Improving fruit size and productivity of Scifresh apple: Effects of regulating whole-tree floral bud density using Centrifugal Training.

Transcription

1 Improving fruit size and productivity of Scifresh apple: Effects of regulating whole-tree floral bud density using Centrifugal Training. Tustin DS, Henriod RE, Dayatilake GA, Breen KC, Oliver MJ May 2008 Report to ENZA Ltd HortResearch Client Report No HortResearch Contract No Tustin DS, Henriod RE, Dayatilake GA, Breen KC, Oliver MJ The Horticulture and Food Research Institute of New Zealand Ltd HortResearch Hawke's Bay Private Bag 1401 Havelock North Hastings 4157 NEW ZEALAND Tel: Fax:

2 DISCLAIMER Unless agreed otherwise, HortResearch does not give any prediction, warranty or assurance in relation to the accuracy of or fitness for any particular use or application of, any information or scientific or other result contained in this report. Neither HortResearch nor any of its employees shall be liable for any cost (including legal costs), claim, liability, loss, damage, injury or the like, which may be suffered or incurred as a direct or indirect result of the reliance by any person on any information contained in this report. COPYRIGHT (2008) The Horticulture and Food Research Institute of New Zealand Ltd, Private Bag , Auckland Mail Centre, Auckland 1142, New Zealand. All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system, transmitted, reported, or copied in any form or by any s electronic, mechanical or otherwise without written permission of the copyright owner. Information contained in this publication is confidential and is not to be disclosed in any form to any party without the prior approval in writing of the Chief Executive Officer, The Horticulture and Food Research Institute of New Zealand Ltd, Private Bag , Auckland Mail Centre, Auckland 1142, New Zealand. LIMITED PROTECTION This report may be reproduced in full, but not in part, without prior consent of the author or of the Chief Executive Officer, The Horticulture and Food Research Institute of New Zealand Ltd, Private Bag , Auckland Mail Centre, Auckland 1142, New Zealand. This report has been prepared by The Horticulture and Food Research Institute of New Zealand Ltd (HortResearch), which has its Head Office at 120 Mt Albert Rd, Mt Albert, AUCKLAND. This report has been approved by: Research Scientist On behalf of Group Leader, Quality Sytems Date: 21 May 2008 Date: 21 May 2008

3 CONTENTS Page EXECUTIVE SUMMARY... 1 INTRODUCTION... 3 MATERIAL AND METHODS... 4 Experimental... 4 General Methods... 5 Year Two Methods... 6 RESULTS AND DISCUSSION... 7 Flowering and Fruit Set... 7 Return Bloom... 7 Floral Spur Characteristics... 8 Fruit Set Behaviour... 9 Tree Productivity Fruit Quality General Comments CONCLUSIONS RECOMMENDATIONS REFERENCES... 20

4

5 1 EXECUTIVE SUMMARY Improving fruit size and quality of Scifresh apple: Effects of regulating whole-tree floral bud density using Centrifugal Training Report to ENZA Ltd Tustin DS, Henriod R, Dayatilake D, Breen K, Oliver MJ May 2008 The current fruit size profile of Scifresh /Jazz does not meet the market requirement and new production technologies are required to improve fruit size. This project investigates how new management methods that reduce the number of floral spurs and their distribution within the tree affects fruit quality. These techniques are thought likely to improve performance of floral spurs and overcome cropping and fruit size problems believed to be related to the characteristic of excessive floral initiation by Scifresh. The main outcomes found after two consecutive years of this study were that: Fruit set of individual spurs was strongly enhanced by Centrifugal Training (CT). The combined fruit set responses of CT and standard Vertical Axe (VA) management treatments from two consecutive years has enabled the development of a fruit set response model for Scifresh apple showing fruit set is enhanced in proportion to reduction in initial floral bud density. The fruit set response model provides strong evidence for the hypothesis that the excessive flower initiation of Scifresh is a cause of reduced fruit set and erratic cropping including variable fruit size. Spur thinning treatments with CT did not alter the composition or size of spur leaves or flower receptacles of floral buds at bloom in the second year of this study. Mean tree yields of 34 to 49 kg in the second year represented productivity of 56 to 80 t/ha gross yields. Yields of 34 to 49 kg per tree were achieved from relatively low final crop densities in a range of 3.9 to 6.1 fruit per cm 2 trunk cross-sectional area. Mean fruit size was largest among the CT crop load treatments, consistently between g in the second year, a g increase from the previous year. It is thought that fruit size responses with CT management may be a result of a larger proportion of floral buds originating as terminals of short to medium length shoots the buds known to produce significantly larger fruit than any other bud type of Scifresh. Fruit traits that are used as harvest maturity attributes were not affected by any of the treatment factors used the study. Fruit from crops in the second year had similar or improved soluble solids concentration, flesh firmness and titratable acidity compared with fruit of the same treatments in the previous year. Internal quality of fruit did not differ greatly between CT and VA management systems. Where differences occurred they appeared to be related to effects of light cropping from biennial bearing within some VA treatments.

