M.P. Digo Mechnicl yield regultion in winegrpes: comprison of erly defolition nd crop thinning. Aust. J. Grpe Wine Res. 18:

134 Verzer, A., G. Tripodi, G. Dim, C. Condurso, A. Sccco, F. Cincott, D.M.L Giglio, T. Sntngelo, nd A. Sprcio Lef removl nd wine composition of Vitis vinifer L. cv. Nero d Avol: the voltile rom constituents. J. Sci. Food Agric. 96: Winterhlter, P. nd R.L. Rouseff Crotenoid-Derived Arom Compounds: An Introduction. In: Crotenoid Derived Arom Compounds; ACS Symposium Series; Winterhlter, P; Rouseff, R.L., Eds.; Americn Chemicl Society: Wshington, DC, Vol. 802, pp Wrolstd, R. E., T.E. Acree, nd E.A. Decker Hndook of Food Anlyticl Chemistry: Pigments, Colornts, Flvors, Texture, nd Bioctive Food Components. Wiley- Interscience: New York, 2005; pp Wolf, T.K., R.M. Pool, nd L.R. Mttick Responses of young Chrdonny grpevines to shoot tipping, etephon, nd sl lef removl. Am. J. Enol. Vitic. 37: Wolf, T.K., ed Wine grpe production guide for estern North Americ. Grpevine Cnopy mngement (A. Reynolds nd T.K. Wolf), pp Ymmoto, H.Y Biochemistry of the violxnthin cycle in higher plnts. Pure Appl. Chem. 51: Ymne, T., nd K. Shiym Effects of chnges in the sensitivity to temperture on colortion of Aki Queen grpe erries. J. Jpn Soc. Hort. Sci. 75: Zhng, M., Q. Xu, C. Dun, W. Qu, nd Y. Wu Comprtive study of romtic compounds in young red wines from Cernet Suvignon, Cernet frnc, nd Cernet gernischet vrieties in Chin. J. Food Sci. 72:

135 Extent nd timing of fruit-zone lef nd lterl shoot removl lters yield components nd fruit composition in Cernet Suvignon grpes. Astrct Bckground nd ims: Aggressive fruit-zone lef removl cn improve disese mngement, prticulrly in humid regions. However, current fruit-zone lef removl prctices tend to e conservtive nd hve seen little refinement over the lst decde or more. We hypothesized tht pre-loom removl of sl leves/lterl shoots would reduce crop yield nd tht ggressive sl lef/lterl shoot removl would improve totl phenolics nd nthocynins in Cernet Suvignon. Methods nd results: Two experiments were conducted to evlute the effects of pre-loom removl of no (PB-NO), four (PB-4), or eight (PB-8) sl leves/lterl shoots (Experiment 1, ) nd post-fruit set removl of no (PFS-NO) nd six (PFS-6) sl leves/lterl shoots (Experiment 2, ) on crop yield components, nd totl grpe phenolics nd nthocynins in Cernet Suvignon. Experiment 1: Pre-loom removl of eight leves/lterl shoots reduced ll yield components to greter extent thn did PB-4. When compred to PB- NO, PB-4 reduced erry numer per cluster y 35-51%, cluster weight y 33-53%, nd crop yield y 51-53% over When compred to PB-NO, PB-8 reduced erry weight y 9-19%, erry numer per cluster y 52-73%, cluster weight y 57-78%, nd crop yield y 55-78% over When compred to PB-NO, yield components tended to e reduced y greter percentge due to re-implementtion of oth PB-4 nd PB-8 in consecutive sesons. Compred to PB-NO, PB-4 reduced cluster compctness y 25-39%, nd PB-8 reduced cluster compctness 58-68% over Botrytis unch rot incidence ws reduced y PB-4 y 87% nd y PB-8 y 100% when compred to PB-NO in Pre-loom removl of eight leves/lterl shoots reduced solule solids in two of three yers, nd oth PB-4 nd PB-8 119

136 reduced titrtle cidity when compred to PB-NO in Both PB-4 nd PB-8 incresed totl grpe phenolics y n verge of 14-31% when compred to PB-NO over While PB-4 incresed totl grpe nthocynins y n verge of 9% when compred to PB-NO in 2014, oth PB-4 nd PB-8 incresed totl grpe nthocynins y n verge of 22% when compred to PB-NO in Experiment 2: Botrytis unch rot incidence ws reduced y 78% y PFS-6. Post-fruit set removl of six leves/lterl shoots reduced solule solids in one yer, nd ph nd titrtle cidity in oth yers. Post-fruit set removl of six leves/lterl shoots incresed totl grpe phenolics in 2014 (13%) nd 2015 (16%), nd incresed totl grpe nthocynins in 2015 (13%). Conclusions: Aggressive removl of fruit-zone leves/lterl shoots tended to increse grpe phenolics nd nthocynins nd reduce otrytis unch rot incidence, regrdless of timing of removl. While pre-loom lef/lterl shoot removl resulted in greter concentrtions of these compounds in grpes compred to post-fruit set lef/lterl shoot removl, pre-loom lef/lterl shoot removl lso reduced crop yield y n verge of 57% compred to no lef removl. Significnce of the study: Aggressive removl of fruit-zone leves/lterl shoots improves the proility of getting disese-free fruit into the winery. Becuse the climte of the humid estern US is not s conducive to heting fruit to criticl tempertures s in other climtes, removing leves/lterl shoots to n equivlent of 0 fruit-zone lef lyers is not deleterious to fruit qulity. As such, disese mngement nd fruit qulity cn e concomitntly improved with ggressive lef removl. However, if removl of leves/lterl shoots occurs efore loom, crop yield cn e drmticlly reduced, depending on the extent of green tissue removl t this criticl stge. 120

137 Introduction Fruit-zone lef removl reduces disese incidence nd severity (English et l. 1989, Wolf et l. 1986). Accordingly, fruit-zone lef removl is lmost uiquitous recommendtion, especilly in humid climtes where fungl diseses re prevlent. Erly works on fruit exposure nd diseses, in ddition to generl cnopy mngement nd environmentl impct studies (Jckson nd Lomrd 1993, Smrt nd Roinson 1991), collectively recommended ggressive lef removl to improve grpe disese mngement nd wine qulity potentil. More recent studies implied tht the more open the fruit-zone, the etter recommendtion my not e true, s excessive fruit exposure cn rdintly het fruit to tempertures tht re prticulrly detrimentl to nthocynins (Bergqvist et l. 2001, Mori et l. 2007, Spyd et l. 2002, Trr et l. 2008). Though these field studies were conducted in reltively wrm/dry growing regions of southestern WA (Spyd et l. 2002, Trr et l. 2008) nd the Sn Joquin Vlley of CA (Bergqvist et l. 2001), fruit-zone mngement recommendtions ecme, or remined, conservtive - even in humid/temperte grpe growing regions (1-2 lef lyers, Wolf 2008), even though nthocynins re ut one clss of the severl clsses of sensory impct compounds in grpes, nd even when mny studies reported grpe nd wine qulity to e incresed with fruit exposure nd reduced with shde (Lee et l. 2007; Ryon et l. 2008, Hunter, 1991, Jckson nd Lomrd 1993, Smith et l. 1988, Smrt et l. 1985, Reynolds et l. 1996, Bledsoe et l. 1988, Cronneu 1985, Di Profio et l. 2011, Stff et l. 1997, Smrt nd Roinson 1991). Excessive fruit exposure cn e deleterious to nthocynins (Bergqvist et l. 2001, Mori et l. 2007, Spyd et l. 2002, Trr et l. 2008). However, ggressive lef removl incresed nthocynins (Di Profio et l. 2011, Stff et l. 2007), even when conducted in regions typiclly drier nd wrmer thn the estern US (Digo et l. 2012, Kotseridis et l. 2012). In Greece, post- 121

138 flowering lef removl incresed nthocynins in Cernet Suvignon nd Merlot, ut not in Sngiovese, puttively function of cultivr-specific tolernce of nthocynins to detrimentl tempertures, or greter source: sink rtio dominted y more photosyntheticlly ctive picl leves (Kotseridis et l. 2012). Post-fruit set lef removl improved nthocynins in Temprnillo grpes in Rioj, Spin, which ws ttriuted to incresed erry skin mss nd/or greter lef re: crop weight rtio (Digo et l. 2012). Similr fctors were t lest prtilly responsile for improving nthocynins nd totl phenols in pre-loom lef removl studies conducted on severl vrieties (Gtti et l. 2012, Pllioti et l. 2011, Poni et l nd 2008, Trdguil et l. 2010). While the enefits of more open fruit-zone microclimte could not e ruled out (Kotseridis et l. 2012, Trdguil et l. 2012), pre-loom lef removl improved grpe nthocynins y other mechnisms, such s reduced erry size nd incresed skin: pulp rtio (Bogicevic et l. 2015, Poni et l. 2006) thicker erry skins (Pllioti et l. 2011, Poni et l. 2008, Trdguil et l. 2010) nd greter lef re: fruit weight rtios (Digo et l. 2012, Intrieri et l. 2008, Gtti et l. 2012, Poni et l. 2006). There re other documented enefits of pre-loom lef removl. Pre-loom lef removl improved cnopy efficiency nd hstened solule solids ccumultion (Pllioti et l. 2011, Poni et l. 2006), reduced cluster compctness (Pllioti et l. 2011) nd unch rot (Trdguil et l nd 2012), nd did not cuse sunurn (Bogicevic et l. 2015) or reduced sunurn incidence compred to lef removl t verison (Pstore et l. 2013). Thus, ggressive pre-loom removl wrrnts considertion for use in more humid, temperte growing regions, where detrimentl grpe tempertures were speculted to occur less frequently, diseses re more prevlent, nd fruit ripening is occsionlly delyed y inclement wether. 122

139 Light is necessry for nthocynin synthesis, prticulrly if uncoupled from temperture extremes (Trr et l. 2008). However, the estern US tends to e cloudier nd more temperte thn regions where the ove cited studies were conducted. Thus, it ws hypothesized tht ggressive fruit-zone lef removl in humid region would increse rdition penetrtion to grpes nd only modestly increse erry tempertures, conditions tht fvor grpe nthocynins reltive to erry temperture extremes in either sun or shde (Trr et l. (2008). Removing leves t the pre-loom stge reduced cluster compctness nd, consequently, unch rot incidence (Hed et l. 2009, Pllioti et l. 2011, Poni et l. 2006, Stini nd Howell 2010, Trdguil et l nd 2012). It ws hypothesized tht fruit-zone lef removl in humid region such s Virgini could e more ggressive thn currently recommended (1-2 lef lyers, Reynolds nd Wolf 2008) in order to improve disese mngement without compromising nthocynins or other spects of fruit chemistry. The gol of this experiment ws to evlute ggressive pre-loom or post-fruit set lef/lterl shoot removl s tool to lter crop yield components, or grpe nthocynins nd phenolics, in Cernet Suvignon. An dditionl gol ws to quntify the tempertures typicl of shded nd exposed erries in humid region. It ws nticipted tht the results of this study would chrcterize how fruit exposure cn impct erry temperture, nthocynins, nd phenolics in the mcroclimte of the humid estern US. Mterils nd Methods Tretments nd experimentl design: Two experiments were conducted, oth of which used Cernet Suvignon ENTAV-INRA clone 337 vines, grfted onto 420-A rootstock, nd grown t Virgini Tech s AHS, Jr. Agriculturl Reserch nd Extension Center (AHS, Jr. AREC) ner Winchester, VA (39 11 N; W). In oth experiments, vines were plnted in My 2006 in 123

140 rows running generlly northest/southwest t 3.0-m (row) x 1.5-m (vine) spcing nd were trined to ilterl cordons with verticlly-positioned shoots. The soil ws Poplimento Hgerstown sndy lom (A. Blckurn, personl communiction, 2013). The inter-row groundcover, estlished in 2001, initilly comprised mixture of orchrd grss (Dctylis glomert) nd tll fescue (Festuc rundince); cv. Shenndoh, with the fescue dominting fter ~ six yers. The intr-row groundcover in the pre-loom lef/lterl shoot removl experiment consisted of perennil creeping red fescue (F. rur), estlished in Sep 2008, nd n 85-cm wide hericide-treted strip in the post-fruit set lef/lterl shoot removl experiment. Lef/lterl shoot removl tretments in the two seprte experiments were s ulleted elow (NOTE: from this point forwrd, tretments will e referred to s lef removl tretments): Expt. I: Pre-loom lef/lterl shoot removl experiment ( ): PB-NO - no leves/lterl shoots removed; lterl shoots mintined t ~3-4 nodes. PB-4 - removl of leves/lterl shoots from primry shoot nodes 1-4 efore loom, [modified Eichorn nd Lorenz (EL) stge (Dry nd Coome 2004)]; distl lterl shoots mintined t ~3-4 nodes. PB-8 - removl of leves/lterl shoots from primry shoot nodes 1-8 efore loom, [modified Eichorn nd Lorenz (EL) stge (Dry nd Coome 2004)]; distl lterl shoots mintined t ~3-4 nodes. Expt. II: Post-fruit set lef/lterl shoot removl experiment ( ): PFS-NO - no leves/lterl shoots removed; lterl shoots mintined t ~3-4 nodes. 124

141 PFS-6 - removl of leves/lterl shoots from primry shoot nodes 1-6 t peerry size / unch closure (modified EL stge 31/32); distl lterl shoots mintined t ~3-4 nodes. A complete lock design ws used in ech experiment, prtitioning the experimentl re into six locks, ech seprted y five-vine order plots within the row nd y ordering uffer rows. Within ech lock, lef removl tretments were rndomly ssigned to either one-vine experimentl units (pre-loom experiment) or two-vine experimentl units (post-fruit set experiment). Pre-loom nd post-fruit set lef removl experimentl units were no more thn ~ 45 m wy from ech other in ech vineyrd lock. Fruit-zone porosity ws monitored weekly, nd mintined from tretment initition through hrvest y removing ny green tissue re-growth into the fruit-zone. Pre-loom lef removl in 2013 ws re-implemented on the sme vines in 2014 nd 2015; these vines re henceforth referred to s PB-4 13re, or PB-8 13re. Both preloom nd post-fruit set lef removl were re-implemented in 2015 on vines initilly used in 2014; only the pre-loom vines re henceforth referred to s PB-4 14re, or PB-8 14re. Figure 1 depicts the chronology of current-seson nd re-implemented pre-loom lef removl sptilly nd temporlly, using the vines in experimentl lock 1 s n exmple. 125

142 2013 ROW POST PB-NO X PB-4 X X PB-8 X X ROW POST 2014 ROW POST PB-NO X PB-4 13re X PB-8 X PB-8 13re X PB-4 X ROW POST 2015 ROW POST PB-NO X PB-4 13re X PB-8 14re X PB-8 13re X PB-4 14re X ROW POST Fig. 1 Chronology of pre-loom lef removl implementtion nd re-implementtion in experimentl lock 1 over the course of the entire study, X = vine; PB-4/8 13re = re-implementtion of pre-loom lef removl tht initilly occurred in 2013; PB-4/8 14re = re-implementtion of pre-loom lef removl tht initilly occurred in Meteorology: Rinfll nd mient ir temperture were recorded hourly from 1-Apr through 24- Jun 2013 using n ET106 wether sttion (Cmpell Scientific, Logn, UT) locted ~150 m from the experimentl vineyrd. Rinfll nd mient ir temperture were lso recorded hourly from 25-Jun 2013 through 31-Oct 2013, nd 1-Apr through 31-Oct 2014 nd 2015 using CS215 temperture/reltive humidity proe (Cmpell Scientific, Logn, UT) nd TE525WS-L rin guge (Texs Electronics, Dlls, TX) locted ~25 nd ~350 m wy from the experimentl vineyrd, respectively. Growing degree dys (GDD) were summed from 1 Apr using se temperture of 10 C. Cnopy chrcteriztion: To chrcterize fruit-zone rchitecture, point qudrt nlysis (PQA) dt were collected etween EL stges 33 nd 35 (verison), s descried in Smrt nd Roinson (1991). A thin metl rod ws inserted into the fruiting zone long the trnsverse xis of the cnopy, using tpe mesure to guide insertions. This process ws repeted 10 times in ech experimentl unit (vine) to quntify fruit-zone lef lyer numer (LLN). Fruit zone sunlight 126

143 penetrtion ws evluted within severl dys of collecting PQA dt using n AccuPAR ceptometer (Model PAR-80, Decgon Devices, Inc., Pullmn, WA). Photosynthetic photon flux density (PPFD) ws ssessed y inserting the ceptometer inside cnopy fruit zones prllel to, nd directly ove, the cordon nd orienting the light interception side of the ceptometer in three different directions (45 est, verticl, 45 west) nd then verging those redings. Two PPFD redings were tken in every experimentl unit etween hrs under cler skies. An mient PPFD reding ws tken with ech experimentl unit reding in order to express fruitzone PPFD s percentge of the mient rdition in order to generte cluster exposure flux vilility (CEFA), with enhnced point qudrt nlysis (EPQA version 1.6.2) softwre (Meyers nd Vnden-Heuvel 2008). The dtlogger vine pnel: The following mesurements were logged t specified intervls in ech respective seson. 2013: erry nd mient temperture, nd fruit-zone PAR were logged on 15-min intervls from 30 Jul 2013 through 9 Oct : erry nd mient temperture, nd fruit-zone nd mient PAR were logged on 15-min intervls from 25 Jul 2014 through 4 Sep 2014, nd on 1-min intervls from 6 Sep 2014 through16 Oct 2014, nd from 22 Jun 2015 through 5 Oct : erry nd mient temperture, nd fruit-zone nd mient PAR were logged on 1- min intervls from 22 Jun 2015 through 5 Oct One experimentl unit (vine) of ech pre-loom lef removl tretment (post-fruit set lef removl vines were never used) in the sme vineyrd pnel (mx. 7.6 m prt) ws sujected to logging of erry temperture nd fruit-zone photosyntheticlly ctive rdition (PAR). This ws done to mesure erry tempertures typicl of our region, nd evlute how light nd 127

144 temperture impct erry temperture. In ddition to these lef removl tretment-specific mesurements, mient temperture nd PAR were logged immeditely surrounding nd ove the dtlogger vine pnel, respectively. The temperture of two outside-fcing erries on oth est nd west cnopy sides of ech lef removl tretment were mesured with mini hypodermic thermocouple (type T, model HYP1/2, Omeg Eng., Stmford, CT) inserted pprox. 0.6 cm eneth erry skins nd ffixed with strong, ll-purpose glue. Fruit-zone PAR ws mesured with SQ-316 quntum sensor (Apogee Instruments, Logn, UT) plced on top, nd prllel to, the orienttion of one cordon in ech lef removl tretment. Amient PAR ws mesured with SQ-316 quntum sensor, mounted pprox. 3 m ove the ground (pprox. 0.6 m ove vine cnopies) t the middle of the vineyrd pnel where ll dt were logged. Amient temperture ws mesured with two ST-100 thermistors (Apogee Instruments, Logn, UT). The thermistors were plced on the north nd south ordering ends (~ 7.6 m prt) of the dtlogger vine pnel, nd mounted t fruit-zone height inside nturlly spirted 6-pnel rdition shield (model A, R.M. Young, Trverse City, MI). Thermocouples were logged with n AM25T solid-stte multiplexer (Cmpell Scientific, Logn, UT) ttched to CR1000 dtlogger (Cmpell Scientific, Logn, UT), which lso logged the PAR nd mient temperture dt. All sensors, excepting the mient temperture sensors, were tken down t the end of ech field seson nd re-mounted in the following seson. Mnul erry temperture: Berry temperture ws mnully mesured on severl dtes over the course of the three sesons in order to etter chrcterize erry tempertures cross ll six experimentl locks (s opposed to just using dt logged in one lock). Berry tempertures were mnully mesured on five dtes etween Jul nd Sep in 2013 (pre-loom lef removl tretments), six dtes etween Jul nd Sep in 2014, nd seven dtes etween Jun nd Sep in

145 (oth pre-loom nd post-fruit set lef removl tretments). Mesurements were tken in every experimentl unit t three different times of dy on ech collection dte: morning (~ hrs), round solr noon (~ hrs), nd lte fternoon (~ hrs). Mesurements were tken y inserting mini hypodermic thermocouple (model HYP1/2, Omeg Eng., Stmford, CT) eneth the skins of erries, which ws connected to hndheld digitl thermometer (model HH 25, Omeg Eng., Stmford, CT). Ech mesurement took ~ 2 sec. At ech time of dy, on ech collection dte, nd on oth est nd west cnopy sides, three erries on the cluster s exterior fce, positioned t the top, center, nd ottom of clusters were mesured on two clusters orne on opposite vine cordons in ech experimentl unit. Thus, totl of twelve totl single-erry temperture mesurements, or six single-erry temperture mesurements on ech cnopy side, were tken in ech experimentl unit. The timefrme tht encompssed mnul erry temperture mesurement in ech experimentl unit ws recorded strting in Aug 2014; this ws done in order to further investigte the reltionship etween erry temperture, nd mient ir temperture nd rdition. Estimted post-verison erry temperture in post-fruit set lef removl plots: An ttempt ws mde to estimte the cumultive length of time tht erries in the post-fruit set removl plots were sujected to two enchmrk tempertures, 30 nd 35 C, in the post-verison period. The percent difference etween verge mnully-mesured erry temperture in pre-loom removl nd post-fruit set lef removl plots during direct-sun times of dy ws clculted to e 4-5%, depending on cnopy side nd seson (Tle 1). This 4-5% increse ws djusted for erries in the PFS-6 over ech hour period etween , s follows: First, the verge hourly percent increse of logged erry temperture in pre-loom lef removl plots compred to mient ir temperture ws clculted etween (Tle 1). The hourly percent 129

146 increse in erry temperture in pre-loom lef removl plots ws normlized ccording to the pek verge percent increse of logged erry temperture, mking nd % (or, 1.00) (Tle 1). The percent increse in mnully mesured post-fruit set erry temperture during direct-sun time periods (4-5%) ws then multiplied y the normlized percent increse of erry temperture in pre-loom lef removl plots for ech respective hour. The result ws n estimtion of the C tht erry temperture ws greter in post-fruit set lef removl compred to pre-loom lef removl plots (Tle 1, furthest column to right). This numer ws dded ck to the verge of every logged temperture of PB-4 nd PB-8 erries in ech respective hour over the entire 2014 nd 2015 post-verison periods. The totl estimted time tht erries in the post-fruit set lef removl plots spent ove 30 nd 35 C ws then summed. 130

147 Tle 1. Method used to estimte hourly percent increse in PFS-6 erry temperture compred to the verge of PB-4/8 erry temperture in 2014 nd % increse in verge logged PB-4/8 erry vs. mient ir temperture Actul % increse Averge % over directsun periods Adjusted sed off pek % increse % increse in PFS-6 vs. PB-4/8 erry temperture Clculted % increse in mnully-mesured erry temperture during direct-sun time periods Adjusted % increse in PFS-6 erry temperture sed off pek % increse of logged PB-4/8 erry vs. mient ir temperture Hour EAST WEST EAST WEST EAST WEST EAST WEST EAST WEST Hour EAST WEST EAST WEST EAST WEST EAST WEST EAST WEST

148 Crop yield components, vine fruitfulness, erry weight, cluster compctness, nd otrytis unch rot nd unch stem necrosis incidence: Crop ws hrvested t commercilly cceptle composition nd integrity. Crop yield per vine ws mesured with field scle t hrvest on 9 Oct 2013, 20 Oct 2014, nd 5 Oct 2015 to otin components of yield. Cluster weight ws estimted from the quotient of crop yield weight nd cluster numer. Berry weight ws determined from 120-erry smples, hlf of which were rndomly collected from ech of est nd west cnopy sides immeditely prior to hrvest. Berry numer per cluster ws derived from the quotient of verge cluster weight nd verge erry weight. Crop lod ws clculted using crop yield weight nd dormnt pruning weight on per-vine sis. Berry weight gin over time ws determined from 120- or 200-erry smples collected from ech experimentl unit t vrious times efore hrvest in Cluster compctness ws mesured in Expt. I to evlute if pre-loom lef removl impcted cluster rchitecture. Cluster compctness ws mesured t hrvest on ten clusters tht were +/- 25% of the verge hrvested cluster weight from ech experimentl unit. Compctness ws indexed y the rtio of totl erry numer y the min rchis length. Berries were not counted, nd rchis sections not mesured, on cluster shoulders or wings in order to reduce confounding the compctness of the min ody of the cluster. Vine fruitfulness ws ssessed in 2014 nd 2015 t the modified EL stge 15/16 y dividing the numer of inflorescences y the numer of shoots originting from oth the cordon (sl) nd one-yer old spurs (count). Crop yield components were lso evluted in experimentl units sujected to pre-loom lef removl re-implementtion. In 2014, the second yer of Expt. I, current-seson pre-loom lef removl tretments were erry smpled five times; re-implemented tretments were not erry smpled until hrvest. Consequently, the solute vlues of components of yield nd 132

149 cluster compctness were greter in the re-implemented (PB-4/8 13re) compred to currentseson plots. Therefore, in 2014, the yield components of PB-4/8 13re were compred to estimted yield components of the frequently-erry smpled PB-NO. Yield components were estimted y dding the product of erry weight t hrvest nd smpled erry numer ck to the ctul hrvest weight for ech experimentl unit; components of yield were then clculted s descried ove. Cluster compctness of PB-NO in 2014 ws estimted y dding the verge erry numer removed per cluster in ech experimentl unit to the verge erry numer per cluster mesured during the cluster compctness ssessment; this ws then divided y the min rchis length. All pre-loom lef removl tretments were erry-smpled with equl frequency in 2015 to eliminte the need for estimtion. Botrytis unch rot (BBR) nd unch stem necrosis (BSN) incidence ws evluted y visul inspection of every hrvested cluster in A severity of ~ 5% of the entire cluster ws mesured s n incident, ut severity of oth BBR nd BSN ws greter in most cses. In ech experimentl unit, percent incidence ws clculted s the numer of clusters with BBR nd/or BSN divided y the totl numer of clusters. Primry fruit chemistry: Primry chemistry ws mesured on juice otined from composite smples of 30 est- nd 30 west- cnopy side positioned erries tht were rndomly collected from ech experimentl unit immeditely prior to hrvest. Juice smples were otined from fresh (non-frozen) erries y eqully hnd-pressing smples in plstic g for ~ two min, nd then immeditely centrifuging for five min t ~ 3500 rpm. Solule solids were mesured with digitl refrctometer (Pocket PAL-1, ATAGO USA, Inc., Bellevue, WA). Juice ph ws mesured, nd TA ws determined y titrtion to n endpoint of ph 8.2, using n 848 Titrino Plus uto-titrtor (Metrohm USA, Riverview, FL) nd 0.1 N NOH se. 133

150 Estimted totl grpe phenolics nd nthocynins: Asorption spectroscopy ws used to estimte totl grpe phenolics nd nthocynins s ffected y pre-loom nd post-fruit lef removl. Immeditely prior to hrvest in 2013 (just pre-loom experiment), 2014, nd 2015 (pre-loom nd post-fruit set experiments), 60-erry smples were rndomly collected from oth est nd west cnopy sides of ech experimentl unit, kept seprte, nd frozen t -20 C until tests commenced. Excepting 2014 post-fruit set smples, which were homogenized with n immersion lender (Wring), erry smples were homogenized for two min with Mgic Bullet (Homelnd Housewres LLC, Los Angeles, CA), nd hd 30 ml of M KCl uffer (ph 1.0) nd 0.4 M sodium cette uffer (ph 4.5) dded to seprte 1.0 g liquots of the erry homogente. The homogente-uffer smples were shken for 10 min nd then centrifuged for five min t ~ 3500 rpm. The superntnt ws pipetted into 10 mm pth length Hellm qurtz cuvette (Thermo Fisher Scientific Inc., Pittsurgh, PA), nd the sornce t 520 nd 700 nm ws mesured in duplicte with Genesys 8 ThermoSpectronic spectrophotometer (Cmridge, UK). The sornce of the smple contining the erry homogente nd M KCl uffer (ph 1.0) ws lso red t 280 nm in duplicte fter diluting to 2:1 smple-uffer rtio. Sttistics: A mixed model ws used in JMP Pro 11 (SAS, Cry NC) to evlute the rndom effect of Block nd fixed effects of pre-loom nd post-fruit set lef removl tretment (α 0.05). Cnopy side nd tretment interction with cnopy side (est nd west) ws dded to the model to evlute estimted totl erry phenolics nd nthocynins. Significnt difference of lest squre mens (α 0.05) ws determined with Tukey s HSD for pre-loom lef removl. Significnt difference of lest squre mens (α 0.05) ws determined with Student s T-test for post-fruit set lef removl, s well s cnopy side effect on estimted totl phenolics nd nthocynins. Mnul erry temperture mesurements were nlyzed seprtely y time of dy 134

151 nd cnopy side on ech dte. Due to differentil erry smpling etween current-seson nd reimplementtion tretments in Expt. I, components of yield dt were nlyzed seprtely in Tretment effect on crop yield t hrvest ws reversed in lock 5 in 2013, nd lock 4 in 2014, when compred to ll other experimentl locks in Expt. I. In 2013, this ws due to no lef removl hving t lest 22% fewer clusters per vine compred to pre-loom lef removl tretments. In 2014, this ws due to no lef removl hving t lest 61% fewer clusters per vine compred to the pre-loom lef removl tretments. As such, ll components of yield dt from lock 5 in 2013 nd lock 4 in 2014, except erry weight, ws removed from the nlyses presented in Tle 4. Simple liner regressions were run to determine the reltionship etween erry weight t hrvest nd totl erry phenolics nd nthocynins hrvest erry weights from ll locks were included in these reltionships. Results Meteorology: Sesonl growing degree dy (GDD) ccumultion in 2015 ws ~ 8% greter thn in 2013 nd 2014, minly ttriuted to greter GDD ccumultion in the post-verison period (Aug nd Sep) (Tle 2). The reltively low GDD ccumultion in Oct 2015 ws unlikely detrimentl to fruit mturtion, s hrvest occurred erlier in 2015 thn in 2013 or in Monthly men temperture ws gretest in Jun 2013, nd Aug nd Sep 2015 compred to the sme months in the other two yers. Monthly mximum tempertures were gretest in Jun nd Oct 2013, nd Aug 2015 compred to the sme months in other yers. Sesonl rinfll ws 25% greter in 2013 compred to 2014 nd 2015 sesons, prtilly function of rinfll in Jul 2013, which ws ~2.6 times greter thn Jul 2014, nd 3.7 times greter thn Jul Rinfll ws lowest in Aug 2015, nd gretest in Sep 2015, compred to these months in 2013 nd

152 Tle 2. Sesonl nd monthly growing degree dy (GDD) nd rinfll ccumultion, nd men nd mximum monthly tempertures for Jul, Aug, Sep, nd Oct t the Alson H. Smith, Jr. AREC in Winchester, VA, Yer Metric Seson totl (1 Apr - 31 Oct) Jul Aug Sep Oct GDD sum Rinfll sum (mm) T men ( C) n/ T mx ( C) n/ GDD sum Rinfll sum (mm) T men ( C) n/ T mx ( C) n/ GDD sum Rinfll sum (mm) T men ( C) n/ T mx ( C) n/ Growing degree dys clculted using se 10 C. Fruit hrvested on 9-Oct-2013, 20-Oct-2014, nd 5-Oct The rnge of sesonl mient PAR ws divided into ritrry 25% increments, nd the time spent in ech qurtile ws summed to determine how mient rdition influenced diurnl est- nd west-side erry temperture. In generl, less time ws spent t lower mient PAR rnges, nd more time t higher mient PAR rnges, when sun ws cst on the est-cnopy side during the morning (Fig. 2 A, C), compred to when the sun ws cst on the west-cnopy side in the fternoon (Fig. 2 B, D). In oth 2014 nd 2015, more time ws spent 25 to 50% mient PAR rnges during compred to (Fig. 2 A-D). Over the course of the entire diurnl period tht erries were heted ove mient ir temperture ( ), mient PAR trends were similr etween 2014 nd 2015, lthough more time ws spent t the 25-75% mient PAR rnges in 2015 (Fig. 2 E, F). 136

153 % of time A PAR (mmol m-2s--1) B PAR (mmol m-2s--1) % of time C PAR (mmol m-2s-1) D PAR (mmol m-2s-1) % of time E PAR (mmol m-2s-1) F PAR (mmol m-2s-1) Figure 2. Percent of time from verison to hrvest spent t 25% photosynthetic ctive rdition (PAR) increments from in 2014 (A) nd 2015 (B), from in 2014 (C) nd 2015 (D), nd from in 2014 (E) nd 2015 (F). 137

154 Cnopy chrcteriztion nd dormnt cne pruning weight: Lef removl resulted in fruitzone lef lyer numer (LLN) of zero in ll yers (Tle 3). Lef removl resulted in t lest three- nd, sometimes, four-fold increse in fruit-zone cluster exposure flux vilility (CEFA) compred to removing no leves, nd there ws no difference in CEFA etween PB-4 nd PB-8. Though differences etween PFS-6 nd PB-4 nd PB-8 could not e sttisticlly nlyzed, their LLN nd CEFA vlues were similr. As such, the pre-loom nd post-fruit set lef removl experiments offered pltform to compre the impct of lef removl timing without the confounding of fruit exposure extent. Pruning weight ws reduced y PB-8 in Pruning weight tended to e reduced over time when pre-loom lef removl ws re-implemented in consecutive sesons, ut only to significnt extent y PB-8 13re compred to PB-NO in 2014, nd oth PB-NO nd PB-4 in Tle 3. Pre-loom nd post-fruit set lef/lterl removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) mesured t verison, nd dormnt cne pruning weights, Tretment LLN CEFA Pruning weight (kg / m row) LLN CEFA Pruning weight (kg / m row) LLN CEFA Pruning weight (kg / m row) PB-NO PB PB PB-4 13re n/ n/ n/ n/ n/ 0.90 n/ n/ 0.79 PB-8 13re n/ n/ n/ n/ n/ 0.66 n/ n/ 0.60 Significnce < < < < < < PFS-NO n/ n/ n/ PFS-6 n/ n/ n/ Significnce n/ n/ n/ < < ns < < ns 2013: PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively; 2014: PB-4 13re, PB-8 13re = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in 2013; 2015: PB-4, PB-8 = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in PFS-NO nd PFS-6 = post-fruit set removl of no nd six leves, respectively. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments) nd Student s T-test (PFS tretments). 138

155 Berry temperture: Berry temperture ws driven y mient ir temperture, ut ws confounded y lef removl tretment (Tle 4). The ir-erry temperture reltionship ws stronger when leves were mintined in the fruit-zone compred to when leves were removed from the fruit-zone, regrdless if erry tempertures were logged or mnully mesured. The ir-erry temperture reltionship ws stronger on the west compred to the est cnopy side, especilly when fruit-zone leves were removed. Accordingly, this reltionship ws weker in the morning (AM), when the sun ws cst on the est cnopy side, compred to the fternoon (PM), when the sun ws cst on the west cnopy side. Berry temperture ws more closely relted to ir temperture t solr noon (NOON) compred to morning or fternoon hours. 139

156 Tle 4. Logged nd mnul-mesured coefficients of determintion for liner reltionships etween mient ir nd erry temperture s function of fruit-zone lef removl, cnopy side, nd time of dy. Logged (R 2 ) Mnul c (R 2 ) Tretment EAST WEST AM NOON PM EAST WEST AM NOON PM PB-NO PB PB PFS-NO n/ n/ n/ n/ n/ PFS-6 n/ n/ n/ n/ n/ PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively; PFS-NO, PFS-6 = post-fruit set removl of no nd six leves, respectively. Dt logged nd verged over 30-Jul Oct-2013, 25-Jul Oct-2014, nd 22-Jun Oct-2015; AM = , NOON = , PM = c Dt collected nd verged on 20-Aug, 28-Aug, 4-Sep, nd 16-Sep in 2014, nd on 29-Jun, 13-Jul, 31-Jul, 13-Aug, 25-Aug, nd 8-Sep in 2015; AM = , NOON = ; PM =

157 Both erry nd mient ir temperture were greter in the fternoon compred to morning hours. The time of dy tht erry temperture ws most different from mient ir temperture ws when the sun ws cst on the fruit-zone, from , or from (Fig. 3). This imodl diurnl trend of elevted erry temperture ws mirrored y imodl diurnl trend of elevted PAR, prticulrly in the PB-4 nd PB-8 tretments (Fig. 3 A, B). Estnd west-cnopy side erry temperture in lef removl tretment plots ws greter thn mient ir temperture t times of dy when the sun ws cst on the fruit-zone, or when the overhed cnopy ws not locking rdition trnsmission to the fruit-zone. For exmple, in PB-4 nd PB- 8 plots, est-side erry temperture ws C greter thn mient ir temperture from , nd west-side erry temperture ws C greter thn mient ir temperture from (Fig. 3 A, B). The greter erry temperture on the est- compred to west- cnopy side ws mirrored y greter mient nd fruit-zone PAR conditions etween ( , nd µmol m -2 s -1, respectively) compred to ( , nd µmol m -2 s -1, respectively). Though mient PAR ws t its gretest, there ws less of lef removl impct on rdint erry heting round solr noon (~1330) compred to the preceding or following severl hours. This ws documented y concomitnt decrese in fruit-zone PAR, nd the mgnitude of difference etween mient ir nd erry temperture, prticulrly oserved in Fig. 3 A, B. Astining from fruit-zone lef removl resulted in lockge of fruit-zone rdition throughout the dy (Fig. 3 C). Consequently, est- nd westside erry temperture ws similr ( C) to mient ir temperture during the sme hours tht erries in lef removl plots were rdintly heted ove mient ir temperture. 141

158 30 25 A Amient ir nd erry temperture, T ( C) Berry T EAST Berry T WEST Amient T Amient PAR Fruit-zone PAR ~ solr noon C B Amient nd fruit-zone photosyntheticlly ctive rdition, PAR ( mol m -2 s -1 ) Hour Fig. 3 Diurnl pttern of mient ir temperture, mient nd fruit-zone PAR, nd erry temperture s ffected y pre-loom removl of eight (A), four (B), nd no (C) fruit-zone leves/lterls. Dt logged on 15- nd 1-min intervls, nd verged over sesons. Amient PAR ws logged on 15- nd 1-min intervls, nd verged over sesons. NOTE: Amient PAR presented s 20% of ctul vlue to ese visuliztion of dt plots

159 Pre-loom nd post-fruit set lef removl hd greter effect on mnully-mesured estside erry temperture during the morning (~ ; Figs. 2 nd 3 A), nd west-side erry temperture during the lte fternoon (~ ; Figs. 2 nd 3 C), when compred to opposite cnopy sides during those periods. Berry temperture ws less ffected y lef removl tretment t NOON, even though mient PAR ws frequently greter t this time compred to the AM or PM. This ws function of rdition lockge to the fruit-zone, s oserved in the diurnl ptterns of logged ir nd erry temperture, nd fruit-zone nd mient PAR shown in Fig. 3, ove. When verged cross ll mornings (AM) from , pre-loom removl of eight (PB-8) nd four (PB-4) leves incresed est-side erry temperture y n verge of 2.3 C, ut west-side erry temperture y n verge of only 0.6 C, when compred to pre-loom removl of no leves (PB-NO) (Fig. 4 A). When verged cross ll fternoons (PM) from , PB-8 nd PB-4 incresed west-side erry temperture y n verge of 2.3 C, ut est-side erry temperture y n verge of only 0.5 C, when compred to PB-NO (Fig. 4 C). Lef removl hd little effect on erry temperture t NOON. When verged cross NOON time periods from , pre-loom lef removl incresed oth verge est- nd west-side erry temperture y 0.5 C, when compred to PB-NO (Fig. 4 B). Pre-loom lef removl to the extent of either four or eight leves significntly incresed mnully-mesured est-nd west-side erry temperture during ll times of dy nd on severl different dtes over (Fig. 4 A-C). Pre-loom lef removl of eight leves incresed erry temperture compred to PB-4 on only 14 of the 102 occsions tht erry temperture ws mnully mesured during over (Fig. 4 A-C). Interestingly, the most common time tht PB-8 hd greter erry temperture thn PB-4 ws when west-side erry temperture ws 143

160 mesured in the AM, which occurred on hlf of the 14 occsions tht erry tempertures differed. By comprison, there were only three occsions tht PB-8 hd greter est-side erry temperture thn PB-4 during the PM. The difference in mnully-mesured erry temperture etween PB-8 nd PB-4 ws never greter thn 2.0 C, nd frequently no greter thn 0.7 C. Pre-loom removl of eight leves incresed mnully-mesured erry temperture more frequently thn PB-4 during time periods tht the sun ws cst on the opposite cnopy side (Fig. 4 A, C). For exmple, when west-side erry temperture ws mesured in the AM, PB-8 incresed erry temperture on 12 dtes, nd PB-4 on only two dtes, over the course of (Fig. 4 A). When est-side erry temperture ws mesured in the PM, PB-8 incresed erry temperture on nine dtes, nd PB-4 gin on only two dtes, over the course of (Fig. 4 C). During NOON, however, est-nd west-side erry temperture ws incresed y PB- 8 nd PB-4 on virtully the sme numer of dtes over the course of (Fig. 4 B). In 2013, PB-8 nd PB-4 incresed est-side erry temperture in the AM on two dtes y rnge of C nd C, respectively, when compred to PB-NO (Fig. 4 A). When compred to PB-NO on the west-cnopy side, PB-8 incresed erry temperture in the PM on five dtes y rnge of C, nd PB-4 incresed erry temperture in the PM on two dtes y rnge of C (Fig. 4 C). In 2014, PB-8 incresed est-side erry temperture in the AM on five dtes y rnge of C, nd PB-4 incresed est side erry temperture in the AM y 1.7 C on one dte, when compred to PB-NO (Fig. 4 A). When compred to PB-NO on the west cnopy side, oth PB-8 nd PB-4 incresed erry temperture in the PM on five dtes y rnge of C nd C, respectively (Fig. 4 C). Lstly, in 2015, oth PB-8 nd PB-4 incresed est-side erry temperture in the AM on five of the six dtes y rnge of C nd C, respectively, when compred to PB-NO 144