6 2 For further information, please contact: Stuart Tustin The Horticulture and Food Research Institute of New Zealand Ltd HortResearch Hawke's Bay Private Bag 1401 Havelock North Hastings 4157 NEW ZEALAND Tel: Fax:

7 3 INTRODUCTION The current crop size profile of Scifresh /Jazz does not meet marketing requirements, with an excess of small-sized fruit commonly being produced from young intensive plantings in New Zealand. Although reducing crop load can improve fruit size, a complementary approach is to improve the quality and performance of the spur flower clusters i.e., to increase the number and size of the spur leaves, the flower size and the bourse development. We have determined that the largest fruit are borne on the terminal flower sites of shoots, which show all the enhanced features just described (Breen et al. 2007; Tustin et al. 2007). This work, commenced in 2006, is to confirm our hypothesis that fruit size within a tree is linked to spur size/strength and then to manipulate spur vigour by a range of management treatments intended to stimulate improved spur vigour. If spur vigour can be manipulated to result in increased fruit size, productivity will be concomitantly improved by the simple increase in fruit weight for an equivalent number of fruit per tree. This project investigates the effects on fruit quality of management methods that reduce the number of floral spurs and their distribution within the tree, as a technique thought likely to enhance the vigour and performance of those floral spurs retained in the tree. Artificial spur extinction by manually reducing the number and distribution of spurs is a practice in a new French tree management technique called Centrifugal Training (CT) applied at the individual branch level, to stimulate spur strength for improving fruit quality and regularity of annual production (Lauri & Lespinasse 1999; Lauri et al. 2004). The theory is that artificial spur extinction (spur removal) largely substitutes for annual pruning to minimise reiterative vegetative growth induced by pruning. The technique depends on training fruiting limbs to a slightly pendant position, understood to be the most optimally reproductive and vegetativelysubdued orientation for fruiting branches in apple. The proposed advantages from CT are that because of subdued vegetative growth, a higher proportion of the total seasonallyaccumulated dry matter resources can be allocated into reproductive development (i.e. flowering, fruit set, fruit development, new floral initiation). One response to reallocation of assimilates is the strengthening of the axillary structure on the reproductive spur (the bourse shoot) into a new reproductive bud, as either a spur or bourse shoot terminal bud. In this way, floral spur development may be enhanced, leading to improved fruit size and annual productivity. What is our rationale behind how CT methods might improve productivity and fruit quality of Scifresh? Scifresh exhibits some characteristic behaviours we associate with low productivity and small fruit size, including a tendency for excessive final fruit drop causing unpredictable fruit set and a very high level of natural annual spur extinction. From observations, we hypothesise that these traits arise from the tendency of Scifresh for excessive floral initiation, which result in many poorly-developed and weak floral spurs that respond as described above. A second feature of this cultivar characteristic may therefore be that many floral sites retain fruit with reduced potential for large fruit size. These combined responses suggest conditions of strong inter-bud competition for resources within Scifresh trees as a consequence of very high floral bud initiation, in particular competition for reserves during the flowering and fruit set period early in the seasonal growth cycle.

8 4 In considering methods to overcome these fruit size and productivity problems, the potential for CT, in particular artificial spur extinction, was identified as a practical management approach that might alter resource allocation and the physiological responses regulating flowering, fruit set, fruit development and productivity at a whole-tree level. In particular the concept of artificial spur extinction could modify resource availability and allocation to be more favourable for individual flowering buds. The study was initiated prior to budbreak in the season with the initial cropping and fruit quality responses from that first season reported to ENZA Ltd in HortResearch Client Report No (Tustin et al. 2007). The work reported here presents the responses of Scifresh trees through the second year after commencement and continuation of CT management compared with standard New Zealand Vertical Axe (VA) tree management in an intensive planting system on M.9 dwarf rootstock. Development responses measured in the second year included flowering and fruit set, fruit quality and tree productivity. MATERIAL AND METHODS EXPERIMENTAL The study was established in August 2006 using a block of six-year-old Scifresh trees grown on M.9 rootstock, planted at 3.5 x 1.75 m (1632 trees per ha) located at Lawn Rd Research Orchard, Hawke s Bay Research Centre of The Horticulture and Food Research Institute of New Zealand Ltd (HortResearch). Trees had been trained since planting as VA dwarf central leader trees. Productivity and fruit quality responses were compared between trees managed using CT methods and trees managed using standard VA training. The experimental layout used was a factorial design with three experimental factors: crop load (three levels), fruit number within crop level (either one or two fruit ) and tree management system (CT versus VA). Treatments were applied as single whole-tree plots in a randomised complete block design, using seven blocks. The factor crop load was defined as the number of fruit per cm 2 branch cross sectional area (BCA), set at four fruit, five fruit or six fruit per cm 2 BCA and set up on an individual branch basis, following methods described for CT management with Gala apple by Lauri et al. (2004). Using the results of Lauri et al. (2004), we calculated theoretical yields per hectare for the crop load treatments to be 53, 66 and 80 t/ha for the crop levels of four, five and six fruit per cm 2 BCA, respectively, as a guide for selecting the crop levels in our study. These calculations indicated average numbers of fruit per tree for each crop load to be 170 fruit for a crop level of four fruit per cm 2 BCA, 205 fruit for a crop level of five fruit per cm 2 BCA and 247 fruit for a crop level of six fruit per cm 2 BCA. In VA trees that had no spur manipulation, these same crop levels were set by hand thinning after final fruit drop. The second experimental factor, fruit number within crop level, was set up by using two levels of spur thinning for setting crop load under CT management. In the first instance, spurs were thinned to the number required to give the required crop level when each spur would bear only a single fruit. In treatments where each spur bore two fruit, the crop load was set by doubling the intensity of spur thinning, resulted in half the number of spurs required when each spur carried only a single fruit. This was a substantially more severe adjustment to the canopy composition of CT-managed trees.