161 (Fig. 4 A). When compred to PB-NO on the west cnopy side, oth PB-8 nd PB-4 incresed erry temperture in the PM on ll six dtes y rnge of C nd C (Fig. 4 C). Pre-loom lef removl occsionlly incresed erry temperture y rnge of only C (2013), 0.3 C 1.7 C (2014), or C (2015) when the fruit-zone did not receive direct sunlight (Fig. 4 A-C). 145

162 Mnully-mesured erry temperture ( C) A B C W PB-8 EAST PB-4 EAST PB-NO EAST PB-8 WEST PB-4 WEST PB-NO WEST Amient PAR E W W c W PB-8 EAST PB-4 EAST PB-NO EAST PB-8 WEST PB-4 WEST PB-NO WEST Amient PAR E W E PB-8 EAST PB-4 EAST PB-NO EAST PB-8 WEST PB-4 WEST PB-NO WEST E W W E W E W W E W c W E W E W W E E W c E W W E W W W E Amient PAR E W c 8-Jul 25-Jul 5-Aug 29-Aug 24-Sep 10-Jul 23-Jul 20-Aug 28-Aug 4-Sep 16-Sep 29-Jun 13-Jul 31-Jul 13-Aug 25-Aug 8-Sep Fig. 4 Logged mient PAR, nd mnully-mesured erry temperture in the AM (A, ), NOON (B, ), nd PM (C, ) on severl dtes over the growing sesons. PB-8, PB-4, PB-NO = pre-loom removl of eight, four, nd no leves, respectively. Dt points re n verge of 36 tempertures; n = 6. Est (E) nd west (W) erry temperture mens within dte not shring letter re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. E W E W E W W E W W W E W E E W E E W W E W W E W W Logged mient photosyntheticlly ctive rdition (PAR, mol m -2 s -1 ) 146

163 Post-fruit set lef removl incresed mnully-mesured erry temperture to comprtively greter mgnitude thn pre-loom lef removl did (Fig. 5 A-C). When verged cross ll mornings, removl of six leves (PFS-6) incresed est-side erry temperture y 4.1 C, nd west-side erry temperture y only 0.5 C compred to removl of no leves (PFS-NO) (Fig. 5 A). When verged cross ll fternoons, PFS-6 incresed west-side erry temperture y 3.7 C nd est-side erry temperture y only 0.5 C compred to PFS-NO (Fig. 5 C). Lef removl tretment hd little effect on erry temperture t NOON. When verged cross ll NOON mesurements, PFS-6 incresed est- nd west-side erry temperture y n verge of 0.5 C when compred to PFS-NO (Fig. 5 B). In 2014, PFS-6 incresed est-side erry temperture in the AM on five of the six dtes y rnge of C when compred to PFS-NO (Fig. 5 A). Post-fruit set removl of six leves incresed west-side erry temperture in the PM on ll six dtes y rnge of C when compred to PFS-NO (Fig. 5 C). Over the course of ll direct-sun time periods (est in AM, west in PM) in 2014, post-fruit set lef removl further incresed est- nd west-side erry temperture y 5% compred to pre-loom lef removl (dt not shown). In 2015, PFS-6 incresed est-side erry temperture in the AM on five of the six dtes y rnge of C when compred to PFS-NO (Fig. 5 A). When compred to PFS-NO on the west cnopy side, PFS-6 incresed erry temperture in the PM on ll six dtes y rnge of C (Fig. 5 C). Over the course of ll direct-sun time periods (est in AM, west in PM) in 2015, post-fruit set lef removl further incresed est-side erry temperture y 4%, nd west-side erry temperture y 5% compred to pre-loom lef removl (dt not shown). By contrst, PFS-6 occsionlly incresed erry temperture y rnge of only C (2014) or C (2015) when erry temperture ws mesured during non-direct sun time periods (Fig. 5 A-C). 147

164 Mnully mesured erry temperture ( C) A B C E W PFS-6 EAST PFS-NO EAST PFS-6 WEST PFS-NO WEST Amient PAR E W PFS-6 EAST PFS-NO NO PFS-6 WEST PFS-NO WEST Amient PAR W E E W E W W E W W E E W W E E W W E W W E E W W E E W E W W E W W E E W W E E W W E E W W Logged mient photosyntheticlly ctive rdition (PAR, mol m -2 s -1 ) PFS-6 EAST PFS-NO EAST PFS-6 WEST PFS-NO WEST Amient PAR Jul 23-Jul 20-Aug 28-Aug 4-Sep 16-Sep 29-Jun 13-Jul 31-Jul 13-Aug 25-Aug 8-Sep Fig. 5 Logged mient PAR, nd mnully-mesured erry temperture in the AM (A, ), NOON (B, ), nd PM (C, ) on severl dtes over the 2014 nd 2015 growing sesons. PFS-6, PFS-NO = post-fruit set removl of six nd no leves, respectively. Dt points re n verge of 72 tempertures; n = 6. Est (E) nd west (W) erry temperture mens within dte not shring letter re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error

165 Berries were heted 30 or35 C for more time etween (est cnopy side) nd (west cnopy side) thn during the middle of the dy ( ) during the post-verison period (Fig. 6). Berries tended to e heted 30 nd 35 C for more time in PB- 4/8 plots compred to PB-NO plots. Berries tended to e heted 30 C for more time in 2013 nd 2015 compred to 2014, nd 35 C for more time in 2015 compred to 2013 nd Berry tempertures 35 C were never logged in PB-NO, nor 30 C outside the diurnl period. Berries in PFS-6 plots were estimted to spend reltively more time 30 nd 35 C compred to the PB-4 nd PB-8 plots (dt not shown): 101 hrs 30 C nd 11 hrs 35 C on the est cnopy side, nd 86 hrs 30 C nd 11 hrs 35 C on the west cnopy side in 2014; 117 hrs 30 C nd 27 hrs 35 C on the est cnopy side, nd 110 hrs 30 C nd 33 hrs 35 C on the west cnopy side in

166 Amient PAR ( mol m -2 s -1 ) Hrs. > 30 C Hrs. > 35 C A C1 EAST Amient PAR Amient PAR to fruit-zone ~solr noon (1330) WEST PB-8 EAST 25 C2 PB-4 EAST PB-NO EAST 20 PB-8 WEST 15 PB-4 WEST PB-NO WEST Fig. 6 Diurnl mient PAR nd percent PAR trnsmission to the fruit-zone logged over (A), nd lef removl tretment effect on logged erry temperture hours ove 30 (B1, C1, D1) nd 35 (B2, C2, D2) deg. C in the post-verison period of 2013 (B), 2014 (C), nd 2015 (D). Amient PAR trnsmission to fruit-zone (%) Hrs. > 35 C Hrs. > 30 C Hrs. > 30 C Hrs. > 35 C PB-8 EAST D2 PB-4 EAST 25 PB-NO EAST 20 PB-8 WEST 15 PB-4 WEST PB-NO WEST B1 B2 PB-8 EAST PB-4 EAST PB-NO EAST PB-8 WEST PB-4 WEST PB-NO WEST D1

167 Crop yield components nd vine fruitfulness: When compred to PB-NO in 2013, PB-8 generlly reduced components of crop yield to greter extent thn did PB-4. Crop yield, cluster weight, nd erry numer per cluster were reduced y PB-4 (30, 43, nd 44%, respectively) when compred to PB-NO (Tle 5). Crop yield, cluster weight, erry numer per cluster, nd erry weight were reduced y PB-8 (62, 62, 57, nd 9%, respectively) when compred to PB-NO, nd crop yield nd cluster weight were reduced y PB-8 (46 nd 33%, respectively) when compred to PB-4. Crop lod ws reduced y PB-4 nd PB-8 when compred to PB-NO. In 2014, crop yield nd components were reduced in similr fshion s in Crop yield, cluster weight, nd erry numer per cluster were reduced y PB-4 (53, 51, nd 48%, respectively) nd PB-8 (69, 65, nd 59%, respectively), when compred to PB-NO. As in 2013, only PB-8 reduced erry weight (14%) compred to PB-NO. Cluster weight nd erry numer per cluster were reduced y PB-4 13re (41, nd 47%, respectively) nd crop yield, cluster weight, nd erry numer per cluster were reduced y PB-8 13re (50, 58, nd 59%), when compred to PB-NOest. As in 2013, oth PB-4 nd PB-8 reduced crop lod when compred to PB-NO. Pre-loom lef removl hd similr effects on crop yield components in 2015, ut to greter extent thn the previous two sesons. Becuse no crop yield component differed etween PB-4/8 13re nd PB-4/8 14re, verges of these tretments were used for comprison to PB-NO. Crop yield, cluster weight, nd erry numer per cluster were reduced y PB-4 (51, 53, nd 51%, respectively) nd y PB-8 (78, 78, nd 73%, respectively) when compred to PB-NO. As in the previous two sesons, erry weight ws only reduced y PB-8 y 19% when compred to PB- NO. Only PB-8 reduced crop yield when compred to PB-NO. 151

168 Tle 5. Pre-loom nd post-fruit set lef removl effects on crop yield components, crop lod, nd count nd sl shoot fruitfulness from Tretment Crop yield (kg/ vine) Cluster numer Cluster weight (g) Berry # /cluster Berry weight (g) Crop lod Fruitfulness (count/sl) PB-NO n/ PB n/ PB c c n/ Significnce c < < n/ PB-NO / PB / 1.11 PB / 1.01 Significnce c ns < ns / ns PB-NO-est n/ PB-4 13re n/ PB-8 13re c 40 c n/ Significnce c ns < < ns ns n/ PFS-NO n/ PFS n/ Significnce c ns ns ns ns ns ns n/ PB-NO / 0.33 PB-4 14re 2.25 c / 0.36 PB-8 14re 1.00 d c 22 c / 0.19 PB-4 13re / 0.33 PB-8 13re 1.09 cd c 22 c / 0.21 Significnce c < ns < < < ns / ns PFS-NO / PFS / 0.25 Significnce c ns ns ns ns < ns ns / ns 2013: PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively; 2014: PB-NO (est.) = estimted yield of PB-NO vines y dding ck hrvest weight of smpled erries throughout seson; PB-4 13re, PB-8 13re = reimplementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in 2013; 2015: PB-4 14re, PB-8 14re = reimplementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in PFS-NO, PFS-6 = post-fruit set removl of no nd six leves, respectively. Presented s cluster numer per shoot; count = one-yer old spur-originting shoot, sl = cordon-originting shoot. Fruitfulness ssessed in yer presented, ut effects ttriuted to previous seson s lef removl. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments) nd Student s T-test (PFS tretments).

169 Pre-loom lef removl tended to further reduce erry numer per cluster, cluster weight, nd crop yield over time when nlyzed s percent reduction compred to PB-NO (Fig. 7). Yield components were generlly reduced to greter solute mount due to PB-8 compred to PB-4. In 2013, erry numer per cluster, cluster weight, nd crop yield ws reduced y PB-4 (44, 43, nd 29%, respectively) nd PB-8 (57, 61, nd 61%, respectively) compred to PB-NO (Fig. 7 A, B, C). The lck of linerity etween these components nd crop yield in PB-4 ppered to e function of the greter cluster numer (see Tle 4). In 2015, erry numer per cluster, cluster weight, nd crop yield were ll reduced y PB-4 (51, 53, nd 48%, respectively) nd y PB-8 (73, 77, nd 76%, respectively). Thus, pre-loom lef removl implementtion in three consecutive yers resulted in erry numer per cluster, cluster weight, nd crop yield to e further reduced y PB-4 (7, 10, nd 19%, respectively) nd y PB-8 (16, 16, nd 15%, respectively). In 2014, erry numer per cluster, cluster weight, nd crop yield were ll reduced y PB-4 (48, 51, nd 54%, respectively) nd y PB-8 (58, 65, nd 65%, respectively) compred to PB-NO (Fig. 7 D, E, F). In 2015, erry numer per cluster, cluster weight, nd crop yield were ll reduced y PB-4 (50, 54, nd 54%, respectively) nd y PB-8 (72, 77, nd 77%, respectively). Thus, pre-loom lef removl implementtion in two consecutive yers resulted in erry numer per cluster, cluster weight, nd crop yield to e only slightly further reduced in PB-4, nd further reduced y PB-8 (14, 12, nd 12%, respectively). 153

170 % reduction in erry # / cluster A PB-8 '13-15 PB-4 '13-15 D PB-8 '14-15 PB-4 '14-15 % reduction in cluster weight B PB-8 '13-15 PB-4 '13-15 E PB-8 '14-15 PB-4 ' C F % reduction in crop yield PB-8 '13-15 PB-4 '13-15 PB-8 '14-15 PB-4 ' Fig. 7. The effect of re-implementtion of pre-loom removl of eight (PB-8) nd four (PB-4) leves in 2013, 2014, nd 2015 (A, B, nd C) nd in 2014 nd 2015 (D, E, nd F) on percent reduction in erry numer per cluster (A, D), cluster weight (B, E), nd crop yield (C, F) compred to removl of no leves; n = 6. Error rs re +/- stndrd error. NOTE: 2014 dt is not presented in A, B, or C due to non-uniform erry smpling from PB-NO, PB-4, nd PB-8. Sesonl erry weight development: Berry weight tended to e reduced y PB-8 ut not y PB-4 or PFS-6 (Fig. 8). In 2013, PB-8 reduced erry weight when compred to oth PB-NO nd PB-4 on 9-Jul (13%), 30-Aug (13 %), nd 9-Oct (10%), nd when compred to PB-4 on 24-Jul (9%) (Fig. 8 A). Averge erry weight increse from 24 Jul to 9 Aug ws 0.02 g nd from 9 Aug to 154

171 30 Aug ws 0.02 g nd 0.06 g greter in PB-NO nd PB-4 compred to PB-8, respectively. In 2014, PB-8 reduced erry weight when compred to oth PB-NO nd PB-4 on 15 Jul (8%; p > F = ) nd 17 Oct (11%; p > F = ), nd only when compred to PB-NO on 14 Aug (17%), 12 Sep (21%), nd 30 Sep (20%). Berry weight increse from 31 Jul to 14 Aug nd from 14 Aug to 12 Sep ws 0.13 g nd 0.14 g more in PB-NO compred to PB-8, respectively. In 2015, PB-8 reduced erry weight on ll dtes y n verge of 15% when compred to the verge erry weight of PB-NO nd PB-4 (Fig. 8 C). Berry weight ws reduced y PB-4 y 8% when compred to PB-NO on 8 Sep only. Re-implementtion of pre-loom lef removl tretments in the third consecutive seson (PB-4/8 13re) did not further reduce erry weight. Berry weight increse from 2 Aug to 8 Sep ws 0.15 g greter in PB-NO, nd 0.7 g greter in PB-4, when compred PB

172 Berry weight (g) A B Jul 24-Jul 9-Aug 30-Aug 9-Oct PB-8 PB-4 PB-NO 15-Jul 31-Jul 14-Aug 12-Sep 30-Sep 17-Oct PB-8 PB-4 PB-NO 1.6 C D Berry weight (g) c c 30-Jun 16-Jul 2-Aug 8-Sep 5-Oct PB-8 PB-4 PB-NO PB-8 '13re PB-4 '13re 14-Aug 12-Sep 30-Sep 17-Oct PFS-6 PFS-NO Fig. 8 Pre-loom (A, B, C) nd post-fruit set (D) lef removl effect on erry weight over time in 2013 (A), 2014 (B, D), nd 2015 (C); Ech dt point is n verge of 120 erries; n = 6. Tretment men erry weight within dte not shring letter re significntly different ( 0.05) using Tukey's HSD. Error rs re +/- stndrd error. Components of cluster compctness: Cluster compctness ws reduced y oth PB-8 (57%) nd PB-4 (25%) when compred to PB-NO in 2013 (Tle 6). Pre-loom removl of eight leves (PB-8) reduced erry numer per cluster (58%) to greter extent thn did PB-4 (35%). The disproportionlly lesser reduction in cluster compctness reltive to erry numer per cluster ws due to shorter rchis lengths in PB-4 in Rchis length ws not ffected y lef removl tretment in 2014, nd erry numer per cluster nd cluster compctness were reduced similrly y PB-8 (59% nd 60%, respectively) nd PB-4 (42% nd 39%, respectively) when compred to PB-NO in In 2014, erries were smpled on five fewer occsions in re-implemented tretments (PB-4/8 13re) compred to current-seson tretments, resulting in no difference in cluster compctness compred to PB-NO. When compred to PB-NO in 2015, PB-4/PB-4 13re 156

173 reduced erry numer per cluster y n verge of 49% nd PB-4 reduced cluster compctness y 38%, nd PB-8/PB-8 13re reduced erry numer per cluster y n verge of 74% nd cluster compctness y n verge of 69%. As in 2013, the greter reduction in erry numer per cluster reltive to cluster compctness in PB-4/PB-4 13re plots ws due to shorter rchis lengths, prticulrly in PB-4 13re. There ws direct, positive reltionship etween reduction in erry numer per cluster nd cluster compctness, ut concomitnt reduction in rchis length confounded this reltionship. Tle 6. Pre-loom lef removl effect on components of cluster compctness from Cluster Tretment Berry # / cluster Rchis length (cm) Compctness (erry # / cm rchis length) PB-NO PB PB-8 34 c c Significnce c < < PB-NO PB-4 34 cd c PB-8 24 d c PB-4 13re PB-8 13re 41 c Significnce c < ns <

174 2015 PB-NO PB c PB-8 21 c cd PB-4 13re PB-8 13re 20 c d Significnce c < < : PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively; 2014: PB-4 13re, PB-8 13re = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in 2013; 2015: PB-4, PB-8 = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in PFS-NO nd PFS-6 = post-fruit set removl of no nd six leves, respectively. NOTE: Clusters smpled non-uniformly in 2014 six times totl in PB-NO/4/8 vines nd only t hrvest in PB-4/8 13revines. c Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments). Botrytis unch rot nd unch stem necrosis incidence: Botrytis unch rot (BBR) incidence ws greter when no leves were removed from the fruit-zones (PB-NO, PFS-NO) compred to ny of the lef removl tretments (PB-4, PB-8, PFS-6) (Tle 7). Though not sttisticlly compred, witing until the post-fruit set period to remove leves incresed the incidence of BBR when compred to removing leves efore loom. Bunch stem necrosis incidence ws not ffected y tretment. Tle 7. Botrytis unch rot (BBR) nd unch stem necrosis (BSN) incidence s ffected y pre-loom nd post-fruit set lef removl in Tretment BBR (%) BSN (%) PB-NO PB PB PB-4 '13re PB-8 '13re

175 Significnce < ns PFS-NO PFS Significnce ns 2013: PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively; 2014: PB-4 13re, PB-8 13re = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in 2013; 2015: PB-4, PB-8 = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in PFS-NO nd PFS-6 = post-fruit set removl of no nd six leves, respectively. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments) nd Student s T-test (PFS tretments). Primry fruit chemistry: Pre-loom removl of eight leves (PB-8) reduced solule solids y 5% when compred to PB-4 in 2013, nd oth current-seson PB-8, nd its re-implementtion (PB- 8 13re), reduced solule solids y n verge of 4% when compred to PB-NO in 2014 (Tle 8). All pre-loom lef removl tretments significntly reduced TA y n verge of 15% when compred to PB-NO in Post-fruit set removl of six leves (PFS-6) reduced ph y 2-3% nd TA y 15-19% in oth yers, nd solule solids y 1% in 2015, when compred to PB-NO. Tle 8. Pre-loom nd post-fruit set lef removl effects on juice solule solids concentrtion (SSC), ph, nd titrtle cidity (TA) in Tretment SSC TA SSC TA SSC TA ph ph ph ( Brix) (g/l) ( Brix) (g/l) ( Brix) (g/l) PB-NO PB PB PB-4 13re n/ n/ n/ PB-8 13re n/ n/ n/ Significnce ns ns ns ns ns ns PFS-NO n/ n/ n/ PFS-6 n/ n/ n/ Significnce n/ n/ n/ ns < : PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively; 2014: PB-4 13re, PB-8 13re = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in 2013; 2015: PB-4, PB-8 = re-implementtion of PB-4 nd PB-8, respectively, on sme vines initilly used in PFS-NO nd PFS-6 = post-fruit set removl of no nd six leves, respectively. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments) nd Student s T-test (PFS tretments). 159

176 Totl erry phenolics nd nthocynins: Pre-loom removl of four leves (PB-4) incresed totl erry nthocynins y n verge of 9% when compred to PB-NO nd PB-8 in 2014 (Tle 9). In 2015, PB-4 nd PB-8 incresed totl erry nthocynins y n verge of 22% when compred to PB-NO. Pre-loom lef removl incresed totl erry phenolics y n verge of 14% compred to PB-NO in oth 2013 nd In 2015, PB-8 incresed totl erry phenolics y 10% compred to PB-4 nd 37% compred to PB-NO; PB-4 incresed totl erry phenolics y 25% compred to PB-NO. Post-fruit set removl of six leves (PFS-6) incresed totl erry phenolics y 13% in 2014 nd 16% in 2015 when compred to PFS-NO. In 2015, PFS-6 incresed totl erry nthocynins y 16% when compred to PFS-NO. Tle 9. Pre-loom nd post-fruit set lef removl effects on totl erry nthocynins (TBA) nd phenolics (TBP) on the est nd west cnopy sides in Lef removl nd cnopy side TBA (mg/g erry) TBP (u/g erry) TBA (mg/g erry) TBP (u/g erry) TBA (mg/g erry) TBP (u/g erry) PB-NO c PB PB EAST WEST Significnce Lef removl (LR) ns < < < Cnopy side (CS) ns ns ns ns ns ns LR*CS ns ns ns ns ns ns PFS-NO n/ n/ PFS-6 n/ n/ EAST n/ n/ WEST n/ n/

177 Significnce n/ n/ Lef removl (LR) n/ n/ ns Cnopy side (CS) n/ n/ ns ns ns ns LR*CS n/ n/ ns ns ns ns PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively. In 2013 nd 2014, current-seson lef removl tretments were used; in 2015, re-implemented tretments from 2014 were used. PFS-NO nd PFS-6 = post-fruit set removl of no nd six leves, respectively. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments) nd Student s T-test (PFS tretments). There ws no reltionship etween erry weight t hrvest nd totl erry phenolics nd nthocynins when evluted y comining the three yers of the pre-loom lef removl study (Fig. 9). Rther, sesons showed grouping of sorts, with 2014 hving the gretest concentrtions of these compounds in grpes cross ll erry weights (Fig. 9 A, E). A higher percentge of dt points from 2013 nd 2015 fell elow the 95% confidence intervl towrd the lower nd higher rnge of erry weight, respectively. There ws no reltionship etween erry weight nd totl erry phenolics nd nthocynins in 2013 (Fig. 9 B, F). While there ws no reltionship etween erry weight nd totl erry nthocynins in 2014, there ws significnt, (p > F = ) negtive, liner reltionship of erry weight nd totl erry phenolics in 2014 (Fig. 9 C, G). There ws significnt, negtive, liner reltionship etween erry weight nd oth totl erry nthocynins (p > F = ) nd phenolics (p > F = ) in 2015 (Fig. 9 D, 161

178 H). A reltively high percentge of dt points from PB-4 fell ove the 95% confidence intervl in oth totl erry phenolics nd nthocynins plots in 2014 nd 2015 due to n increse in metolite concentrtion without concomitnt decrese in erry weight. The grouping of tretment dt points ws well-defined only in the totl erry phenolics reltionship in 2014, nd in oth totl erry phenolics nd nthocynins reltionships 2015: PB-8 hd lower erry weight nd higher metolite concentrtion, nd vice-vers for PB-NO. 162

179 A y = x R 2 = E y = x R 2 = Liner regression 95% conf idence interv l Liner regression 95% conf idence interv l B y = x R 2 = F y = x R 2 = Totl erry nthocynins (mg / g erry) C PB-8 PB-4 PB-NO Liner regression Totl erry phenolics (u / g erry) 95% conf idence interv l 40 PB-8 PB-4 PB-NO Liner regression y = x 95% conf idence interv l R 2 = G PB-8 PB-4 PB-NO Liner regression 95% conf idence interv l PB-8 PB-4 PB-NO Liner regression 95% conf idence interv l y = x R 2 = D y = x R 2 = H y = x R 2 = PB-8 PB-4 PB-NO Liner regression 95% conf idence interv l PB-8 PB-4 PB-NO Liner regression 95% conf idence intervl Berry weight (g) Berry weight (g) Fig. 9 The liner reltionship etween hrvest erry weight nd totl erry nthocynins (A-D) nd phenolics (E-H) over (A, E), nd s ffected y pre-loom lef removl tretment in 2013 (B, F), 2014 (C, G), nd 2015 (D, H). PB-8, PB-4, nd PB-NO = pre-loom removl of eight, four, nd no leves, respectively. Ech dt point represents n verge of 120 erries; n =

180 There ws no reltionship etween erry weight t hrvest nd totl erry phenolics nd nthocynins when evluted y comining the two yers of the post-fruit set lef removl study (Fig. 10). Sesons were grouped, with 2014 hving lower metolite concentrtions s well s lower erry weights, wheres 2015 hd comprtively higher metolite concentrtions cross roder rnge of erry weights (Fig. 10 A, E). There ws no reltionship etween erry weight nd totl erry phenolics nd nthocynins in 2014, the yer tht post-fruit set lef removl did not ffect erry weight (Fig. 10 B, E). There ws significnt, negtive, liner reltionship etween erry weight nd oth totl erry nthocynins (p > F = ) nd phenolics (p > F = ) in 2015 (Fig. 10 C, F), the yer tht PFS-6 reduced erry weight compred to PFS-NO. In 2015, tretment dt point groups were well-defined, with PFS-6 hving lower erry weight nd higher metolite concentrtion compred to PFS-NO. 164

181 A y = x R 2 = D y = x R 2 = Liner regression 95% conf idence intervl Liner regression 95% conf idence intervl Totl erry nthocynins (mg / g erry) B y = x R 2 = Totl erry phenolics (u / g erry) PFS-6 PFS-NO 40 Liner regression 95% conf idence intervl 35 E y = x R 2 = PFS-6 PFS-NO Liner regression 95% conf idence interv l C y = x R 2 = F y = x R 2 = PFS-6 PFS-NO Liner regression 95% conf idence interv l PFS-6 PFS-NO Liner regression 95% conf idence interv l Berry weight (g) Berry weight (g) Fig. 10 The liner reltionship etween hrvest erry weight nd totl erry nthocynins (A-C) nd phenolics (D-F) over (A, D), nd s ffected y post-fruit set lef removl of six (PFS-6) nd no (PFS-NO) leves in 2014 (B, E) nd 2015 (C, F). Ech dt point represents n verge of 120 erries; n =

182 Discussion The gol of these studies ws to evlute if exposing fruit to greter extent nd/or t n erlier time thn currently recommended would fvorly lter yield components nd fruit composition of Cernet Suvignon. Most components of crop yield were reduced to greter extent y pre-loom removl of eight compred to four leves, while post-fruit set lef removl hd miniml effect on components of yield. Fruit-zone lef removl tended to increse totl grpe phenolics nd nthocynins, regrdless of extent or timing of tretment execution. Berry temperture nd rdition: Comprehensive studies nd reviews hve documented tht temperture nd light re two of the most importnt meteorologicl determinnts of fruit composition, wine qulity, nd vine growth nd helth (Bergqvist et l. 2001, Dokoozlin nd Kliewer 1995, Downey et l. 2006, Jckson nd Lomrd 1993, Smrt nd Roinson 1991, Spyd et l. 2002). Yet, populr techniques developed to quntify cnopy porosity nd the mount of incident rdition reching the fruit-zone do not quntify erry temperture (Smrt nd Roinson 1991, Meyers nd Vnden-Heuvel 2008). It ws therefore justified to chrcterize mient ir temperture nd rdition ptterns, nd quntify how these relted to erry temperture nd, ultimtely, composition of grpes from oth shded nd exposed fruit-zones. Rdition trnsmission to the fruit-zone ws required to het erries ove mient ir temperture, s shown efore (Bergqvist et l. 2001). This ws temporl phenomenon, however, given the diurnl chnge in solr ngle nd the decrese in fruit-zone rdition round solr noon. Trends of the current study showed tht the percent rdition trnsmission to the fruit-zone typiclly declined s the solr ngle trcked closer to noon due to the overhed cnopy (Fig. 3). As such, the diurnl period of gretest rdint heting potentil of grpes my e voided in verticlly-shoot positioned vineyrds. These findings contrst Bergqvist et l. (2001), 166

183 who reported 7-10 C difference in erry temperture etween exposed nd shded cluster t middy, nd no middy depression in the difference etween the temperture of mient ir nd exposed erries, lthough tht study ws conducted in est-west oriented rows. It ws possile tht cnopies were sprser, nd likely tht sunlight ws more intense, in the reltively wrmer/sunnier climte of the centrl Sn Joquin Vlley of Cliforni compred to the northern Shenndoh Vlley of Virgini in which the current study ws conducted. The regionl contrst in climte etween these two studies mke mient rdition n importnt considertion s vrile when modelling erry temperture, prticulrly under vrile sky conditions (Col et l. 2009). Direct sunlight ws shown to increse grpe temperture to greter extent thn indirect sunlight (Bergqvist et l. 2001), reveling the sptiotemporl influence of solr ngle on erry temperture in given trining system. The most ggressive fruit-zone lef removl tretment frequently incresed erry temperture when the sun ws cst on the opposite cnopy. However, ecuse direct light hets plnt tissues to greter extent thn diffuse light (Smrt nd Sinclir 1976), erry temperture ws gretest when sun ws cst on the est cnopy side in the AM nd on the west cnopy side in the PM. During times of dy tht the sun ws not locked y the ove-hed cnopy, 4-5% further increse in erry temperture ws oserved in post-fruit set lef removl plots compred to pre-loom lef removl plots. This ws puttively function of erries wrming to greter mgnitude when touching neighoring erries in compct clusters (Keller et l. 2010), phenomenon tht might e mitigted with reduced cluster compctness fforded y pre-loom lef removl. As such, pre-loom lef removl my e useful tool to increse the mount of surfce re of erries tht is exposed to ir movement nd, thus, ultimtely reduce erry heting. 167

184 Mnully mesured erry temperture often showed sttisticl seprtion, ut questionle iologicl importnce, nd were merely point-in-time oservtions. Lef removl cn iologiclly impct nthocynins if erry tempertures routinely exceed C (Bergqvist et l. 2001, Mori et l. 2007, Spyd et l. 2002, Trr et l. 2008). Although temperture is known to ffect nthocynin development, it ws uncler whether sustined periods of elevted temperture were more or less importnt thn shorter periods of possily higher temperture (Trr et l. 2008). Assuming time spent C impcted grpe nthocynin concentrtions, the est determinnt for tretment impct ws logged erry temperture. Lef removl resulted in erry temperture ccumultion for 90 hrs 30 C ut only 10 hrs 35 C, when verged cross pre-loom removl extent nd cnopy side over three post-verison periods ( ). Comprtively, erry tempertures were logged for 163 hrs 30 C nd 63 hrs 35 C when verged cross cnopy side over three post-verison periods under the more rid conditions used y Trr et l. (2008). This comprison shows the reltively shorter time spent t these criticl temperture thresholds in the current study, even when Trr et l. (2008) did not employ fruit-zone lef removl, which likely would hve further incresed the mount of time erry temperture ws 30 nd 35 C. The comprison of erry temperture results etween Trr et l. (2008) nd the current study revels tht climte lrgely drives erry temperture. There my e less difference etween exposed nd shded grpes in wrm nd cool climtes (Winkler et l. 1974), s cited in Bergqvist (2001), who reported point-in-time mient ir tempertures 35 C t middy. By contrst, mient temperture never exceeded 33.6 C in the course of the current study. The mount of time tht erry temperture ws 30 or 35 C did not gretly differ etween est nd west cnopy sides in this study. While the mount of time tht erry 168

185 temperture ws 30 C ws not different, erries spent 52 less hrs 35 C on the est compred to west cnopy side in Trr et l. (2008). Amient rdition in this study ws often ttenuted in the fternoon compred to the morning (Figs. 2 nd 3), generlly s function of greter fternoon cloudiness. Though west-exposed erries hd greter verge erry temperture, rdition ppered to influence erry temperture on the west-side to lesser degree thn it did on the est-side. Whether this ws due to greter cloud coverge in the fternoon, or tht erries t greter current temperture re less susceptile to rdint heting-induced temperture increse, is eyond the scope of this study. However, the net effect ws less of cnopy-side difference in time spent t or ove criticl temperture thresholds in this study compred to others conducted in drier growing regions (Bergqvist et l. 2001, Spyd et l. 2002, Trr et l. 2008). Vrile cloudiness, prticulrly in the fternoon hours, is more common in Virgini thn in the rid West. Thus, in humid growing regions, it my e erroneous to refer to specific cnopy side, prticulrly the west cnopy side of north-south oriented rows, s the hot cnopy side. Rther, the cnopy side receiving more sunlight my vry from seson-to-seson in humid regions. Components of yield: The close reltionship etween percent reduction in crop yield nd percent reduction in erry numer per cluster nd cluster weight reveled tht pre-loom lef removl reduced crop yield vi reduction in erry numer per cluster more so thn reduction in erry weight or cluster numer per vine. While fruit set ws not mesured in our study, the sustntil reduction in erry numer per cluster in pre-loom plots ws highly indictive of reduced set. Yield reduction ws perhps function of source-induced cron deficit tht incurred efore, nd ws mintined throughout, flowering (Cndolfi-Vsconcelos nd Kolet 1990, Cspri nd Lng 1996, Coome 1962). Further, loom-time wether is frequently cool, cloudy, nd/or riny 169

186 in the estern US, conditions tht re unfvorle for pollen viility nd germintion rtes (Kolet 1966). Optimum temperture for pollen germintion is etween 25.0 nd 30.0 C, wheres temperture elow 10 or ove 35 C prevent germintion (Keller 2010). Further, pollen tue growth is slowed elow C (Keller 2010). Averge temperture in the week surrounding loom ws 19.7 C in 2013 nd 2014, nd 16.7 C in Berry numer per cluster ws ccordingly lower in 2015 compred to 2013 nd Thus, source tissue reduction nd su-optiml germintion tempertures t loom my hve reduced fruit set to greter extent thn would e expected to occur in regions with reltively wrmer loom periods. Reduction in crop yield ws consistent response to pre-loom lef removl (Digo et l. 2012, Pllioti et l. 2011, Pstore et l. 2013, Poni et l. 2006, 2009, Gtti et l. 2012, Intrieri et l. 2008, Trdguil et l nd 2012). In those previous studies, pre-loom lef removl reduced fruit set y 8-37%, erry numer per cluster y 14-64%, cluster weight y 20-69%, nd crop yield y 30-71%. Those rod response rnges represent different lef removl techniques, such s mnul (Gtti et l. 2012, Pllioti et l. 2011) nd mechnicl pre-loom defolition (Trdguil et l. 2012), or oth (Intrieri et l. 2008, Trdguil et l. 2010), nd were conducted in potted nd field settings (Poni et l. 2006), nd in severl cultivrs, including Temprnillo (Digo et l. 2012, Trdguil et l. 2012), Sngiovese (Gtti et l. 2012, Intrieri et l. 2008, Pllioti et l. 2011, Poni et ll. 2006), Grcino nd Crignn (Trdguil et l. 2010), Treino (Poni et l. 2006), nd Brer nd Lmrusco slmino (Poni et l. 2009). Only one other preloom lef removl study ws known to e conducted in region nlogous to the current study - Michign, US. (Stini nd Howell 2010). In the Michign study, pre-loom lef removl reduced fruit set, erry numer per cluster, cluster weight, nd crop yield. Effects were more pronounced in Vignoles, French-Americn hyrid, response tht ws ttriuted to the 170

187 stronger shoot growth of Vignoles reltive to Pinot noir nd Pinot gris (Stini nd Howell 2012). Cernet Suvignon, used in the current study, is n inherently vigorous vriety. As such, the strong shoot growth my hve served s strong sink competing with the flowering nd fruit set process. Few other thn Stini nd Howell (2010) studied the severity of pre-loom lef removl on yield components. In tht study, there ws strong ( ) positive reltionship etween the numer of leves removed t trce loom nd the extent of cluster weight reduction in different vrieties (Stini nd Howell 2010). When yers were comined in the current study, there ws positive liner reltionship (R 2 = ) etween the extent of preloom lef removl nd the reduction in erry numer per cluster, cluster weight, nd crop yield (dt not shown). When evluted y yer, this reltionship ws stronger in 2013 nd 2015 (R 2 = ) thn in 2014 (R 2 = ) (dt not shown). This ws likely ecuse there ws less solute reduction in crop yield components due to pre-loom lef removl in 2014 compred to 2013 nd The reltionship etween pre-loom lef removl extent nd crop yield hd lower R 2 vlues compred to erry numer per cluster nd cluster weight. This ws likely due to the dditionl influence of erry weight nd cluster numer per vine on this metric. Our results suggest tht there my not lwys e strong liner reltionship etween pre-loom source tissue removl nd reduction in fruit set/cluster weight, nd this reltionship my vry etween vrieties, nd cross growing sesons nd regions. The extent of fruit set reduction my e more proportionl to the percent of leves retined, rther thn the percent of leves removed, efore loom. This reltionship ws not mesured in the current study, nd would e difficult to quntify given the rpid increse in lef re during loom nd fruit set. 171

188 Pre-loom lef removl further reduced crop yield when re-implemented over the following two sesons. However, when re-implemented once (2015), pre-loom removl of eight, ut not four, leves further reduced crop yield. Reduction in crop yield ws prlleled y reduction in erry numer per cluster nd cluster weight. While some studies oserved no further reduction in crop yield over multiple yers (Poni et l. 2006, Intrieri et l. 2008), others oservtions were similr to the current study (Gtti et l. 2012, Pllioti et l. 2011, Stini nd Howell 2010, Trdguil et l nd 2012). For exmple, fruit set ws further reduced y 24 nd 88%, nd crop yield y 55 nd 83% due to re-implementtion of pre-loom removl of four nd six leves, respectively (Stini nd Howell 2010). If ggressive pre-loom lef removl is nnully conducted on the sme vines, crop yield my continue to decline due to reduced erries per cluster nd cluster weights (Gtti et l. 2012, Pllioti et l. 2011), nd s function of reduced set (Stini nd Howell 2010). Reduced crohydrte sttus cn limit fruit set (Cndolfi-Vsconcelos nd Kolet 1990, Cspri nd Lng 1996, Coome 1962). Accordingly, the further reduction in erry numer per cluster over time oserved in the current study ws possily indictive of lower crohydrte sttus in the pre-loom lef removl vines. Crohydrte reserves in the permnent vine prts cn e continully depleted due to reduced source tissues during the previous-seson s fruit ripening period s well s efore the flowering/fruit set period of the current-seson (Cndolfi- Vsconcelos nd Kolet 1990). Pre-loom lef removl decresed pruning weight (Pllioti et l. 2011), potentilly due to pre-loom source tissue limittion s well s the continul depletion of storge reserves in permnent vine prts (Pllioti et l. 2010). In the current study, reimplementtion of pre-loom lef removl further reduced pruning weights over-time, ut to 172

189 significnt within-seson extent due only with pre-loom removl of eight leves. Therefore, repeted pre-loom lef removl cn reduce vine cpcity. Defolition of 50% of lef re reduced finl erry size when conducted 12 dys fter flowering, ut hs reduced effect when executed t 35 nd 58 dys fter flowering (Kliewer 1970). The reduction in erry nd pericrp cell size ws likely due to reduced ssimilte supply (Ollt nd Gudillere 1998). Yet, the smller erry size my hve lso een due to decresed trnsloction of lef-derived hormones to the grpes from erly defolited tretments (Kliewer 1970). Erly lef removl my not completely limit erry growth potentil (Ollt nd Gudillere 1998), s erry growth rtes were similr etween defolited nd folited vines during the lg nd second growth phse (Kliewer 1970). The current study confirms tht erly nd ggressive source tissue removl cn limit erry size, ut witing until severl weeks fter fruit set hs little impct on erry weight. The most ggressive pre-loom lef removl tretment in the current study reduced the rte of erry weight gin, especilly during the lg nd second growth phse. This delyed erry weight gin ws nlogous to tht of the low source tissue tretment implemented immeditely fter fruit set in Ollt nd Gudillere (1998). Since erry cell division occurs during the first two weeks fter flowering (Jon nd Bott 1988) it is possile tht witing until 12 dys fter flowering hd little effect on erry cell division, resulting in less difference in erry weight gin during the lg nd second growth phse, s in Kliewer et l. (1970). It ws likely tht only the most ggressive pre-loom lef removl tretment in the current study concomitntly reduced erry cell size nd cell numer enough to result in n initil, nd sustined, erry weight reduction throughout the seson. It ws speculted tht tht the preloom lef removl threshold for reducing erry size ws etween four nd eight leves. 173