9 5 The third experimental factor, tree management, required the setting up of trees using CT management techniques, in particular the use of spur removal (artificial spur extinction) on a branch-by-branch basis to achieve the required crop load treatment and where required restraining branches into a slightly pendant ( ) position. The trees retained under conventional VA central leader training were also considered to be the experimental control for the tree management factor. Treatments set up in the were repeated on the same trees in the season. GENERAL METHODS Spur thinning for CT treatments was carried out in the late dormant to early budbreak period annually. As well as spur thinning, all axillary buds on annual shoots were removed except that a deblossomed axillary bud was retained approximately every 10 cm for future spur replacement. Axillary bud removal was done during early bloom. At the time of setting up branch spur densities in CT treatments, any branches within these trees that were oriented above a horizontal position were tied down to a slightly pendant orientation for reasons described in the introduction, related to reducing and controlling branch vigour. Trees under VA management were not manipulated at this time, other than receiving their normal winter pruning. All trees had their trunk cross-sectional area determined from measurements of their trunk circumference 20 cm above the bud union. No chemical thinning was carried out. Fruit set was measured after final drop, prior to hand thinning adjustments to set final crop level treatments. One major limb in each of the lower east and west sides of trees was used as the fruit set monitored limb. The basal circumference of these limbs was measured for calculating branch cross-sectional area. Fruit set was recorded as the number of fruit that set on each individual floral spur, including recording zero set for spurs that flowered and set no fruit. Immediately after final set and fruit drop in late November, crop levels equivalent to those calculated and set by spur thinning with CT management at budbreak were set up in VAmanaged trees. All fruit set on axillary bud sites were also removed in VA-managed trees. At the same time, bud sites in CT-managed trees that had set more than one or two fruit in their respective treatments were thinned to the required number of one or two fruit. Measurements of yield, fruit size and fruit quality attributes were made once fruit were harvested. Fruit were harvested according to ENZA-defined commercial maturity and quality standards, which required three selective picks over an interval of days from the first to the third harvest. Harvested fruit from each tree was passed over an electronic fruit sizer, which counted the number and recorded the mass of each individual fruit. As fruit from each replicate tree were discharged from the sizer, a random sample of ten fruit was collected to use for harvest quality assessment. The following harvest quality attributes were measured: 1. Blush coverage (%) 2. Background colour (0 9) using the industry Braeburn background colour swatch 3. Fresh weight (g) 4. Fruit firmness (kgf) measured using an 11.1 mm head with the Güss fruit texture analyser

10 6 5. Soluble solids concentration ( Brix) measured with an Atago PAL-1 digital refractometer 6. Titratable acidity (% malic acid) measured by titrating juice with 0.1 M NaOH using a bulk juice sample extracted from two vertical wedges of tissue from opposite sides of all ten fruit (Metrohm 730 auto-titrator) 7. Starch pattern index using the 0 6 scale of the industry standard starch pattern chart. YEAR TWO METHODS In August of 2007, trees under CT management were lightly pruned as required and branch spur densities were reset to their respective experimental crop loads x fruit number levels. Trees under VA management were pruned as required according to standard industry practice for dwarf trees. Floral spur quality was evaluated using four floral spurs, terminals and axillary buds per tree collected at mid bloom. Floral buds were all collected on the same date at exactly the same stage of anthesis when the king flower had opened but all lateral flowers were unopened. All floral buds were collected from the west side of the tree, from limbs m above the ground. Spurs were dissected into floral and vegetative components which were measured as follows: 1. Total leaf number 2. Total leaf area (cm 2 ) using a LiCor 3100 area meter 3. Total number of flowers 4. Total receptacle dry weight after drying to constant weight at 60 o C in a hot air oven. During bloom, the return bloom flowering responses of all treatments was measured by counting the number of floral spurs on two major limbs, one in each of the lower east and west sides of trees. The basal circumference of these limbs was measured for calculating branch cross-sectional area. These same monitored limbs were used for measuring fruit set, recorded at the completion of final fruit drop in late November. Fruit set was measured as the number of fruit that set on each individual floral spur, including recording zero set for spurs that flowered and set no fruit. These data also enabled calculation of total number of spurs setting fruit and overall fruit set density. After fruit set was recorded, all trees were hand thinned to crop loads and fruit number per spur according to their respective treatments, as described above in methods for the first year of the study. In this second year, in particular with VA trees from the higher crop load treatments, biennial bearing was evident, reducing flowering on spur and terminal buds to less than required for some treatment crop loads. It was decided for VA trees in these situations to use fruit set on axillary buds to endeavour to achieve the required fruit numbers on individual branches. This correction was not possible with CT trees as the methods involved removal of most axillary buds prior to flowering and fruit set. Fruit harvest, fruit quality and yield measurements in the second year were conducted following the procedures and methods used in the first year as presented above.

11 7 RESULTS AND DISCUSSION FLOWERING AND FRUIT SET Return Bloom Centrifugal training imposes upper limits to floral bud density of 4, 5 or 6 flower clusters per cm 2 BCA for crop loads using one fruit and 2, 2.5 or 3 flower clusters per cm 2 BCA for crop loads using two fruit. Return bloom floral bud density was less than these limits in all CT treatments indicating that not all spurs selected during spur thinning were floral (Table 1). In fact, return bloom floral bud density was similar for all crop loads within CT management using one fruit and was lower than in the previous year (Tustin et al. 2007). Return bloom of CT treatments using two fruit was closest to the upper limits imposed by spur thinning, indicating that most of these spurs retained were floral. Floral bud density of VA trees was similar to CT trees for the lowest crop load treatment but declined significantly as crop load was increased. Trees trained as VA within the crop load treatments five or six fruit per cm 2 BCA had floral bud densities of less than two flower clusters per cm 2 BCA and were visibly light flowering. In comparison, CT and VA trees which achieved floral bud densities above 2.5 flower clusters per cm 2 BCA appeared to provide a satisfactory return bloom and were not considered to be biennially off-flowering. The patterns of return bloom observed in 2007 were generally related to the actual cropping levels among the various management treatments from the previous season, where bloom density declined in response to higher tree yields, which is an expected behaviour of apple trees. Table 1. Return bloom ( floral bud density of spur and terminal buds number of clusters per cm 2 branch cross-sectional area (BCA)) in spring of the second year, for sevenyear-old trees of Scifresh apple managed using annual Centrifugal Training set up in the previous season and Vertical Axe training (control treatment), where in the previous season, three crop loads were set according to fruit number per cm 2 BCA, within which fruit were thinned to either one or two fruit. Floral bud density (no. flower clusters per cm 2 BCA) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05): = ns; Management = 0.577; Fruit = 0.577; x tree management interaction = 0.99.

12 8 Floral Spur Characteristics Flowering buds from spur and terminal sites were dissected into leaf and floral components at mid bloom to determine if tree and crop management treatments from the previous season influenced the size and vigour of flowering spurs early in the ensuing season. The number of leaves on floral spurs was very similar among all treatments, within the range of six to eight leaves (data not presented). Total spur leaf area of floral spurs and terminals at bloom did not differ, with a range of 22 to 24 cm 2. Differences in spur leaf area were associated with the interaction of tree management with crop load but not with the number of fruit (Table 2). Overall VA managed trees had spurs with greatest leaf area at bloom although declining in higher crop load treatments. This pattern was not evident with CT managed trees apart from having similar or smaller leaf area than VA trees for similar treatments. Although significant differences associated with tree or crop management were found, the absolute differences were not large and the typical leaf area of Scifresh spurs at bloom is small compared with other commercial apple cultivars (Wunsche & Palmer unpublished data). Table 2. Total spur leaf area (cm 2 ) of spur and terminal floral buds at the mid bloom period of the second year, for seven-year-old trees of Scifresh apple managed using annual Centrifugal Training set up in the previous season and Vertical Axe training (control treatment), where in the previous season, three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Total spur leaf area (cm 2 ) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05): = 2.53; Management = 2.07; Fruit = ns; x tree management interaction = The number of flowers in floral buds was slightly greater for terminal buds compared with spurs (a of 5.6 compared with 5.1 flowers) but flower number per bud was unaffected by tree management and crop load treatments. When the total dry weight of floral receptacles per bud was compared, terminals appeared to produce significantly larger total receptacle dry weight. When corrected for flower number per bud, however, it became clear that there was no difference in flower receptacle dry weight (~20 mg) among spur and terminal floral buds. Therefore, tree training and cropping treatments applied in the previous season had no effect on the morphology of floral buds when measured at bloom in the next season. This indicates that those treatments did not have any measurable effect on altering the size of individual floral buds of Scifresh in ways likely to improve initial fruit set.