190 While some pre-loom lef removl studies reported reduction in erry weight (Gtti et l. 2012, Pllioti et l. 2011, Trdguil et l. 2010), others did not (Intrieri et l. 2008, Pstore et l. 2013, Poni et l. 2009, Stini nd Howell 2010, Trdguil et l nd 2012). The ltter group cited tht erry growth ws compenstory due to reduction in erry numer per cluster (Intrieri et l. 2008, Poni et l. 2009) nd the null effect tht erly mechnicl defolition hd on erry weight ws function of source tissue limittion efore cell division in erry growth stge I (Trdguil et l. 2012). Since inflorescence primordi re initited one yer efore they flower, erly lef removl cn reduce vine fruitfulness (cluster numer / unit of mesure) nd, thus, crop yield in oth the current- nd following seson (Cndolfi-Vsconcelos nd Kolet 1990). Pre-loom lef removl reduced current-seson fruitfulness y 6%, nd susequent-seson fruitfulness y 21-30% (Stini nd Howell 2010, Trdguil et l. 2012). However, in the current study, s well s others (Intrieri et l. 2008, Gtti et l. 2012, Pllioti et l. 2011, Pstore et l. 2013), preloom lef removl did not reduce fruitfulness. While limiting source tissues t/efore loom in the previous seson my negtively ffect ud fertility, the increse in rdition trnsmission to next yer s fruiting uds cn improve fruitfulness (Perez nd Kliewer 1990, Snchez nd Dokoozlin 2005), potentilly negting the impct of source tissue limittion on ud fertility. There is likely threshold t which the lef re reduction negtively impcts ud fertility nd fruitfulness, ut this will likely vry s function of timing nd physiologicl sttus of the vines t the time of lef removl. Besides smll reduction in erry weight in one of two yers, crop yield components were not negtively ffected y post-fruit set lef removl, even to the ggressive extent of six leves removed. This response ws nticipted, however, s fruit set nd crop yield were 174

191 unffected y lef removl t four weeks post-loom (Cndolfi-Vsconcelos nd Kolet 1990), nor due to lef removl s erly s t fruit set (Bledsoe et l. 1988). While not evluting lef removl extent, lef removl tht occurred further fter loom reduced crop yield to lesser extent thn did lef removl performed closer to loom (Cndolfi-Vsconcelos et l. 1990). Results of our pre-loom nd post-fruit set lef removl experiments support tht crop yield cn e regulted y the timing nd extent of lef removl (Intrieri et l. 2008, Kliewer 1970, Pstore et l. 2013, Trdguil et l. 2010) primrily due to the sensitive period surrounding loom, t which time fruit set is limited y crohydrte supply to the flowering/setting cluster (Cndolfi- Vsconcelos nd Kolet 1990, Cspri nd Lng 1996, Coome 1962). Primry fruit chemistry: Severl pre-loom lef removl studies oserved incresed solule solids (Digo et l. 2012, Gtti et l. 2012, Intrieri et l. 2008, Pllioti et l. 2011, Pstore et l. 2013, Poni et l. 2006, 2009, nd 2010, Stini nd Howell 2010, Trdguil et l. 2012). Preloom lef removl incresed solule solids due to: greter lef re: fruit weight rtios (Gtti et l. 2012, Trdguil et l nd 2012), regulted vegettive vigor, improved cnopy nd wter use efficiency, nd regulted crop lod (Pllioti et l. 2011), nd restored nd more efficient lef re due to lterl re-growth in the post-verison period, nd hstened ssimilte trnsloction to the fruit (Intrieri et l. 2008, Pstore et l. 2013, Poni et l nd 2009); the lst found efore (Kolet et l. 1993, Quinln nd Wever 1970). Most of the ove-cited studies were conducted in drier regions, such s Itly (Gtti et l. 2012, Intrieri et l. 2008, Poni et l nd 2009, Pllioti et l. 2011) nd Spin (Digo et l. 2012, Trdguil et l. 2012), ll of which likely hve drier nd sunnier ripening periods thn typiclly oserved in humid regions. 175

192 In the current study, pre-loom removl of eight leves reduced solule solids, while removl of four nd six leves did not, regrdless of when they were removed. Crop lod tended to e lower in pre-loom lef removl plots, ut lso reltively low in the trditionl sense in ll other tretments (Brvdo et l. 1985). Though lef re ws not mesured in the current study, fruit-zone nd lterl re-growth were eliminted to minimize fruit-zone shding, perhps limiting the pool of crohydrtes ville for fruit ripening, even though remining leves on defolited shoots re more efficient ssimiltors (Buttrose 1966, Kliewer et l. 1970, Cndolfi-Vsconcelos nd Kolet 1991, Poni et l. 2006). Becuse of the supr-optiml wter vilility in humid regions, ggressive lef removl my not hve reduced vegettive vigor to point tht crohydrte competition with ripening fruit ws llevited. Therefore, the source sink lnce ws unfvorle for sugr trnsloction to ripening grpes potentilly due to growing shoot tips serving s competitive crohydrte sink (Winkler et l. 1974). Erly-seson (pre-loom, immeditely post-fruit set) lef removl cn dely ripening onset due to n initil lef re deficit. Ripening ws delyed prticulrly when leves were removed fter fruit set to result in smller erries s sinks (Ollt nd Gudillere 1998). Genetic expression nlyses indicted tht pre-loom nd verison-time lef removl cn result in dely in the trnscriptionl ripening progrm of erries (Pstore et l. 2013). However, erry solule solids cn resume ccumulting when source tissues re restored (Ollt nd Gudillere 1998, Pstore et l. 2013). In the current study, crop level nd erry size my hve een reduced so much in the PB-8 plots tht tht sink strength nd ssimilte trnsloction to fruit ws reduced (Ollt nd Gudillere 1998). The comined effects of (1) mintined lef re deficit from preloom through hrvest, (2) smll crop sink/strength, nd (3) su-optiml ssimiltion rtes during ripening my hve resulted in reduced solule solids. Given the inconsistent reduction of 176

193 solule solids in PB-8 plots cross ll yers, it is possile tht the source-sink mechnisms tht mitigted solule solids ccumultion in this tretment my hve een slightly offset y the lesser solute sugr required to rech given concentrtion in smller erries (Ollt nd Gudillere 1998). Nonetheless, ggressive pre-loom lef removl my e useful tool to dely sugr ccumultion in regions prone to hstened sugr ccumultion, nd in vrieties prone to high Brix, such s Petit mnseng or Frontenc. This tctic my result in fruit with etter enologicl lnce of primry fruit chemistry nd vrietl chrcter from secondry metolites. In the current study, pre-loom lef removl inconsistently reduced titrtle cidity (TA), while post-fruit set removl of leves reduced TA more consistently, nd to reltively greter mgnitude, thn pre-loom lef removl. Pre-loom lef removl reduced TA in some cses (Pstore et l. 2013, Poni et l. 2009, Trdguil et l. 2012), ut either incresed (Poni et l. 2006), or did not ffect it in others (Gtti et l. 2012, Intrieri et l. 2008, Pllioti et l. 2011). Though genetic nlysis suggested tht trtric cid synthesis in sun-exposed grpes from preloom lef removl plots ws prolonged, higher erry tempertures likely incresed mlic cid degrdtion, ultimtely resulting in lower titrtle cidity compred to shded grpes (Pstore et l. 2013). Interestingly, two pre-loom lef removl studies tht did not oserve TA reduction lso permitted the re-growth of lterls from sl nodes (Intrieri et l. 2008, Pllioti et l. 2010). In those studies, lterl lef shding my hve mitigted the temperture-driven mlic cid respirtion frequently oserved in exposed fruit-zones (Jckson nd Lomrd 1993, Kliewer nd Schultz 1964), emphsizing the importnce of reporting fruit-zone rchitecture, s y cluster exposure flux vilility nd/or lef lyer numer. Pre-loom lef removl s vrile effect on TA ws potentilly due to incresed trtric cid synthesis (Gtti et l. 2012, Poni et l. 2006), s greter cron ccumultion into trtric cid occurs in sun versus shde erries (Kliewer nd 177

194 Schultz 1964). Further, the optimum temperture rnge for orgnic cid nd mlic cid ccumultion is etween 20 nd 25 C, nd mlic cid levels decline ove 38 C (Kliewer 1964, Lkso nd Kliewer 1975). During the pre-verison period of orgnic cid ccumultion in grpes (Kliewer 1965), verge logged erry temperture in oth pre-loom lef removl nd no lef removl plots ws within optiml rnge ( C). Thus, in humid climtes, the temperture of exposed nd shded erries during pre-verison my e optiml, or t lest not inhiitory, for orgnic cid synthesis; this my not e true in wrmer climtes. Unlike mlic cid, trtric cid degrdes very little throughout erry ripening (Ruffner, 1982), nd the pprent reduction in trtrte levels is ctully dilution due to wter nd sugr influx into the erry (Johnson nd Crroll 1973). Tht pre-loom lef removl hd less consistent effects on TA thn post-fruit set lef removl in the current study my e explined y comintion of the lst few points. The reltively erlier exposure of grpes in the pre-loom plots my hve enhnced orgnic cid ccumultion in grpes compred to the severl-week dely in grpe exposure in the post-fruit set lef removl plots. Furthermore, ny high temperture-induced mlic cid degrdtion my hve een offset y the lck of trtric cid dilution in smller erries from the pre-loom removl of eight lef-plots. Totl erry phenolics nd nthocynins: Totl phenolics were incresed every yer due to lef removl, regrdless of timing or mgnitude. While the mechnisms were not mesured, greter fruit-zone rdition nd the concentrting effect of smller erries were likely cndidtes, the ltter prticulrly in the pre-loom removl of eight leves tretment. There ws stronger negtive reltionship etween erry weight nd phenolics thn erry weight nd nthocynins. However, if decrese in erry weight ws solely responsile for n increse in erry phenolics, then phenolics should hve een incresed only y PB-8. This ws not the cse, however, s PB- 178

195 4 lso incresed phenolics in ll yers, ut did not retrd erry weight increse. The 2015 results, where phenolics were greter in PB-8 nd PB-4 compred to PB-NO, ut lso greter in PB-8 compred to PB-4, were interpreted s concentrtion hving prtil, ut not sole, role of incresing erry phenolics. In this prticulr cse, it ws speculted tht reltively higher incident rdition to the fruit-zone eqully incresed phenolics in oth PB-4 nd PB-8, nd the greter phenolics oserved in PB-8 plots ws due to the dditionl concentrtion effect. Other studies similrly found tht totl erry phenolics were incresed with pre-loom lef removl (Intrieri et l. 2008, Poni et l. 2006, Trdguil et l nd 2012), leit inconsistently t times (Digo et l. 2012). Mechnisms ehind incresed grpe phenolics were incresed skin: pulp rtios (Poni et l. 2006), thicker erry skins (Pllioti et l. 2011, Poni et l. 2008, Trdguil et l. 2010), greter nd more efficient lef re: fruit weight rtios (Digo et l. 2012, Intrieri et l. 2008, Gtti et l. 2012, Poni et l. 2006), nd the positive effects of sun exposure on grpe phenolics (Gtti et l. 2012, Trdguil et l. 2012) such s skin flvonols nd nthocynins (Pstore et l. 2013). Phenolics include severl flvonoid nd non-flvonoid suclsses, the former suclss contining compounds tht ccumulte t different times during erry development nd re differentilly ffected y environmentl conditions such s light nd temperture (Downey et l. 2006). It ws, therefore, difficult to pinpoint specific environmentl effect or mechnism ehind the incresed grpe phenolics oserved in lef removl plots. Nonetheless, lef removl t either the pre-loom or t the post-fruit set stge incresed rdition to the fruit-zone, nd incresed erry phenols, correltive response reported in other studies s well (Jckson nd Lomrd 1993, Price et l. 1995, Smrt nd Roinson 1991). Other fctors, such s erry skin thickening, greter lef re: fruit weight rtios, more 179

196 efficient remining-lef ssimiltion, nd/or hstened ssimilte trnsloction to clusters might lso hve een responsile for dditionl increses in erry phenolics. Removl of fruit-zone leves efore loom (Digo et l. 2012, Gtti et l. 2012, Intrieri et l. 2008, Pllioti et l. 2011, Poni et l. 2009, Trdguil et l nd 2012) nd fter loom (Digo et l. 2012; Kotseridis et l. 2012, Poni et l. 2006) hs incresed nthocynins. There were severl postulted mechnisms ehind incresed nthocynins in pre-loom lef removl plots: greter nd more efficient lef re nd greter lef re: fruit weight rtios (Gtti et l. 2012, Intrieri et l. 2008, Pllioti et l. 2011, Digo et l. 2012), lterl shoot-shding of the fruitzone (Intrieri et l. 2008, Pllioti et l. 2011), modifictions to the flvonoid iosynthetic pthwy (Pstore et l. 2013), greter skin: pulp rtios (Poni et l. 2006), nd greter erry skin mss (Pstore et l. 2013, Poni et l nd 2010), regrdless of erry size (Poni et l. 2008). Yet, the enefit of open fruit-zones could not e ruled out, even though fruit-zone light nd erry tempertures were not well-chrcterized (Gtti et l. 2012, Kotseridis et l. 2012, Pllioti et l. 2011, Trdguil et l. 2012). None of these postulted mechnisms tht incresed grpe nthocynins were mesured in the current study. The inconsistent reltionship oserved etween erry weight nd nthocynin concentrtion in the current study suggests tht nthocynins were not simply incresed in concentrtion due to erry weight reductions. The incresed nthocynins oserved here with ggressive lef removl ws contrst to the well-documented reduction in nthocynin synthesis (Ymne et l. 2006) or increse in degrdtion (Mori et l. 2007) due to excessive erry tempertures ttined when grpes re exposed to the sun (Bergqvist et l. 2001, Downey et l. 2006, Spyd et l , Trr et l. 2008). Severl studies oserved decresed grpe nthocynins t C, the commonly cited temperture thresholds tht re detrimentl to 180

197 nthocynin ccumultion (Bergqvist et l. 2001, Mori et l. 2007, Spyd et l. 2002, Trr et l. 2008, Ymne et l. 2006). mrna levels of flvonoid iosynthetic pthwy enzyme genes nd the VvmyA1 trnscription fctor, responsile for enhnced expression of genes in the flvonoid iosynthetic pthwy, were lower in grpes held t 30 compred to 20 C (Ymne et l. 2006). However, in pple, it ws shown tht similr trnscription fctor responsile for nthocynin iosynthesis in red-skinned pple cultivrs ws induced y light exposure (Tkos et l. 2006). These studies collectively showed tht light is necessry, ut extreme tempertures deleterious, for nthocynin iosynthesis nd ccumultion. Thus, the generl mechnism of incresed nthocynins in the current study ppered to e greter fruit-zone rdition coupled with sun-exposed erries tht spent less time t criticl temperture thresholds compred to other studies (Spyd et l. 2002, Trr et l. 2008, Pstore et l. 2013). Previous studies hve reported the durtion of time tht erries spent ove criticl temperture thresholds for grpe nthocynin degrdtion (Spyd et l. 2002, Trr et l. 2008). Working in the sunny Ykim Vlley of Wshington Stte, Trr et l. (2008) found tht exposed erries tht hd lower nthocynins spent 166 hrs 30 C nd 89 hrs 35 C during the post-verison period (Trr et l. 2008). However, nthocynins were reduced when shded grpes were rtificilly heted for comprtively less time (142 hrs 30 C nd 15 hrs 35 C) during the post-verison period (Trr et l. 2008). In nother study in the Ykim Vlley, sunexposed erries on the west cnopy side tht spent hrs > 30 C nd hrs > 35 C from unch closure to hrvest hd lower totl grpe color nd nthocynins thn est-exposed erries tht spent hrs > 30 C nd 3-8 hrs > 35 C from unch closure to hrvest. In study conducted in Bologn, Itly, erry tempertures 30 C were logged for ~ 300 hrs in oth pre-loom nd verison lef removl tretments, ut nthocynins were incresed in the pre- 181

198 loom lef removl nd decresed in the verison lef removl tretment compred to nondefolited control (Pstore et l. 2013). Comprtively, in the most extreme cses of the current study, exposed grpes on oth cnopy sides spent only hrs 30 C nd hrs 35 C, or when estimted in other tretments, hrs 30 C nd hrs 35 C. The two weeks following color chnge is the criticl period of susceptiility for temperture-induced color loss in red grpes (Ymne et l. 2006). The results of the current study suggest tht erry temperture must e 30 C for more thn 57 hrs nd/or 35 C for more thn 12 hrs during the two weeks following color chnge, s these were the most extreme conditions experienced y erries tht did not experience reduced nthocynins. Collectively, the ove studies suggest tht sun-exposed erries cn e sujected to tempertures 30 nd 35 C for longer period of time thn shded erries efore nthocynin ccumultion is inhiited or degrdtion is incresed. If shded erries re simultneously heted, they re more susceptile to nthocynin degrdtion thn re exposed erries (Spyd et l. 2002, Trr et l. 2008). However, the results of the current study generlly showed tht shded grpes were seldom heted ove mient temperture, t lest in humid, frequently cloudy region in the estern US. The more relevnt temperture for reducing grpe nthocynins ppers to e 35 C rther thn 30 C, s nthocynins were not reduced when erries spent 270 hrs 30 C in non-defolited tretment (Pstore et l. 2013). Further, west-exposed erries tht hd lower nthocynins spent more hrs 35 C, nd only 7-33 more hrs 30 C, when compred to est-exposed erries (Spyd et l. 2002, Trr et l. 2008). It is unknown, however, if comintion of these two tempertures is etter determinnt of grpe nthocynins thn either lone. As shown in the current study nd others (i.e. Pstore et l. 2013), the timing of lef removl nd therefore the history of erry exposure my hve 182

199 ering on the fte of nthocynin ccumultion nd degrdtion. To this effect, pre-loom lef removl pprently promotes nthocynins reltive to removing leves lter, either due to the cooling effect of looser clusters, or other mechnisms, such s the promotion of erry skin growth (Poni et l. 2006). However, in mny prts of the estern US, even witing until fter fruit set to remove leves will likely result in fruit experiencing comintion of moderte tempertures nd incresed sunlight, eneficil comintion for grpe nthocynins (Trr et l. 2008). Becuse nthocynins nd phenolics tended to e incresed, sl lef removl in redfruited vrieties in humid regions (1) cn only enefit fruit composition; (2) need not e s conservtive s currently prcticed; (3) need not e focused on the est cnopy side, nd (4) my not result in sunurn, prticulrly if lef removl occurs efore loom (Pstore et l. 2013). Conclusion: Post-fruit set lef removl did not reduce crop yield nd concurrently incresed erry phenolics nd nthocynins nd decresed lte seson unch rot incidence. Aggressive, pre-loom lef removl is not recommended in humid regions unless crop yield reduction is cceptle. Removl of no more thn four leves efore loom would e recommended strting point. On the other hnd, sl lef removl need not e s conservtive s once thought due to frequent cloud cover nd moderte tempertures experienced in humid compred to drier regions. Thus, fruit-zones cn e mintined with fewer lef lyers thn currently recommended in order to concomitntly improve disese mngement nd increse totl grpe phenolics nd nthocynins. This work offers refinement for uiquitous prctice tht hs received little modifiction nd ttention in humid growing regions over the lst two decdes. Future work ought to investigte long-term, cultivr-specific responses to pre-loom lef removl cross severl humid meso-climtes, such s those in south nd estern Virgini, North Crolin, Georgi, Tennessee, nd Missouri. 183