13 9 Fruit Set Behaviour The patterns of fruit set of CT trees spur-thinned to provide four to six fruit per cm 2 BCA were very similar to patterns of fruit set recorded in the previous season. Only 15% to 20% of buds completely failed to set any fruit whilst 50% of buds set two or more fruit (Figures 1 3). The pattern of fruit set was more-or-less consistent across the three crop levels set by CT management. Frequency distribution CT4x1 CT4x2 VA4x1 VA4x Sp0 Sp1 Sp2 Sp3 Sp4 Sp5 Sp6 Fruit no. Figure 1. The proportions of flowering spurs setting zero, one, two, three, four, five or six fruit on seven-year-old trees of Scifresh apple set to a spur density of four fruit per cm 2 branch cross-sectional area using Centrifugal Training (CT) and the comparable response from equivalent unmodified trees managed using conventional Vertical Axe (VA) tree management. In the treatment legend x1 refers to one fruit and x2 refers to two fruit within the crop level. Frequency distribution CT5x1 CT5x2 VA5x1 VA5x Sp0 Sp1 Sp2 Sp3 Sp4 Sp5 Sp6 Fruit no. Figure 2. The proportions of flowering spurs setting zero, one, two, three, four, five or six fruit on seven-year-old trees of Scifresh apple set to a spur density of five fruit per cm 2 branch cross-sectional area using Centrifugal Training (CT) and the comparable response from equivalent unmodified trees managed using conventional Vertical Axe (VA) tree management. In the treatment legend x1 refers to one fruit and x2 refers to two fruit within the crop level.

14 10 Frequency distribution CT6x1 CT6x2 VA6x1 VA6x Sp0 Sp1 Sp2 Sp3 Sp4 Sp5 Sp6 Fruit no. Figure 3. The proportions of flowering spurs setting zero, one, two, three, four, five or six fruit on seven-year-old trees of Scifresh apple set to a spur density of six fruit per cm 2 branch cross-sectional area using Centrifugal Training (CT) and the comparable response from equivalent unmodified trees managed using conventional Vertical Axe (VA) tree management. In the treatment legend x1 refers to one fruit and x2 refers to two fruit within the crop level. Fruit set of VA-managed trees described a different pattern from CT-managed trees. The proportion of floral spurs failing to set any fruit declined from 20% to 10% as crop treatment level increased from four to six fruit per cm 2 BCA (Figures 1 3). This contrasts with levels of 50% to 60% recorded in the previous year. At the same time the proportion of spurs setting multiple fruits increased from 50% to 80% of spurs over the same range of crop level (Figures 1 3). These fruit set responses appear to be related to differences in floral bud densities arising from return bloom responses of both CT and VA treatments (see Table 1). Return bloom in the second year was reduced, especially on VA trees, more-or-less in proportion to the previous season s increasing crop level. These responses in the second year appear to support our hypothesis that fruit set may be improved as a result of less competition for allocation of stored reserves in spring when there are fewer developing floral meristematic sites. In this second year, the return bloom responses of different treatments provided two scenarios to observe fruit set behaviour at lower floral densities: reduction by spur thinning with CT management; and reduction of floral spur density caused by biennial bearing in higher cropped VA managed trees. By combining fruit set data from 2006 and 2007, we have derived a generalised fruit set response model for Scifresh cultivar demonstrating an association with initial floral bud density (Figure 4). As floral bud density decreases there are two complementary responses, both which enhance fruit set; the reduction in the proportion of spurs that fail to set fruit at all and an increase in the proportion of spurs that exhibit strong fruit set. These relationships appear to support our hypothesis that excessive flowering may be a cause of poor and erratic fruit set because of intense competition for available reserves/assimilates early in the season. Further, this fundamental relationship demonstrates that the fruit set potential of Scifresh should be easily manipulated by simply modifying the initial floral bud density of branches/trees.

15 11 Buds failing to set (%) Buds setting 2+fruit (%) Floral bud density (no./cm2 bca) A B Floral bud density (no./cm2 bca Figure 4. The relationship between initial floral bud density (spurs and terminals only) and the proportion of buds that fail to set fruit (A) and the proportion of buds that set two or more fruit per bud (B) on Scifresh apple on M.9 rootstock. Floral bud densities are a combination of s of natural flowering of unmodified Scifresh trees (VA management) and from artificial spur extinction treatments to reduce floral spur numbers (CT management). Data from 2008 show lower floral bud densities caused by biennial bearing in some treatments. Trend lines (A= linear regression, B= polynomial regression) describe 2007 data. TREE PRODUCTIVITY At harvest, the number of fruit per tree was between 195 and 260 fruit among the tree management x crop load x fruit number treatments. This range was more compact than in 2007 as well as having a higher minimum number of fruit per tree and a lower maximum. The crop density of CT-managed trees was very comparable with the previous year (Table 3). In responses similar to the first year, the final crop density of CT trees was somewhat lower than planned because 15 20% the retained buds failing to set any fruit and possibly some spurs did not flower. A second trend noted in the previous year was again evident in the second season, where the crop density of CT-managed trees did not increase by