200 Literture cited Bergqvist, J., N.K. Dokoozlin, nd N. Eisud Sunlight exposure nd temperture effects on erry growth nd composition of Cernet Suvignon nd Grenche in the centrl Sn Joquin Vlley of Cliforni. Am. J. Enol. Vitic. 52: 1-7. Bledsoe, A.M., W.M. Kliewer, nd J.J. Mrios Effects of timing nd severity of lef removl on yield nd fruit composition of Suvignon lnc grpevines. Am. J. Enol. Vitic Bogicevic, M., V. Mrs, M. Mugos, V. Kodzulovic, J. Ricevic, S. Sucur, nd O. Fill The effect of erly lef removl nd cluster thinning tretments on erry growth nd grpe composition in cultivrs Vrnc nd Cernet Suvignon. Chem Biol Techn Agric 2: 1-8. Brvdo, B., Y. Hepner, C. Loinger, S. Cohen, nd H. Tcmn Effect of crop level nd crop lod on growth, yield, must nd wine composition, nd qulity of Cernet Suvignon. Am. J. Enol. Vitic. 36: Buttrose, M.S The effect of reducing lef re on the growth of roots, stems, nd erries of Gordo grpe vine. Vitis 5: Cndolfi-Vsconcelos, M.C., nd W. Kolet Yield, fruit qulity, ud fertility nd strch reserves of the wood s function of lef removl in Vitis vinifer Evidence of compenstion nd stress recovering. Vitis 29: Cronneu, A The erly selection of grpevine rootstocks for resistnce to drought conditions. Am. J. Enol. Vitic. 36: Cspri, H.W., nd A. Lng Crohydrte supply limits fruit-set in commercil Suvignon lnc grpes. In: Proceedings for the Fourth Interntionl Symposium on Cool Climte Enology nd Viticulture. T. Henick-Kling et l., Eds. New York Stte Agriculturl Experiment Sttion, pp Chorti, E., S. Guidoni, A. Ferrndino, nd V. Novello Effect of different cluster sunlight exposure levels on ripening nd nthocynin ccumultion in Neiolo grpes. 61: Col, G., O. Fill, nd L. Mrini BerryTone A simultion model for the dily course of grpe erry temperture. Agric. For. Meteorol. 149: Coome, B.G The effect of removing leves, flowers, nd shoot tips on fruit set in Vitis vinifer L. J. Hortic. Sci. 37: Digo, M.P., B. Ayestrn, Z. Gudlupe, S. Poni, nd J. Trdguil Impct of preloom nd fruit set sl lef removl on the flvonol nd nthocynin composition of Temprnillo grpes. Am. J. Enol. Vitic. 63: Di Profio, F., A.G. Reynolds, nd A. Ksimos Cnopy mngement nd enzyme impcts on Merlot, Cernet frnc, nd Cernet Suvignon. II. Wine composition nd qulity. Am. J. Enol. Vitic. 62: Dokoozlin, N.K., nd W.M. Kliewer The light environment within grpevine cnopies. II. Influence of lef re density on fruit zone light environment nd some cnopy ssessment prmeters. Am. J. Enol. Vitic. 46: Downey, M.O., N.K. Dokoozlin, nd M.P. Krstic, M.P Culturl prctice nd environmentl impcts on flvonoid composition of grpes nd wine: review of recent reserch. Am. J. Enol. Vitic. 57:

201 Dry, P., nd B. Coome, eds Viticulture 1 Resources, 2 nd Ed. Revised version of grpevine growth stges The modified E-L system. Winetitles, Adelide, Austrli. English, J.T., C.S. Thoms, J.J. Mrois, nd W.D. Guler Microclimtes of grpevine cnopies ssocited with lef removl nd control of Botrytis unch rot. Phytopthology 79: Gtti, M., F. Bernizzoni, S. Civrdi, nd S. Poni Effects of cluster thinning nd preflowering lef removl on growth nd grpe composition in cv. Sngiovese. Am. J. Enol. Vitic. 63: Hed, B Reltionship etween cluster compctness nd unch rot in Vignoles grpes. Plnt Dis. 93: Hunter, J., O.T. de Villiers, nd J.E. Wtts The effect of prtil defolition on qulity chrcteristics of Vitis vinifer L. cv. Cernet Suvignon grpes II. Skin colour, skin sugr nd wine qulity. Am. J. Enol. Vitic. 1991, 42: Intrieri, C, I. Filippetti, G. Allegro, M. Centinri, nd S. Poni Erly defolition (hnd vs mechnicl) for improved crop control nd grpe composition in Sngiovese (Vitis vinifer L.). Aus. J. Grpe Wine Res. 14: Jckson, D.I., nd P.B. Lomrd Environmentl nd mngement prctices ffecting grpe composition nd wine qulity A Review. Am. J. Enol. Vitic. 44: Johnson, L.A., nd D.E. Crroll Orgnic cid nd sugr contents of scuppernong grpes during ripening. J. Food. Sci 38: Jon, R., nd R. Bott Fruit set nd erly erry development in two grpevine cultivrs. Isrel J. Bot. 307: Keller, M The Science of Grpevines: Antomy nd Physiology, 1 st Ed. Acdemic Press. Kliewer WM Influence of environment on metolism of orgnic cids nd crohydrtes in Vitis vinifer. I. Temperture. Plnt Physiol. 39: Kliewer, W.M., nd H.B. Schultz Influence of environment on metolism of orgnic cids nd crohydrtes in Vitis vinifer. II. Light. Am. J. Enol. Vitic Kliewer, W.M Chnges in the concentrtion of trtrtes, mltes, nd totl free cids in flowers nd erries of Vitis vinifer L. Am. J. Enol. Vitic. 16: Kliewer, W.M Effect of time nd severity of defolition on growth nd composition of Thompson seedless grpes. Am. J. Enol. Vitic. 21: Kolet, W Berry set of grpe vines relted to shoot tretments nd climtic fctors. Wein- Wiss 21: Kolet, W. M.C. Cndolfi-Vsconcelos, E, Aeschimnn, nd G.S. Howell Influence of defolition, rootstock, nd trining system on Pinot noir grpevines. I. Moilistion nd reccumultion of ssimiltes in woody tissue. Wein-Wiss. 48: Kotseridis, Y., A. Georgidou, P. Tikos, S. Kllithrk, nd S. Koundours Effects of severity of post-flowering lef removl on erry growth nd composition of three red Vitis vinifer L. cultivrs grown under semirid conditions. J. Agric. Food Chem. 60: Lee, S-H., M.J. Seo, M. Riu, J.P. Cott, D.E. Block, N.K. Dokoozlin, nd S.E. Eeler Vine microclimte nd norisoprenoid concentrtion in Cernet Suvignon grpes nd wines. Am. J. Enol. Vitic. 58: Meyers, J.M., nd J.E. Vnden Heuvel Enhncing the precision nd sptil cuity of point qudrt nlysis vi clirted exposure mpping. Am. J. Enol. Vitic. 59:

202 Mori, K., S. Sugy, nd H. Gemm Decresed nthocynin iosynthesis in grpe erries grown under elevted night tempertures. Sci. Hortic. 105: Ollt, N., nd J.P. Gudillere The effect of limiting lef re during stge I of erry growth nd development nd composition of erries of Vitis vinifer L. cv. Cernet Suvignon. Am. J. Enol. Vitic. 49: Pllioti, A., M. Gtti, S. Poni Erly lef removl to improve vineyrd efficiency: gs exchnge, source-to-sink lnce, nd reserve storge responses. Am. J. Enol. Vitic. 62: Pstore, C., S. Zenoni, M. Fsoli, M. Pezzotti, G. Bttist Tornielli, nd I. Filipetti Selective defolition ffects plnt growth, fruit trnscriptionl ripening progrm nd flvonoid metolism in grpevine. BMC Plnt Biol 13:1-16. Perez, J., nd W.M. Kliewer Effect of shding on ud necrosis nd ud fruitfulness of Thompson seedless grpevines. Am. J. Enol. Vitic. 41: Poni, S., L. Cslini, F. Bernizzoni, S. Civrdi, nd C. Intrieri Effects of erly defolition on shoot photosynthesis, yield components, nd grpe composition. Am. J. Enol. Vitic. 57: Poni, S., F. Bernizzoni, S. Civrdi, nd N. Lielli Effects of pre-loom lef removl on growth of erry tissues nd must composition in two red Vitis vinifer L. cultivrs. Aust. J. Grpe Wine Res. 15: Poni, S., F. Bernizzoni, S. Civrdi, nd N. Lielli Effects of pre-loom lef removl on growth of erry tissues nd must composition in two red Vitis vinifer L. cultivrs. Aust. J. Grpe Wine Res. 15: Poni, S., nd F. Bernizzoni A three-yer survey on the impct of pre-flowering lef removl on erry growth components nd grpe composition in cv. Brer vines. J. Int. Sci Vigne Vin. 44: Price, S.F., P.J. Breen, M. Vlldo, nd B.T. Wtson Cluster sun exposure nd quercetin in Pinot noir grpes nd wine. Am. J. Enol. Vitic. 46: Quinln, J.D., nd J.R. Wever Modifiction of pttern of the photosynthetic movement within nd etween shoots of Vitis vinifer L. Plnt Phys. 46: Reynolds A.G., D.A. Wrdle, nd A.P. Nylor Impct of trining system, vine spcing, nd sl lef removl on Riesling vine performnce, erry composition, cnopy microclimte, nd vineyrd lor requirements. Am. J. Enol. Vitic. 47: Ruffner, H.P. Metolism of trtric nd mlic cid in Vitis: review Prt B. Vitis 21: Ryon, I., B.S. Pn, D.S. Intrigliolio, A.N. Lkso, nd G.L. Scks Effects of cluster light exposure on 3-isoutyl-2-methoxypyrzine ccumultion nd degrdtion ptterns in red wine grpes (Vitis vinifer L. cv. Cernet Frnc). J. Agric. Food Chem. 56: Stini, P., nd G.S. Howell Effects of erly defolition on yield, fruit composition, nd hrvest seson cluster rot complex of grpevines. HortScience 45: Snchez, L.A., nd N.K. Dokoozlin Bud microclimte nd fruitfulness in Vitis vinifer L. Am. J. Enol. Vitic. 56: Smrt, R.E., nd T.R. Sinclir Solr heting of grpe erries nd other sphericl fruits. Agric. Meteorol. 17: Smrt, R.E., J.G. Roinson, G.R. Due, nd C.J. Brien Cnopy microclimte modifiction for the cultivr Shirz II. Effects on must nd wine composition. Vitis 24:

203 Smrt, R., nd M. Roinson Sunlight into Wine. A Hndook for Winegrpe Cnopy Mngement. Winetitles, Adelide, Austrli. Smith S., I. Codrington, M. Roertson, nd R.E. Smrt Viticulturl nd oenologicl implictions of lef removl for New Zelnd vineyrds. In: Proc. Second Int. Cool Climte Viticulture nd Oenology Symp., Aucklnd, NZ. R.E. Smrt, R.J. Thorton, S.B. Rodriguez nd J.E. Young Eds. pp New Zelnd Society of Enology nd Viticulture, Aucklnd. Spyd, S., J.M. Trr, D.L. Mee, nd J.C. Ferguson Seprtion of sunlight nd temperture effects on the composition of Vitis vinifer cv. Merlot erries. Am. J. Enol. Vitic. 53: Stff, S.L., D.C. Percivl, J.A. Sullivn, nd K.H. Fisher Fruit zone lef removl influences vegettive, yield, disese, fruit composition, nd wine sensory ttriutes of Vitis vinifer L. Optim nd Cernet frnc. Cn. J. Plnt Sci. 77: Tkos, A.M., F.W. Jffe, S.R. Jco, J. Bogs, S.P. Roinson, nd A.R. Wlker Light- Induced Expression of MYB Gene Regultes Anthocynin Biosynthesis in Red Apples. Plnt Physiol. 142: Trr, J., J.M. Lee, S.E. Spyd, nd C.F. Scgel Berry temperture nd solr rdition lter cyltion, proportion nd concentrtion of nthocynins in Merlot grpes. Amer. J. Enol. Vitic. 59: Trdguil, J., F. Fernndo Mrtinez de Tod, S. Poni, nd M.P. Digo Impct of erly lef removl on yield nd fruit nd wine composition of Vitis vinifer L. Grcino nd Crignn. Amer. J. Enol. Vitic. 61: Trdguil, J., J.A. Blnco, S. Poni, S., M.P. Digo Mechnicl yield regultion in winegrpes: comprison of erly defolition nd crop thinning. Aust. J. Grpe Wine Res. 18: Winkler, A.J., J.A. Cook, nd W.M. Kliewer Generl Viticulture, 2 nd Ed. University of Cliforni Press, Berkeley, CA. Wolf, T.K., R.M. Pool, nd L.R. Mttick Responses of young Chrdonny grpevines to shoot tipping, ethephon, nd sl lef removl. Am. J. Enol. Vitic. 37: Wolf, T.K., ed Wine grpe production guide for estern North Americ. Grpevine Cnopy mngement (A. Reynolds nd T.K. Wolf), pp Ymne, T., nd K. Shiym Effects of chnges in the sensitivity to temperture on colortion of Aki Queen grpe erries. J. Jpn Soc. Hort. Sci. 75: Zoecklein BW, T.K. Wolf, N.W. Duncn, J.M. Judge, nd M.K. Cook Effects of fruitzone lef removl on yield, fruit composition, nd fruit rot incidence of Chrdonny nd White Riesling (Vitis-vinifer L.) grpes. Am. J. Enol. Vitic. 43:

204 Preliminry ssessment of n hourly mient temperture nd rdition-driven model of grpe erry temperture under vrile sky conditions of the estern US Astrct Bckground nd ims: Light nd temperture re importnt determinnts of rom nd flvor compounds in grpes nd, thus, wines. Metrics hve een developed, nd enhnced, tht chrcterize fruit-zone rdition, quntify the physicl nture nd sptil distriution of the fruitzone, nd relte to fruit composition. However, few models re known to e ville tht cn ccurtely predict grpe temperture, even though it is ffected y oth fruit-zone rchitecture, nd mient ir temperture nd rdition. It ws sought to develop grpe tempertureprediction model to id in the meteorologicl risk ssessment of over-heting well-exposed grpes to known criticl temperture thresholds, such s those for nthocynins (30-35 C). Methods nd results: Amient nd fruit-zone photosynthetic ctive rdition (PAR), mient nd erry temperture, nd mient reltive humidity were logged over three consecutive growing sesons. Berry temperture extremes were ccounted for y logging the temperture of erries on the exterior fce of grpe clusters. The results were grower friendly models for ech 15 hour ngle tht predict erry temperture differentil from mient ir temperture from est nd west cnopy sides, in well-exposed nd shded fruit-zones, nd using mient rdition nd hour ngle. The difference etween mnully mesured nd predicted erry temperture rnged [0.17] to [2.84] C using the model. Conclusions: When fruit-zones re shded, erry temperture is highly predictle using mient ir temperture s the independent vrile. The degree to which exposed erry temperture is heted ove mient ir temperture is dependent on severl fctors, including, ut not limited to, solr rdition, diurnl hour ngle, nd current mient ir temperture. The 188

205 ility to predict erry temperture differentil from mient ir temperture ws complicted y fruit-zone lef removl prctice, cnopy side, hour ngle, nd mient rdition, prticulrly t diurnl periods of direct rdition penetrtion to the fruit-zone. Significnce of study: Hving the ility to predict erry temperture will permit growers to hve etter understnding of their site-specific risk of reching criticl grpe tempertures (i.e. for nthocynins). Growers cn djust their future fruit-zone mngement prctices ccordingly. Introduction Cnopy mngement studies hve een n ctive re of viticulture reserch since the 1960 s (Shulis et l. 1966). Since, severl studies demonstrted tht fruit exposure impcts fruit nd wine qulity. In generl, fruit exposure to sunlight positively impcts (Bledsoe et l. 1988; Cronneu 1985; Di Profio et l. 2011; Reynolds et l. 1996; Stff et l. 1997) nd shding negtively impcts (Hunter, 1991; Jckson nd Lomrd 1993; Ryon et l. 2008; Smrt et l. 1985; Smrt nd Roinson 1991) fruit nd wine qulity. But, compounds cn e ffected differentilly y fruit exposure. Highly exposed fruit cn reduce methoxypyrzines (Ryon et l. 2008), nthocynins (Spyd et l. 2002), nd cidity (Jckson nd Lomrd 1993), ut increse flvonols (Price et l. 1995; Spyd et l. 2002). Light nd temperture cn differentilly ffect grpe compound synthesis. Thus, the impct of light nd temperture on grpe composition (Dokoozlin nd Kliewer 1995), prticulrly nthocynins nd other flvonoids (Bond nd Sdrs 2015; Downey et l. 2006), ut lso voltile compounds (Roinson et l. 2014), hs een of high interest over the lst decde. Becuse exposed fruit is sujected to concurrent increse in light nd temperture, n importnt question ws if prticulr grpe compounds were ffected more or less y temperture or light. It ws determined tht temperture prticulrly ffects grpe nthocynins (Bergqvist et 189

206 l. 2001; Spyd et l. 2002; Trr et l. 2008), eing detrimentl t ~ 35 C (Downey et l. 2006; Mori et l. 2007; Trr et l. 2008). Temperture, in ddition to light, cn impct the ctivity of glycoside-producing enzymes in fruit (Gerdes et l. 2002). Further, temperture ws relted to C13-norisoprenoids (Lee et l. 2007) nd cn ffect grpe crotenoids (Bureu et l. 1998). Though temperture highly impcts grpe compounds, some studies evluting the effects of fruit exposure on sensory impct compounds nd their precursors hve not mesured grpe erry temperture (Kwsniweski et l. 2010) or not elorted on its impct, even when there ws 5 C difference in temperture etween sun-exposed nd shded clusters (Rzungles et l. 1998). Environmentl fctors cn ffect pprox. 18% of grpe genes (Dl Snto et l. 2013). Temperture, long with sunlight, re the environmentl fctors tht vry most cross vineyrd sites nd sesons nd, therefore, ccount for flvonoid differences etween sites nd seson (Downey et l. 2006). While oth temperture nd sunlight cn ffect romtic compounds nd their precursors (Gerdes et l. 2002), temperture ws only riefly mentioned in terms of its impct on rom compounds in recent review (Roinson et l. 2014). This suggests tht reltively few studies nlyzing grpe composition hve sought to mesure grpe erry temperture, potentilly ecuse it is difficult to experimentlly seprte light nd temperture effects in field setting (Bond nd Sdrs 2015, s in Spyd et l nd Trr et l. 2008), or ecuse the effect of temperture on grpe composition is not pprecited s much s the effects of sunlight. Strtegies to chieve optiml cnopies nd metrics to quntify them, clled point-qudrt nlyses (PQA), hve een used nd cited extensively over the lst 25 yers (Smrt nd Roinson 1991). Proportion of exterior (from the cnopy) fruit ws given high rting for its 190

207 impct on wine qulity, citing tht shded fruit is inferior for winemking. An enhncement of PQA, EPQA, dded photon flux ssessment, llowing quntifiction of cnopy iomss distriution, light environment, nd the efficcy of tretment (i.e. lef removl, shoot thinning) implementtion (Meyers nd Vnden-Heuvel 2008). Cluster exposure flux vilility, n EPQA metric, hd the highest predictive strength for glycosylted rom compounds in Riesling when compred to other metrics (Meyers et l. 2013). Given the impct tht the work of Smrt nd Roinson (1991) nd Meyers nd Vnden-Heuvel (2008) hs hd on shping viticulture prctice nd reserch, it is evident tht the fruit-zone microclimte cn hve greter impct on grpe composition compred to ny other environmentl feture of the vineyrd. Yet PQA or EPQA metrics do not quntify erry temperture. Indirect methods to evlute erry temperture re often imprecise (Bond nd Sdrs 2015). Furthermore, cnopy mngement prctices tilored for prticulr mcroclimte might not e pproprite in other mcroclimtes. In order to etter understnd how erry temperture ffects grpe erry composition, quntittive nd direct temperture mesures re criticl, nd should e done in the field to limit confounding effects ssocited with growth chmers or rtificil heting or cooling of fruit (Bond nd Sdrs 2015). The lst sentence of recent review on the effect of temperture on grpe composition red: Modeling remins n underdeveloped pproch to investigte erry composition in response to temperture (Bond nd Sdrs 2015). Only one other erry temperture prediction model ws known to e developed (Col et l. 2009). This model ws sed on technique clled energy lnce to simulte therml regimes of living odies, nd only required minimum nd mximum dily mient ir temperture s inputs. The uthors cited tht direct erry temperture mesurement ws costly nd complicted, nd tht other meteorologicl metrics, such s solr rdition nd 191

208 reltive humidity, were difficult to otin. However, gin, indirect methods to evlute erry temperture re inconclusive nd suject to misinterprettion (Bond nd Sdrs 2015). Temperture highly impcts fruit composition, yet reltively few fruit exposure studies hve quntified grpe temperture. Thus, we sought to develop model to predict grpe erry temperture, prticulrly y using redily ville meteorologicl metrics of mient light nd temperture s inputs. Permitting the prediction of grpe erry temperture will llow ccurte nd quntittive comprisons to e mde etween regions tht gretly differ in temperture nd sunlight conditions. Consequently, fruit-zone mngement cn ecome fine-tuned, nd more vineyrd- nd region-specific. Our gol ws to produce model tht ws perhps more prgmtic nd dt driven y tking more oots-on-the-ground, rther thn mth-driven pproch of nother erry temperture model study (in Col et l. 2009). This ws ccomplished y logging erry temperture nd meteorology t the scle of the cnopy. It ws hypothesized tht erry temperture could e predicted using only mient ir temperture nd rdition, ut tht erry temperture my lso e ffected y reltive humidity, erry color, nd cnopy-side position. All of these metrics were thus logged over the course of the study. Mterils nd Methods Tretments nd experimentl design: Two seprte timings of lef removl tretments were implemented, oth of which used Cernet Suvignon ENTAV-INRA clone 337 vines, grfted onto 420-A rootstock, nd grown t Virgini Tech s AHS, Jr. Agriculturl Reserch nd Extension Center (AHS, Jr. AREC) ner Winchester, VA (39 11 N; W). Vines were plnted in My 2006 in rows running generlly northest/southwest t 3.0-m (row) x 1.5-m (vine) spcing nd were trined to ilterl cordons with verticlly-positioned shoots. The soil ws Poplimento Hgerstown sndy lom (A. Blckurn, personl communiction, 2013). The 192