16 12 the increments expected according to the three set levels of crop load. Despite this, overall crop density did increase in response to increasing crop load treatments. Compared with the first year, VA-trained trees expressed a marked decrease in crop density which was particularly evident at the higher crop load treatments where biennial bearing caused reduced flowering (Tables 1 and 3). As a result the crop density of VA trees was similar or only slightly greater than for CT trees. Table 3. Crop density (number of fruit per cm 2 trunk cross-sectional area (TCA)) at harvest from seven-year-old trees of Scifresh apple managed using annual Centrifugal Training, initially set up at budbreak in the previous season and Vertical Axe training, in which three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Crop density (no. fruit per cm 2 TCA) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05): = ns; Management = 0.53; Fruit = ns; x tree management interaction = 0.46 Table 4. Mean total tree yield (kg) at harvest from seven-year-old trees of Scifresh apple managed using annual Centrifugal Training initially set up in the previous season and Vertical Axe training (control treatment), where three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Mean tree yield (kg) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05): = ns; Management = ns; Fruit = ns; x tree management interaction = 6.5

17 13 Mean total yield per tree ranged between 34 and 49 kg per tree among the management x crop load x fruit number treatments in the second year (Table 4). As well as being less variable than in the first year, most crop treatments fell into a narrow range of kg per tree. Converting tree yields to tonnes per hectare indicated productivity from 56 to 80 t/ha, which represented 3000 to 4300 gross carton equivalents per hectare (tce). These yields demonstrate that using the range of cultural practices investigated in this study, orchard productivity comparable with other commercial cultivars was readily achievable with Scifresh apple. It is also worth highlighting that these yields were derived from whole-tree crop densities in the range of only 3.9 to 6.1 fruit per cm 2 TCA. Considering fruit were selectively borne on spurs and terminals, and axillary fruit were completely eliminated, it is evident that floral spur manipulation treatments have practical possibilities for optimising productivity of Scifresh apple to overcome cropping problems caused by the tendency for excessive floral initiation. FRUIT QUALITY Mean fruit weight at harvest differed among treatments by as much as 36 g (compared with 23 g in 2007) and was responsive to the combination of crop load and tree management factors (Table 5). Fruit from CT treatments were generally g larger than from the previous season and were larger than expected when compared with the previous year s fruit size responses to crop load. In contrast, fruit weight from some VA treatments was smaller than anticipated, in particular at the lower crop densities. Table 5. Mean fruit weight (g) at harvest from seven-year-old trees of Scifresh apple managed using annual Centrifugal Training initially set up in the previous season and Vertical Axe training (control treatment), where three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Mean fruit weight (g) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05) = 3.7; Tree management = 3.0; Fruit number = 3.0; x tree management interaction = ns; x tree management x fruit interaction = 7.4. Carryover effects of cropping and tree management treatments from the previous year are thought to have contributed to fruit size responses. For CT treatments, fruit weight was relatively constant between g as crop density increased from 3.9 to 5.5 fruit per cm 2 TCA, with buds cropped as either one or two fruit (Figure 5). Improved fruit size could be, in part, the result of CT management inducing greater numbers of terminal buds on

18 14 short to medium length shoots in the previous season (Tustin unpublished data). Terminal buds are known to produce significantly larger fruit than spur or axillary buds on Scifresh apple (Breen et al. 2007). 210 Mean fruit weight (g) CT singles CT doubles VA singles VA doubles Crop density (no.per cm2 TCA) Figure 5. Mean fruit size (g) response to whole tree crop load (number per cm 2 trunk crosssectional area (TCA)) with Centrifugal (CT)-trained and Vertical Axe (VA)-trained trees of Scifresh apple, in which crop loads were set by thinning to either one (singles) or two (doubles) fruit. Data are responses (n=7) measured in the second year after commencing Centrifugal tree management. With VA treatments, those which showed greatest effects of biennial bearing in their return bloom and final fruit density had fruit of significantly smaller size, which was also smaller than expected when compared with the previous season s fruit size response to crop load (Figure 5). In these treatments, where branch units lacked sufficient spur and terminal fruit when thinned to their final crop load, axillary fruits were retained to achieve the respective branch fruit numbers for the crop load treatments. It is possible that retaining a small proportion of axillary fruit contributed to smaller fruit weight in VA treatments. Breen et al. (2007) reported that fruit at harvest from axillary buds were slightly smaller in size than fruit from spurs, when sampled from trees cropped using commercial methods of selective hand thinning. Having found cumulative effects of tree management and crop load influencing potential yield and fruit weight in the second year, did treatment differences also affect other fruit quality traits? Fruit traits used to indicate harvest maturity (SPI and fruit background colour) and those which are considered to be quality determinants (percentage blush coverage) showed very small differences among treatments. Differences among treatment factors that were significant were so small that they had no practical commercial implications. Fruit were harvested at a SPI of 3.3, background colour scale of 5 6 with average blush coverage of 65 70%.

19 15 Overall, fruit soluble solids concentrations (SSC) at harvest did not differ between the CT and VA management treatments in 2008 (Table 6). Compared with the previous year, SSC of fruit from VA-trained trees was 0.5 to 1% higher, which may be explained by the reduction in whole-tree crop load. Fruit SSC from CT treatments were very similar in both years. Unlike with the crop in year one, SSC at harvest was not strongly related to whole-tree crop load but showed a reasonably strong relationship with total tree yield. Overall fruit SSC levels in the second year were in the high end of the range expected for Scifresh apple. A commercially important outcome was that in the second year, especially with CT management, fruit size was markedly increased whilst at the same time these fruit attained SSC that were in the upper range for Scifresh apple. Taking 2007 and 2008 production together, SSC of Scifresh apple fruit at harvest appears to be largely determined by the size of the annual crop borne by the tree. Table 6. Fruit soluble solids concentration (%) at harvest from seven-year-old trees of Scifresh apple managed using annual Centrifugal Training initially set up in the previous season and Vertical Axe training (control treatment), where three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Fruit soluble solids concentration (%) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05) = ns; Tree management = ns; Fruit number = 0.17; x tree management interaction = 0.3. Fruit firmness in the second year was higher across all treatments than in the previous year. As well, fruit from VA-trained trees was generally firmer than from CT managed trees (Table 7). The greatest differences in fruit firmness between years occurred where crop load was substantially reduced in the second year compared with in the first year. This is evident within the VA treatments that had high crop loads in 2007 and showed obvious biennial bearing at flowering in 2008, resulting in reduced yields from the combination of lower crop load together with smaller fruit size. Fruit from these treatments had uniformly very high firmness around 10 kgf irrespective of crop load or tree yield. Fruit firmness from CT treatments was on average 0.5 kgf less firm than VA treatments. Within the CT crop load treatments, fruit firmness declined as crop load increased across the limited range of crop loads experienced (adjusted R 2 =0.64, P=0.03).

20 16 Table 7. Fruit flesh firmness (kgf) at harvest from seven-year-old trees of Scifresh apple managed using annual Centrifugal Training initially set up in the previous season and Vertical Axe training (control treatment), where three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Fruit flesh firmness (kgf) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05) = ns; Tree management = 0.16; Fruit number = ns; x tree management interaction = ns. Differences in titratable acidity were evident between CT and VA management, following the same trends as other quality traits that were identified to be associated with the effects of biennial bearing on cropping in VA treatments that had higher crop loads in 2007 (Table 8). Titratable acidity levels were lower in fruit from CT trees than VA trees and also lower than from CT treatments in the previous season. However, many of the quantitative differences in titratable acidity between treatment factors were less than 0.08%, the smallest difference that trained consumer panellists can perceive as a difference in acid taste between apples (Harker et al. 2001). Notwithstanding that, the effect of the ratio of sugars to acid is important for taste perception and the intense taste characteristics of Scifresh is important to its market differentiation and is associated with both sweetness and acidity characteristics.

21 17 Table 8. Titratable acidity (% malic acid) at harvest from seven-year-old trees of Scifresh apple managed using annual Centrifugal Training initially set up in the previous season and Vertical Axe training (control treatment), where three crop loads were set according to fruit number per cm 2 branch cross-sectional area (BCA), within which fruit were thinned to either one or two fruit. Fruit titratable acidity (% malic acid) Centrifugal management Vertical Axe management no./ cm 2 BCA Mean Mean Four fruit Five fruit Six fruit Management Fruit per spur LSD (P=0.05) = ns; Tree management = 0.03; Fruit number = 0.03; x tree management interaction = GENERAL COMMENTS A feature of this study was the contrasting management techniques between the conventional VA and the experimental CT methods. For experimental purposes, a number of standard practices were followed. In particular, no use of chemical thinning and withholding hand thinning until after final fruit drop in late November enabled natural fruit set to be measured. These practices have had some influence on results especially with respect to return flowering and degrees of biennial bearing in the second year. This was observed to affect VA-managed trees more severely. Scifresh trees managed using CT appeared to flower abundantly in the second year and indeed, one objective of the development of CT in France was to overcome biennial bearing tendencies. In the second year, CT management utilised floral spurs that had developed under the CT regime during the previous season. We think that results in the second year may demonstrate optimal responses to CT management because flowering spurs had developed within the CT environment where competition among individual spurs may be less because of the reduction in the total number of spurs per branch. A second response to CT management is that a greater proportion of floral buds selected in the second year were probably terminal buds of short and medium shoots. Artificial spur extinction to a predetermined spur number per branch is the technique of CT management intended to promote spur development, in particular to stimulate the development of short-to-medium length bourse shoots that terminate into a strong floral bud (Lauri et al. 2004). With Scifresh where it has been found that terminal floral buds produce significantly larger fruit than spurs or axillary buds, a component of the fruit size response and improvements in fruit quality attributes from CT management are likely to be associated with a proportionally greater contribution to the total crop from floral buds of terminal shoots. Under CT management it is practical to thin by hand very early in the season and this may enable the complete avoidance of chemical thinning without high risk of biennial bearing. The use of chemical thinning in association with CT management has not yet been examined