209 inter-row groundcover, estlished in 2001, initilly comprised mixture of orchrd grss (Dctylis glomert) nd tll fescue (Festuc rundince); cv. Shenndoh, with the fescue dominting fter ~ six yers. The intr-row groundcover in the pre-loom lef/lterl shoot removl plots consisted of perennil creeping red fescue (F. rur), estlished in Sep 2008, nd n 85-cm wide hericide-treted strip in the post-fruit set lef/lterl shoot removl plots. Lef/lterl shoot removl tretments were s ulleted elow: Expt. I: Pre-loom lef/lterl shoot removl tretments ( ): PB-NO - no leves/lterl shoots removed; lterl shoots mintined t ~3-4 nodes. PB-4 - removl of leves/lterl shoots from primry shoot nodes 1-4 efore loom, [modified Eichorn nd Lorenz (EL) stge (Dry nd Coome 2004)]; distl lterl shoots mintined t ~3-4 nodes. PB-8 - removl of leves/lterl shoots from primry shoot nodes 1-8 efore loom, [modified Eichorn nd Lorenz (EL) stge (Dry nd Coome 2004)]; distl lterl shoots mintined t ~3-4 nodes. Expt. II: Post-fruit set lef/lterl shoot removl tretments ( ): PFS-NO - no leves/lterl shoots removed; lterl shoots mintined t ~3-4 nodes. PFS-6 - removl of leves/lterl shoots from primry shoot nodes 1-6 t pe-erry size / unch closure (modified EL stge 31/32); distl lterl shoots mintined t ~3-4 nodes. A complete lock design ws used for ech set of lef removl timing, prtitioning the experimentl re into six locks, ech seprted y five-vine order plots within the row nd y ordering uffer rows. Within ech lock, lef removl tretments were rndomly ssigned to either one-vine experimentl units (pre-loom experiment) or two-vine experimentl units (post- 193

210 fruit set experiment). Pre-loom nd post-fruit set lef removl experimentl units were no more thn ~ 45 m wy from ech other in ech vineyrd lock. Generl vine mngement: Cordons were spur-pruned ech winter. Shoot density ws djusted, iming for ~ shoots per m of cordon. Shoots were mintined verticlly upright with the id of ctch wires. Shoots were shoot-hedged efore they extended more thn ~0.9 m ove the top ctch wire. Frequent re-visits to experimentl units were mde to ensure tht cnopy porosity ws mintined through hrvest. No vine nutritionl deficiency symptoms were evident, nd disese mngement ws stndrd for the region. Cnopy chrcteriztion: Point qudrt nlysis (PQA) dt ws collected etween EL stges 33 nd 35 (verison), s descried in Smrt nd Roinson (1991), in order to chrcterize fruit-zone rchitecture. This ws performed in the Cernet Suvignon pre-loom study from , the Cernet Suvignon post-fruit set study from A thin metl rod ws inserted into the fruiting zone long the trnsverse xis of the cnopy, using tpe mesure to guide insertions. This process ws repeted 10 times in ech experimentl unit (vine) in order to quntify fruit-zone lef lyer numer (LLN). Photosynthetic photon flux density (PPFD) ws ssessed y inserting the ceptometer inside cnopy fruit zones prllel to, nd directly ove, the cordon nd orienting the light interception side of the ceptometer in three different directions (45 est, verticl, 45 west) nd then verging those redings. Two PPFD redings were tken in every experimentl unit (one reding ove ech cordon) etween ~ hrs under consistent mient rdition conditions. An mient PPFD reding ws tken etween with ech fruit-zone reding in order to express fruit-zone PPFD s percentge of the mient rdition. This dt ws used to generte the cluster exposure flux vilility (CEFA) metric with 194

211 enhnced point qudrt nlysis (EPQA version 1.6.2) softwre (Meyers nd Vnden-Heuvel 2008). The dtlogger vine pnel: The following mesurements were logged t specified intervls in ech respective seson. 2013: erry nd mient temperture, nd fruit-zone PAR were logged on 15-min intervls from 30 Jul 2013 through 9 Oct : erry nd mient temperture, nd fruit-zone nd mient PAR were logged on 15-min intervls from 25 Jul 2014 through 4 Sep 2014, nd on 1-min intervls from 6 Sep 2014 through16 Oct 2014, nd from 22 Jun 2015 through 5 Oct : erry nd mient temperture, nd fruit-zone nd mient PAR were logged on 1-min intervls from 22 Jun 2015 through 5 Oct One experimentl unit (vine) of ech pre-loom lef removl tretment (post-fruit set lef removl vines were never logged) in the sme Cernet Suvignon vineyrd pnel (mx. 7.6 m prt) ws sujected to logging of erry temperture nd fruit-zone photosyntheticlly ctive rdition (PAR). This ws done to mesure erry tempertures typicl of our region, nd evlute how light nd temperture impct erry temperture. In ddition to these lef removl tretment-specific mesurements, mient temperture nd PAR were logged immeditely surrounding nd ove the dtlogger vine pnel, respectively. The temperture of two outside-fcing erries on oth est nd west cnopy sides of ech lef removl tretment were mesured with mini hypodermic thermocouple (type T, model HYP1/2, Omeg Eng., Stmford, CT) inserted pprox. 0.6 cm eneth erry skins nd ffixed with strong, ll-purpose glue. Fruit-zone PAR ws mesured with SQ-316 quntum sensor (Apogee Instruments, Logn, UT) plced on top, nd prllel to, the orienttion of one cordon in ech lef removl tretment. Amient PAR ws mesured with SQ-316 quntum sensor, mounted pprox. 3 m ove the ground (pprox. 0.6 m ove vine cnopies) t the middle of the vineyrd pnel where 195

212 ll dt were logged. Amient temperture ws mesured with two ST-100 thermistors (Apogee Instruments, Logn, UT). The thermistors were plced on the north nd south ordering ends (~ 7.6 m prt) of the dtlogger vine pnel, nd mounted t fruit-zone height inside nturlly spirted 6-pnel rdition shield (model A, R.M. Young, Trverse City, MI). Thermocouples were logged with n AM25T solid-stte multiplexer (Cmpell Scientific, Logn, UT) ttched to CR1000 dtlogger (Cmpell Scientific, Logn, UT), which lso logged the PAR nd mient temperture dt. All sensors, excepting the mient temperture sensors, were tken down t the end of ech field seson nd re-mounted in the following seson. Logged green nd red erry temperture: One Cernet Suvignon (red fruit) vine pnel nd one Petit mnseng (green fruit) vine pnel were locted directly cross the row from the ovementioned dtlogger vine pnel locted in the Cernet Suvignon vineyrd The sl leves nd lterl shoots were removed up to the second ctch wire (i.e. ~ 10 sl leves nd lterls) from the primry shoots growing from one cordon of one vine in ech vriety pnel to prevent erry temperture from eing differentilly ffected y fruit exposure extent. On the lef-pulled hlf of the vine, the temperture of one Cernet Suvignon nd one Petit mnseng erry positioned on the exterior fce of n est- nd west- exposed cluster ws logged on 1 min intervls from 3 Oct to 16 Oct 2014 ( totl of 12,356 totl logged mesurements of color/cnopy side specific erry temperture), nd from 22 Aug to 25 Sep 2015 ( totl of 38,645 totl logged mesurements of color/cnopy side specific erry temperture). The logged Cernet Suvignon nd Petit mnseng erries were ~0.6 m wy from ech other. Berry temperture ws mesured with mini hypodermic thermocouple (type T, model HYP1/2, Omeg Eng., Stmford, CT) inserted pprox. 0.6 cm eneth erry skins nd ffixed with n ll-purpose glue. Dt ws logged with the sme CR1000 dtlogger (Cmpell Scientific, Logn, UT) mentioned ove. 196

213 Mnul erry temperture mesurement for model vlidtion: Berry temperture ws mnully mesured on 29 Jun, 31 Jul, 25 Aug in 2015 in the sme Cernet Suvignon vineyrd tht the dtlogger ws locted. Mesurements were tken in every experimentl unit t three different times of dy on ech collection dte: morning (~ hrs), round solr noon (~ hrs), nd lte fternoon (~ hrs). Mesurements were tken y inserting mini hypodermic thermocouple (model HYP1/2, Omeg Eng., Stmford, CT) eneth the skins of erries, which ws connected to hndheld digitl thermometer (model HH 25, Omeg Eng., Stmford, CT). Ech mesurement took ~ 2 sec. At ech time of dy, on ech collection dte, nd on oth est nd west cnopy sides, three erries on the cluster s exterior fce, positioned t the top, center, nd ottom of the clusters, were mesured on two clusters orne on opposite vine cordons in ech experimentl unit. Thus, totl of 12 single-erry temperture mesurements, or six single-erry temperture mesurements on ech cnopy side, were tken in ech experimentl unit. The timefrme tht encompssed mnul erry temperture mesurement ws recorded in order to vlidte the prediction of erry temperture with mient ir temperture nd PAR using the herein developed model. The mient ir temperture nd PAR were retrieved from the dtlogger pnel, nd logged on one-min intervls. Model development: Hour ngle for ech logged dt point ws computed in Excel (Microsoft Corportion, Redmond, Wshington, USA) using formuls dpted from spredsheet posted y the Ntionl Ocenic & Atmospheric Administrtion Erth System Reserch Lortory ( Hour ngle converts locl solr time into degrees with which the sun moves cross the sky. Hour ngle is negtive in the morning, 0 t solr noon, nd positive in the fternoon. Accordingly, negtive nd positive hour ngles were those ssocited with rdition directed on the est nd west cnopy sides, respectively. Ech hour wy from solr 197

214 noon corresponds to 15 ngulr motion of the sun in the sky. Hour ngle ws rounded to 15 intervls to correspond to hourly chnges in locl solr time s the seson progressed. Berry temperture differentil ws clculted for ech logged dt point y sutrcting mient temperture from cnopy side-specific erry temperture. Summry sttistics, nd liner correltions etween mient PAR nd erry temperture differentil, were clculted using MATLAB (The Mthworks, Ntick, Msschusetts) for ech tretment-hour ngle comintion in 2014 nd 2015 only (NOTE: mient PAR ws not logged in 2013). Sttistics: All liner reltionships were nlyzed with simple one-wy ANOVA using the ivrite nlysis function in JMP Pro 11 (SAS, Cry, NC). Liner reltionship grphs were uilt using the grph uilder function in JMP Pro 11, nd hourly reltionship grphs were uilt using Sigm Plot 12.5 (Systt Softwre, Inc., Sn Jose, CA). 198

215 Results Cnopy chrcteriztion in plots used for mnul erry temperture mesurement nd model vlidtion: Lef removl in the Cernet Suvignon plots resulted in zero fruit-zone lef lyers (LLN) t verison in ll yers (Tle 1). Lef removl resulted in t lest three- nd, sometimes, four-fold increse in fruit-zone cluster exposure flux vilility (CEFA) t verison compred to removing no leves, nd there ws no difference in CEFA etween PB-4 nd PB-8. Tle 1. Pre-loom nd post-fruit set lef/lterl removl effect on fruit-zone lef lyer numer (LLN) nd cluster exposure flux vilility (CEFA) mesured t verison in Cernet Suvignon over Tretment LLN CEFA LLN CEFA LLN CEFA PB-NO PB PB Significnce < < < < < < PFS-NO n/ n/ PFS-6 n/ n/ Significnce n/ n/ < < < < PB-NO, PB-4, PB-8 = pre-loom lef removl of no, four, nd eight leves, respectively. Significnce of tretment effects (p > F; ns = not significnt t 0.05 level). *Mens in sme tretment group (columns) not shring letter re significntly different t 0.05 level sed on djusted p-vlues using Tukey HSD (PB tretments) nd Student s T-test (PFS tretments). Amient ir temperture vs. erry temperture: Berry temperture ws highly correlted with ir temperture, s shown y the highly significnt (p > F = < ), positive, liner reltionship etween these two vriles (Fig. 1); this reltionship will henceforth e referred to s the irerry temperture reltionship. NOTE: ll liner reltionships henceforth presented re significnt (p > F = < ). The ir-erry temperture reltionship ws mintined in ech of the three yers tht oth of these metrics were logged (Fig. 2). The ir-erry temperture reltionship ws reltively stronger on the west compred to est-cnopy side in ll three yers (Fig. 3). 199

216 Fig. 1. The liner reltionship etween mient ir temperture nd erry temperture over Fig. 2. The liner reltionship etween mient ir temperture nd erry temperture in ech of the three yers of dt collection,

217 Fig. 3. The liner reltionship etween mient ir temperture nd est- nd west- cnopy side erry temperture in ech of the three yers of dt collection, A reltively greter numer of dt points fell ove the ir-erry temperture trend line compred to elow the ir-erry temperture trend line (Figs. 1-3). Thus, it ws suspected tht there ws nother meteorologicl fctor tht incresed erry temperture without concomitntly incresing ir temperture. This ppered to e function of removing leves nd lterl shoots from the fruit-zone, s reltively more dt points fell ove the ir-erry temperture trend line in the pre-loom removl plots (PB-4/PB-8) compred to when no leves were removed (PB- NO) (Fig. 4). Accordingly, the ir-erry temperture reltionship ws stronger when no leves were removed from the fruit-zone compred to when leves were removed from the fruit-zone (Fig. 5). The ir-erry temperture reltionship in the fruit-zone of ech lef removl tretment ws consistent cross ll three yers of the study (Fig. 6). 201

218 Fig. 4. The liner reltionship etween mient ir temperture nd erry temperture, ccounting for pre-loom lef removl tretments over Fig. 5. The liner reltionship etween mient ir temperture nd erry temperture in ech pre-loom lef removl tretment over

219 Fig. 6. The liner reltionship etween mient ir temperture nd erry temperture in ech pre-loom lef removl tretment in ech of the three yers of dt collection, Cnopy side hd little effect on the ir-erry temperture reltionship in PB-NO plots, ut, s oserved when ll tretments were comined (Fig. 3), the ir-erry temperture reltionship ws stronger on the west compred to est cnopy side in PB-4 nd PB-8 plots (Fig. 7). Collectively, Figs. 1-7 showed tht erry temperture could e highly predicted (t lest R ) y mient ir temperture lone, nd without regrd to time of dy, or ny other meteorologicl fctor. 203

220 Fig. 7. The liner reltionship etween mient ir temperture nd est- nd west-cnopy side erry temperture in ech of the lef removl tretment plots over Amient photosyntheticlly ctive rdition (PAR) vs. erry temperture: The reltionship etween mient PAR nd erry temperture ws much weker thn the ir-erry temperture reltionship (Fig. 8). This reltionship will henceforth e referred to s the mient PAR-erry temperture reltionship. The mient PAR-erry temperture reltionship hd similr R 2 vlues in 2014 nd 2015, the two yers tht oth of these metrics were logged (Fig. 9). Opposite to the ir-erry temperture reltionship, the mient PAR-erry temperture reltionship ws reltively stronger on the est- compred to west-cnopy side over 2014 nd 2015 (Fig. 10). 204

221 Fig. 8. The liner reltionship etween mient photosyntheticlly ctive rdition (PAR) nd erry temperture over Fig. 9. The liner reltionship etween mient PAR nd erry temperture in ech of the two yers of dt collection,

222 Fig. 10. The liner reltionship etween mient PAR nd est- nd west- cnopy side erry temperture in ech of the two yers of dt collection, Unlike the ir-erry temperture reltionship, there ws little evident seprtion etween the numer of dt points tht fell ove or elow the mient PAR-erry temperture trend line (Figs. 8-10). Thus, it did not pper tht nother fctor ltered the reltionship etween mient PAR nd erry temperture. Nonetheless, there ws less cler seprtion of dt points etween PB-4/PB-8 nd PB-NO in the mient PAR-erry temperture reltionship (Fig. 11) compred to when the ir-erry temperture reltionship ws seprted y tretment (Fig. 4). However, PB-4 nd PB-8 dt points were further ove the mient PAR-erry temperture trend line in comprison to PB-NO dt points. Yet, the mient PAR-erry temperture reltionship ws reltively stronger in the PB-4 nd PB-8 plots compred to PB-NO plots (Fig. 12). This differed from the ir-erry temperture reltionship, which ws stronger in the PB-NO plots compred to PB-4 nd PB-8 plots (Fig. 5). 206

223 Fig. 11. The liner reltionship etween mient PAR nd erry temperture, ccounting for pre-loom lef removl tretments over Fig. 12. The liner reltionship etween mient PAR nd erry temperture in ech pre-loom lef removl tretment over

224 The mient PAR-erry temperture reltionship in fruit-zone lef removl tretments ws consistent cross the two yers of the study in which oth of these metrics were logged together (Fig. 13). Cnopy side hd little effect on the mient PAR-erry temperture reltionship in PB-NO plots, ut, s oserved when ll tretments were comined (Fig. 10), the mient PAR-erry temperture reltionship ws stronger on the est compred to west cnopy side in PB-4 nd PB-8 plots (Fig. 14). Collectively, Figs showed tht erry temperture ws not very predictle (0.338 R ) with mient PAR lone, nd without regrd to time of dy, or ny other meteorologicl fctor. Fig. 13. The liner reltionship etween mient PAR nd erry temperture in ech pre-loom lef removl tretment in ech of the two yers of dt collection,

225 Fig. 14. The liner reltionship etween mient PAR nd est- nd west-cnopy erry temperture in ech of the lef removl tretment plots over Fruit-zone photosyntheticlly ctive rdition (PAR) vs. erry temperture: The reltionship etween fruit-zone PAR nd erry temperture ws much weker thn the ir-erry temperture reltionship (Fig. 15). This reltionship will henceforth e referred to s the fruit-zone PARerry temperture reltionship. The fruit-zone PAR-erry temperture reltionship hd similr R 2 vlues over (Fig. 16). The fruit-zone PAR-erry temperture reltionship ws reltively stronger on the est- compred to west-cnopy side in ll yers tht dt ws logged, similr to the mient PAR-erry temperture reltionship (Fig. 17). 209

226 Fig. 15. The liner reltionship etween fruit-zone photosyntheticlly ctive rdition (PAR) nd erry temperture over Fig. 16. The liner reltionship etween fruit-zone PAR nd erry temperture in ech of the three yers of dt collection,

227 Fig. 17. The liner reltionship etween fruit-zone PAR nd est- nd west- cnopy side erry temperture in ech of the three yers of dt collection, There ws cler seprtion etween the dt points from the PB-4/PB-8 nd PB-NO tretments in the fruit-zone PAR-erry temperture reltionship (Fig. 18). However, the seprtion ws not ove or elow the fruit-zone PAR-erry temperture trend line. Rther, there were reltively more PB-4 nd PB-8 dt points s fruit-zone PAR incresed from left to right, while most PB-NO dt points fell in the lower fruit-zone PAR rnge. The strength of the fruit-zone PAR-erry temperture reltionship ws PB-8 > PB-4 > PB-NO (Fig. 19), nd this reltionship ws consistent over (Fig. 20). The fruit-zone PAR-erry temperture reltionship ws reltively stronger on the est- compred to west-cnopy side in PB-4 nd PB-8 plots, ut cnopy side hd lesser ering on this reltionship in PB-NO plots (Fig. 21). Collectively, Figs showed tht fruit-zone PAR hd more of n effect on erry temperture in lef removl plots compred to no lef removl plots. However, fruit-zone PAR lone ws not 211

228 prticulrly effective predictor of erry temperture. These figures lso showed tht fruit-zone PAR hd reltively greter effect on est- compred to west-side erry temperture. The reltionship etween fruit-zone PAR nd erry temperture ws consistent over Fig. 18. The liner reltionship etween fruit-zone PAR nd erry temperture, ccounting for pre-loom lef removl tretments over

229 Fig. 19. The liner reltionship etween fruit-zone PAR nd erry temperture in ech pre-loom lef removl tretment over Fig. 20. The liner reltionship etween fruit-zone PAR nd erry temperture in ech pre-loom lef removl tretment in ech of the three yers of dt collection,

230 Fig. 21. The liner reltionship etween fruit-zone PAR nd est- nd west-cnopy erry temperture in ech of the lef removl tretment plots over Fruit-zone reltive humidity vs. erry temperture: There ws wek, negtive reltionship etween fruit-zone reltive humidity (RH) nd erry temperture (Fig. 22). This reltionship will henceforth e referred to s the fruit-zone RH-erry temperture reltionship. The fruitzone RH-erry temperture reltionship ws mintined over , ut ws stronger in 2015 (R 2 = 0.415) compred to 2014 (R 2 = 0.242) (dt not shown). The fruit-zone RH-erry temperture reltionship ws similr etween cnopy sides, s R 2 vlues rnged cross cnopy sides in 2014, nd rnged cross cnopy sides 2015 (dt not shown). There ws no evident difference in the fruit-zone RH-erry temperture reltionship etween pre-loom lef removl tretment other thn the reltively greter erry tempertures oserved in PB-4 nd PB-8 compred to PB-NO plots t the sme fruit-zone RH (dt not shown). The fruitzone RH-erry temperture reltionship ws similr etween ll pre-loom lef removl 214