22 18 and requires investigation. With enhanced fruit set through CT it is possible that more predictable chemical thinning responses may be achievable compared with standard VA tree management where chemical thinning responses can be erratic, unpredictable and easily overdone. CONCLUSIONS Several novel tree and crop management approaches have now been examined over two consecutive years as techniques for improving the fruit size, quality and productivity of Scifresh apple based on enhancing fruiting spur quality and productivity. These included: Setting crop levels using a branch level crop load model rather than a whole tree method Examining techniques of Centrifugal Training that set crop density by regulating the distribution and density of fruiting spurs, set up in late winter in the tree canopy, using spur removal (artificial spur extinction) that largely substitutes for pruning Evaluating the influence of cropping spurs with either one or two fruit within the two crop management approaches CT compared with standard VA. The main outcomes found after two consecutive years of this study were that: 1. Fruit set of individual spurs observed in the first year to be strongly enhanced by CT management was again evident in the second year. The combined fruit set responses of CT- and VA-managed treatments from two consecutive years has enabled the development of a fundamental fruit set response model for Scifresh apple. The fruit set response model provides strong evidence for the hypothesis that the excessive flower initiation of Scifresh is a fundamental cause of reduced fruit set and erratic cropping including variable fruit size. The model demonstrates that fruit set is enhanced in proportion to reduction in initial floral bud density. Reducing floral bud density rather than enhancing new floral spur development appeared to be the main mechanism by which CT improved fruit set and fruit development. 2. Mean tree yields of 34 to 49 kg equated with productivity of between 56 and 80 t/ha gross yields ( tce) in the second year. These yields are easily comparable with most commercial cultivars, demonstrating that under the range of cultural practices examined in this study, high orchard productivity is achievable with Scifresh apple. The yields of 56 to 80 t/ha were achieved from surprisingly low final crop densities in the range of 3.9 to 6.1 fruit per cm 2 TCA. 3. Mean fruit size was largest among the CT crop load treatments and was consistently between g across crop densities from 3.9 to 5.5 fruit per cm 2 TCA, in the second year. It is thought that the fruit size responses with CT management are partly because a larger proportion of floral buds may have been terminals of short to medium length shoots the buds known to produce significantly larger fruit than any other bud type of Scifresh. 4. Fruit traits that are used as harvest maturity attributes were not affected by any of the treatment factors used the study.

23 19 5. Fruit from the second year exhibited similar or improved soluble solids concentration, fruit firmness and titratable acidity compared with the first year. In general fruit produced under CT and VA management systems had similar internal qualities, and where differences were found they appeared to be largely related to effects of biennial bearing from VA treatments. The internal fruit quality from CT-managed trees was maintained or improved in parallel with increases in fruit size. RECOMMENDATIONS Results of novel tree and crop manipulation studies over two consecutive years have provided important insight of Scifresh fruit set behaviour and have demonstrated new orchard management techniques with potential to improve Scifresh apple fruit size, quality and orchard productivity. There remains further significant opportunities to confirm the most important and valuable responses to these new management methods from continuing the study through its third consecutive year. In particular in the third year, VA trees will produce an on-flowering response with very high floral bud densities and crop load potential and CT trees will provide responses induced by three successive years of CT management that will add considerably to the information already developed. These are important considerations for ensuring the technologies are sufficiently well developed to actively promote their uptake and implementation, as the definitive step to significantly improve and optimise the fruit size profile of Scifresh for market requirements.

24 20 REFERENCES Breen K, Palmer J, Birken E, Tustin, S, Henriod R, Seymour S, Diack R, Dayatilake D, Oliver M Influence of fruiting bud structure on fruit size and quality in Scifresh apple. NZ Institute of Agriculture and Horticulture Science Convention, August, Lincoln University, Canterbury. Abstract 3.7. Harker FR, Marsh KB, Young H, Murray SH, Gunson FA, Walker SB Sensory interpretation of instrumental measurements 2: sweet and acid taste of apple fruit. Postharvest Biology and Technology 24: Lauri P-E, Lespinasse J-M Apple tree training in France: current concepts and practical implications. Fruits 54: Lauri P-E, Willaume M, Larrive G, Lespinasse J-M The concept of centrifugal training in apple aimed at optimising the relationship between growth and fruiting. Acta Horticulturae 636: Tustin DS, Palmer JW, Henriod R, Dayatilake D, Breen K, Seymour SM, Diack R Scifresh fruit size and quality: Improving fruit size via improved spur quality. Report to ENZA Ltd. HortResearch Client Report No p.