231 tretments, s there ws smll R 2 vlues rnge ( ) cross tretments, nd this smll rnge ws mintined cross tretments in 2014 ( ) nd 2015 ( ) (dt not shown). Lstly, when verged over , there ws little difference (R 2 = ) in the fruit-zone RH-erry temperture reltionship when evluted y every cnopy side nd preloom lef removl tretment comintion (dt not shown). This dt showed tht there ws wek, negtive liner reltionship etween fruit-zone RH nd erry temperture. While there ws no pprent difference in the fruit-zone RH-erry temperture reltionship etween tretment or cnopy side, this reltionship ws reltively stronger in 2015 compred to Fig. 22. The liner reltionship etween fruit-zone RH nd erry temperture over Averge fruit-zone RH nd mient ir temperture: The ove section showed tht the fruitzone RH ws the only logged meteorologicl metric tht ws negtively relted to erry temperture. However, the reltionship of fruit-zone RH to erry temperture (Fig. 22) ws very similr to the reltionship of fruit-zone RH to mient ir temperture (Fig. 23). There ws lso strong inverse reltionship etween fruit-zone RH nd mient ir temperture throughout the 215

232 dy (Fig. 24). Further, the diurnl fruit-zone RH-erry temperture nd fruit-zone RH-ir temperture reltionships prlleled ech other, excepting from , when R 2 vlues were reltively greter in the fruit-zone RH-erry temperture reltionship (Fig. 25). The sme fctor ws ssumed to e responsile for the reltively greter R 2 vlues in the fruit-zone RH-erry temperture reltionship, nd the reltively lower R 2 vlues of the ir-erry temperture reltionship, s these occurred t the sme times of dy. Collectively, these reltionships showed tht fruit-zone RH ws not necessrily inversely relted to erry temperture. Rther, fruit-zone RH decresed s ir temperture, nd consequently erry temperture, incresed throughout the dy, nd vice-vers. Fig. 23. The liner reltionship etween fruit-zone RH nd mient ir temperture over

233 95 26 Averge fruit-zone RH (%) Averge fruit-zone RH Amient ir temperture Amient ir nd erry temperture ( C) Hour Fig. 24. The diurnl trend of verge fruit-zone RH, nd mient ir nd erry temperture over R 2 (fruit-zone RH vs. erry / miient ir temperture) Fruit-zone RH-erry temperture Fruit-zone RH-ir temperture Amient ir temperture-erry temperture R 2 (mient ir temperture vs. erry temperture) Hour Fig. 25. The hourly R 2 of the diurnl reltionship etween fruit-zone RH nd mient ir temperture, fruit-zone RH nd erry temperture, nd mient ir temperture nd erry temperture over

234 Amient ir temperture vs. green nd red erry temperture: The liner ir-erry temperture reltionship ws strong when evluted y comining logged green nd red erry temperture (Fig. 26). The temperture of red erries tended to e mrginlly greter thn green erries t given mient ir temperture. Accordingly, the ir-erry temperture reltionship ws slightly stronger in green erries (R 2 = 0.914) compred to red erries (R 2 = 0.900) (dt not shown). This trend ws mintined in oth yers, s the ir-red erry temperture reltionship hd lower R 2 vlues when compred to the ir-green erry temperture reltionship over the course of 2014 nd 2015 (dt not shown). When evluted y cnopy side, the ir-erry temperture reltionship did not gretly differ etween green nd red erries, lthough, s shown ove, this reltionship ws stronger on the west cnopy side (R 2 = cross red nd green erries, respectively) compred to the est cnopy side (R 2 = cross red nd green erries, respectively) (dt not shown). This dt reveled tht the ir-erry temperture reltionship ws mrginlly ffected y erry color, ws mintined in oth yers, nd tht cnopy side did not differentilly ffect this reltionship. 218

235 Fig. 26. The liner reltionship etween mient ir temperture nd erry temperture, ccounting for green nd red erry color over Amient PAR vs. green nd red erry temperture: There ws positive, liner reltionship etween mient PAR nd green nd red erry temperture (Fig. 27), lthough this reltionship ws much weker thn the ir-erry temperture reltionship. However, unlike in the ir-erry temperture reltionship, there ws no cler trend in seprtion of dt points etween green nd red erries. The mient PAR-erry temperture reltionship ws slightly stronger in red (R 2 = 0.479) erries compred to green erries (R 2 = 0.463), nd this trend ws mintined in oth 2014 (R 2 = ) nd 2015 (R 2 = ) (dt not shown). When evluted y cnopy side, the mient PAR-erry temperture reltionship did not gretly differ etween green nd red erries, lthough, s shown ove, this reltionship ws stronger on the est cnopy side (R 2 = cross green nd red erries, respectively) compred to the west cnopy side (R 2 = cross red nd green erries, respectively) (dt not shown). 219

236 Collectively, this dt showed tht the mient PAR-erry temperture reltionship ws not gretly ffected y erry color, ws mintined in oth yers, nd tht cnopy side did not differentilly ffect this reltionship. Fig. 27. The liner reltionship etween mient PAR nd erry temperture, ccounting for green nd red erry color over

237 Diurnl ptterns of the reltionship of erry temperture with mient ir temperture, mient PAR, nd fruit-zone PAR: Amient ir temperture hd the strongest liner reltionship to erry temperture compred to ll other mesured meteorologicl metrics, nd there ws strong evidence tht dt shifted ove the trend line, s well s evidence of consistent, cnopy-side specific response to the ir-erry temperture reltionship. Amient nd fruit-zone PAR hd weker, positive, reltionships to erry temperture. However, the mient/fruit-zone PARerry temperture reltionships were stronger in opposite scenrios compred to the ir-erry temperture reltionship; this ws suggestive tht rdition ws the fctor tht chnged the reltionship etween mient ir nd erry temperture. For exmple, the mient/fruit-zone PAR-erry temperture reltionship ws reltively stronger on the est cnopy side nd in lef removl plots, while the ir-erry temperture reltionship ws stronger on the west cnopy side nd in no lef removl plots. While RH ppered to e negtively relted to erry temperture, no trend ws oserved tht suggested it ws fctor tht ffected ir-erry temperture reltionship. Further investigtion reveled tht RH merely decresed s mient ir temperture, nd thus erry temperture, incresed; nd vice-vers. Therefore, it ws concluded tht erry temperture ws not impcted y RH. Berry color did not chnge the wy tht erry temperture ws ffected y mient ir temperture nd PAR, regrdless of cnopy side. It ws similrly concluded tht erry temperture ws not impcted y erry color. Collectively, these findings suggested tht erry temperture ws primrily function of mient ir temperture, ut tht rdition could ffect this reltionship y incresing erry temperture when mient ir temperture ws not concomitntly incresed; this ws prticulrly true when leves were removed from the fruit-zone. The liner reltionships did not tke time into ccount, ut mient ir temperture nd PAR re known to chnge throughout the dy. 221

238 Thus, it ws suspected tht the ir-erry temperture reltionship would chnge over the course of the dy, nd this would e different etween lef removl tretments nd cnopy sides. Thus, the reltionship of est- nd west-side erry temperture with mient ir temperture, mient PAR, nd fruit-zone PAR were diurnlly investigted. The diurnl ir-erry temperture reltionship showed tht erry temperture ws lwys highly predictle with ir temperture in no lef removl plots (R ), regrdless of cnopy side or time of dy (Fig. 28 C). By contrst, erry temperture prediction with ir temperture ws dependent on time of dy nd cnopy side in lef removl plots (Fig. 28 A, B). Est-side erry temperture ws highly predictle with ir temperture until 0700, predictility decresed until 1400, nd ws gin highly predictle with ir temperture therefter. Westside erry temperture ws highly predictle with ir temperture until 1400, fter which predictility decresed until 1900, nd ws gin highly predictle with ir temperture therefter. There were slight decreses in the ir-erry temperture reltionship on the west-side in the morning nd the est-side in the fternoon, nd these were reltively greter in eight lef removl plots (Fig. 28 A) compred to four lef removl plots (Fig. 28 B). 222

239 R 2 (mient ir temperture vs. erry temperture) EAST WEST EAST WEST A B EAST WEST C Hour Fig. 28. The hourly R 2 of the cnopy-side specific diurnl reltionship etween mient ir nd erry temperture, s ffected y pre-loom removl of eight (A), four (B), nd no (C) leves over

240 The diurnl mient PAR-erry temperture reltionship ws similr in lef removl nd no lef removl plots (Fig. 29 A-C). Est-side erry temperture ws unffected y mient PAR until 0600, fter which R 2 vlues incresed, reched morning mximum etween , dropped until 1400, reched n fternoon mximum t 1700, then dropped off therefter until erry temperture ws no longer influenced y mient PAR t West-side erry temperture ws similrly influenced y diurnl mient PAR when compred to est erry temperture. However, the mximum fternoon R 2 pek ws greter on the west cnopy side, wheres the mximum morning R 2 pek ws greter on the est cnopy side. The est-side morning R 2 peks were slightly greter thn the west-side fternoon R 2 peks, suggestive tht rdition ws consistently greter in the morning compred to fternoon hours. Berry temperture ws ffected y mient PAR to greter extent in lef removl compred to no lef removl plots, s shown y the reltively greter R 2 vlues in lef removl plots. There ws greter difference etween the morning nd fternoon R 2 peks on ech cnopy side in the lef removl compred to no lef removl plots. These lst two points suggested tht there ws reltively greter direct nd indirect rdition in lef removl compred to no lef removl plots. The diurnl mient PAR-erry temperture reltionship ws essentilly inverse to tht of the diurnl ir-erry temperture reltionship s R 2 vlues were reduced in the ir-erry temperture reltionship, R 2 vlues incresed in the mient PAR-erry temperture reltionship, nd vice-vers. Collectively, these trends provided further evidence tht rdition ws the primry fctor tht chnged the wy ir temperture ws relted to erry temperture. 224

241 A EAST WEST R 2 (mient PAR vs. erry temperture) B EAST WEST C EAST WEST Hour Fig. 29. The hourly R2 of the cnopy-side specific diurnl reltionship etween mient PAR nd erry temperture, s ffected y pre-loom removl of eight (A), four (B), nd no (C) leves over

242 The diurnl pttern of the fruit-zone PAR-erry temperture reltionship ws similr to the diurnl pttern of mient PAR-erry temperture reltionship (Fig. 30). Est-side erry temperture tended to e ffected y fruit-zone PAR to greter extent in the morning, nd vicevers for west-side erries in the fternoon, s evidenced y higher R 2 vlues in these respective periods. Berry temperture ws ffected y fruit-zone PAR to greter extent in the pre-loom of eight compred to four lef removl plots, nd to greter extent in the pre-loom removl four compred to no lef removl plots (Fig. 30 A-C). The R 2 vlues were lower in the fruitzone PAR-erry temperture reltionship compred to the mient PAR-erry temperture reltionship. Further, the middy depression in the R 2 vlues of the fruit-zone PAR-erry temperture reltionship were greter nd lsted longer thn those of the mient PAR-erry temperture reltionship. These lst two points were suspected to e function of the ovecordon (i.e. center of fruit-zone) plcement of the fruit-zone quntum sensors. For the former point, the logged temperture of outside-fcing erries ws uncoupled from the logged rdition trnsmission to the center of the fruit-zone. For the ltter point, rdition trnsmission to the center of the fruit-zone ws lrgely locked y the cnopy during middy. Though weker thn the mient PAR-erry temperture reltionship, the diurnl trend of the fruit-zone PAR-erry temperture reltionship ws inverse to tht of the diurnl ir-erry temperture reltionship. This, gin, supported tht rdition ws the fctor tht decresed the ility to predict erry temperture with mient ir temperture. 226

243 A EAST WEST R 2 (fruit-zone PAR vs. erry temperture) B EAST WEST C EAST WEST Hour Fig. 30. The hourly R 2 of the cnopy-side specific diurnl reltionship etween fruit-zone PAR nd erry temperture, s ffected y pre-loom lef removl of eight (A), four (B), nd no (C) leves. 227

244 Diurnl ptterns of mient ir temperture, erry temperture, nd mient nd fruit-zone PAR: Collectively, the ove plots showed tht erry temperture is highly predictle with, nd thus minly ffected y, mient ir temperture. However, erry temperture ws lso ffected y other fctors, such s temporl chnges in hour ngle over the course of the dy, fruit-zone lef nd lterl shoot removl, nd the cnopy side tht the erry ws exposed to. Diurnl ptterns of mient ir temperture, mient nd fruit-zone PAR, nd erry temperture re displyed s function of oth lef removl tretment nd cnopy side (Fig 31). The diurnl reltionship plots presented ove, nd the hourly men vlues presented in Fig. 31, showed tht there re only certin times of the dy tht erry temperture is different from mient ir temperture. Further, tht erry temperture is primrily ffected y rdition when leves re removed from the fruit-zones, when the ove-hed cnopy is not locking rdition (i.e. surrounding solr noon), nd when tht cnopy side is receiving direct rdition. 228

245 30 25 A Amient ir nd erry temperture, T ( C) Berry T EAST Berry T WEST Amient T Amient PAR Fruit-zone PAR ~ solr noon C B Amient nd fruit-zone photosyntheticlly ctive rdition, PAR ( mol m -2 s -1 ) Hour Fig. 31. Diurnl pttern of mient ir temperture, mient nd fruit-zone PAR, nd erry temperture s ffected y pre-loom removl of eight (A), four (B), nd no (C) fruit-zone leves/lterls. Dt logged on 15- nd 1-min intervls over sesons. Amient PAR is presented s 20% of ctul vlue, nd ws logged on 15- nd 1-min intervls over sesons

246 Berry temperture prediction models: The dt henceforth presented used 15 hour ngles s solr rdition-time reference, s time of dy per se ws not responsile for ffecting the irerry temperture reltionship. Rther, hour ngle chnged over the course of the dy nd seson, which chnged the ility to predict erry temperture with mient ir temperture. The use of hour ngle ws included to mke the models more pplicle cross geogrphic regions. Since erry temperture could e predicted with mient ir temperture, except during the times of dy tht rdition lso ffected erry temperture, liner erry temperture prediction models were developed with oth mient ir temperture nd rdition. While oth mient nd fruit-zone PAR ffected erry temperture, mient PAR ws used s the independent vrile to develop models. Compred to fruit-zone PAR, mient PAR (1) hd stronger reltionship to erry temperture, (2) could e used with the hour ngle time reference, nd (3) cn often e retrieved from pulic wether sttions. As such, mient PAR ws expected to e more pplicle for use y growers nd reserchers like. Most forthcoming models were developed seprtely y cnopy side ecuse rdition differentilly ffected the temperture of erries exposed to est nd west cnopy sides. NOTE: the difference etween erry temperture nd mient ir temperture will henceforth e referred to s the erry temperture differentil. 230

247 Cnopy side-specific, lef removl tretment non-specific, erry temperture-prediction models: The ility to predict est-side erry temperture differentil when verged cross ll lef removl tretments ws greter from -105 to -15 hour ngles (R 2 = ) compred to 15 to 105 hour ngles (R 2 = ) (Appendix A). Men, mximum, nd stndrd devition in erry temperture differentil tended to e greter from -75 to 15 hour ngles compred to hour ngles outside of this rnge. Men erry temperture differentil ws 1.12 to 3.53 C from -75 to 15 hour ngles. The ility to predict west-side erry temperture differentil when verged cross ll lef removl tretments ws greter from 30 to 75 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix B). Men, mximum, nd stndrd devition in erry temperture differentil ll tended to e greter from 30 to 75 hour ngles compred to hour ngles outside of this rnge. Men erry temperture differentil ws 1.19 to 2.55 C from 30 to 75 hour ngles. Cnopy side-specific, lef removl tretment-specific, erry temperture-prediction models: The ility to predict est-side erry temperture differentil in PB-NO plots ws greter from -60 to 0 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix C). Men nd stndrd devition in erry temperture differentil tended to e greter from -60 to 0 hour ngles compred to hour ngles outside of this rnge, while mximum erry temperture differentil hd inconsistent trends cross hour ngles. Men erry temperture differentil ws 0.70 to 1.10 C from -60 to 0 hour ngles. The ility to predict west-side erry temperture differentil in PB-NO plots ws greter from -75 to -45 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix D). Mximum nd men erry temperture differentil hd inconsistent trends cross hour ngles, nd were t most 6.15 nd 0.83 C, respectively. Stndrd devition in erry 231

248 temperture differentil ws greter from -75 to -45 hour ngles compred to hour ngles outside of this rnge. The ility to predict est-side erry temperture differentil in PB-4 plots ws greter from -105 to -15 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix E). The predictility of est-side erry temperture differentil in PB-4 plots from -105 to -15 hour ngles ws greter thn the predictility of est-side erry temperture differentil in PB-NO plots from -60 to 0 hour ngles. Men, mximum, nd stndrd devition in erry temperture differentil tended to e greter from -75 to 0 hour ngles compred to hour ngles outside of this rnge. Men erry temperture differentil ws 1.86 to 4.28 C over the -75 to 0 hour ngles, nd these tempertures were greter thn the men est-side erry temperture differentil over the -60 to 0 hour ngles in the PB-NO plots. The ility to predict west-side erry temperture differentil in PB-4 plots ws greter from 30 to 90 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix F). Men, mximum, nd stndrd devition in erry temperture differentil were greter from 30 to 75 hour ngles compred to hour ngles outside of this rnge. Men erry temperture differentil ws 1.36 to 3.21 C over the 30 to 75 hour ngles, nd these tempertures were greter thn men west-side erry temperture differentil t ny hour ngle in the PB-NO plots. Similr to the hour ngle rnge tht est-side erry temperture differentil in the PB-4 plots ws est predicted, there ws greter ility to predict est-side erry temperture differentil in the PB-8 plots from -105 to -15 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix G). Mximum, men, nd stndrd devition in erry temperture differentil tended to e greter from -75 to 0 hour 232

249 ngles compred to hour ngles outside of this rnge. Men erry temperture differentil ws 1.68 to 4.64 C over the -75 to 0 hour ngle rnge; these tempertures were similr to men est-side erry temperture differentil over the -75 to 0 hour ngle rnge in the PB-4 plots, nd greter thn the men est-side erry temperture differentil over the -60 to 0 hour ngles in the PB-NO plots. There ws greter ility to predict west-side erry temperture differentil in the PB-8 plots from 30 to -75 hour ngles (R 2 = ) compred to hour ngles outside of this rnge (R 2 = ) (Appendix H). Mximum erry temperture differentil tended to e inconsistent cross hour ngles, ut ws consistently high from 15 to 90 hour ngles. Men nd stndrd devition in erry temperture differentil were greter over the 30 to 90 hour ngles compred to hour ngles outside of this rnge. Men erry temperture differentil ws 1.40 to 3.37 C over the 30 to 75 hour ngle rnge; these tempertures were similr to the men west-side men erry temperture differentil in the PB-4 plots over the 30 to 75 hour ngle rnge, nd greter thn men west-side erry temperture differentil t ny hour ngle in the PB-NO plots. Mnully mesured erry temperture for model vlidtion: Amient ir temperture tended to e greter fter the 0 hour ngle compred to efore the 0 hour ngle over the three dtes tht mnully mesured erry temperture ws compred to predicted erry temperture (Fig. 32). The stndrd devition of mient ir temperture ws prticulrly high t the -60 to 0 hour ngles on 31 Jul 2015 compred to other hour ngles on this dte, s well s on 29 Jun nd 25 Aug Amient PAR tended to e greter t the 0 nd 15 hour ngles, leit mient PAR t the 0 hour ngle on 31 Jul 2015 ws typiclly low (Fig. 33). Similr to the stndrd devition of mient ir temperture, the stndrd devition of mient PAR ws prticulrly high t the -60 to 0 hour ngles on 31 Jul 2015 compred to other hour ngles on this dte, s 233

250 well s on 29 Jun nd 25 Aug Thus, it ws inferred tht mient ir temperture nd PAR ws highly vrile on 31 Jul compred to other dtes. Fig. 32. The men mient ir temperture t ech hour ngle dte tht mnul erry temperture ws mesured for purposes of comprison to predicted erry temperture. Error rs re ± one stndrd devition from the men. Fig. 33. The men mient PAR t ech hour ngle dte tht mnul erry temperture ws mesured for purposes of comprison to predicted erry temperture. Error rs re ± one stndrd devition from the men. 234