HORTSCIENCE 53(7):968–977. 2018. https://doi.org/10.21273/HORTSCI12925-18 density can decrease orchard profitability (Lordan et al., 2017a; Robinson, 2008a; Robinson et al., 2007). Dwarfing apple root- Long-term Effects of Training Systems stocks, especially ‘M.9’ and ‘M.26’, have made possible the transition of entire fruit- and Rootstocks on ‘McIntosh’ and growing sectors to higher tree density and training systems over the last 50 years. New ‘Honeycrisp’ Performance, a 15-year varieties such as Honeycrisp require a reeval- uation of promising rootstocks because the scion cultivar has low vigor (Robinson et al., Study in a Northern Cold Climate— 2011b). The Geneva Apple Rootstock Breed- ing Program has developed rootstock geno- Part 1: Agronomic Analysis types that are better adapted to biotic stresses 2 common in eastern North America (Cummins Jaume Lordan and Aldwinckle, 1983; Fazio et al., 2015). Department of Horticulture, NYSAES, Cornell University, Geneva, NY 14456 Among them, ‘G.30’ has been especially useful 1 in cold climates with short growing seasons, Anna Wallis whereas ‘G.16’ has a similar growth and vigor Cornell Cooperative Extension, Eastern NY Commercial Horticulture to ‘M.9’ clones (Autio et al., 2011; Robinson Program, Cornell University, Plattsburgh, NY 12901 and Hoying, 2004; Robinson et al., 2003). Most orchards in New York State are Poliana Francescatto and Terence L. Robinson replanted on old orchard sites. The severity Department of Horticulture, NYSAES, Cornell University, Geneva, NY 14456 of apple replant disease (ARD) is not known 3 although research efforts by Merwin et al. Additional index words. crop load, fruit quality, Malus domestica, tree size, yield, yield (2001) have attempted to assess the prob- efficiency, cold climate apple, eastern North America lem. In addition, recent regulations on fu- Abstract. Choice of cultivar, training system, planting density, and rootstock affect migation have resulted in not a single orchard performance and profitability. To provide guidance to growers in northern cold licensed fumigation company working in climates on these choices, a field trial was established in Peru, Clinton County, NY, in New York State. Rootstock resistance to ARD could be a more sustainable solution to 2002, with two apple cultivars (Honeycrisp and McIntosh). From 2002 through 2016, we Ò compared Central Leader on ‘M.M.111’; Slender Pyramid on ‘M.26’ and ‘GenevaÒ 30’ ARD than fumigation. Several of the Geneva (‘G.30’); Vertical Axis on ‘M.9 (NicÒ 29)’ (‘M.9’), ‘Budagovsky 9’ (‘B.9’), and ‘G.16’; rootstocks have shown some tolerance to SolAxe on ‘M.9’, ‘B.9’, and ‘G.16’; and Tall Spindle on ‘M.9’, ‘B.9’, and ‘G.16’. Central ARD, especially ‘G.30’ (Fazio et al., 2015; Leader was planted at 539 trees/ha, Slender Pyramid at 1097 trees/ha, Vertical Axis and Kviklys et al., 2016; Robinson et al., 2012). SolAxe at 1794 trees/ha, and Tall Spindle at 3230 trees/ha. Cumulative yield was higher Northern cold climates with fewer heat with ‘McIntosh’ than with ‘Honeycrisp’. High planting densities (Tall Spindle) gave the units result in slower tree growth and usually highest cumulative yields (593 t·haL1 on ‘McIntosh’ and 341 t·haL1 on ‘Honeycrisp’). lower yield than more temperate climates Tall Spindle (3230 trees/ha) on ‘M.9’ appeared to be the best option for ‘McIntosh’. On with more heat units (Robinson et al., 2011a). the other hand, for a weak-growing cultivar such as ‘Honeycrisp’, Tall Spindle on ‘B.9’ This has meant that much of the data on (366 t·haL1) and Slender Pyramid (1097 trees/ha) on ‘G.30’ (354 t·haL1) were the two orchard system performance is not directly combinations with the highest cumulative yield, largest fruit size (220–235 g), and applicable to colder growing regions. greatest efficiency index (4.6–3.9 kg·cmL2). Because there are many different factors that affect orchard profitability (Badiu et al., 2015; Balkhoven-Baart et al., 2000; Bradshaw et al., 2016; Lordan et al., 2018b; Sojkova and The success of a new orchard investment 2007; Sansavini and Musacchi, 2002; Walsh Adamickova, 2011; Weber, 2001) there is depends on yield performance and the market et al., 2011; White and DeMarree, 1992). a need to evaluate performance for different value of the crop. Therefore, factors such as Location is another key factor that deter- situations of climate cultivar, training system, cultivar, rootstock, training system, planting mines orchard success; for instance, localized rootstock, planting density, location, and density, precocity, and fruit quality, and price weather of a site can enhance biotic stresses economic condition. play key roles in overall orchard profitability favoring diseases such as fire blight (Erwinia This project was intentionally performed (Bravin et al., 2009; DeMaree, 1995; DeMarree amylovora Burill) or winter damage, com- in a cold climate on old orchard soil without et al., 2003; Elkins et al., 2008; Goedegebure, promising orchard productivity. ‘McIn- fumigation to simulate the actual orchard 1993; Heijerman et al., 2015; Robinson et al., tosh’, because of its cold-hardiness, is one replanting carried out in northern New York of the main cultivars grown in northeastern State and to determine if any combination of United States and eastern Canada (Ferree rootstock and planting system could be suc- and Warrington, 2003). In recent years, cessful in both the climatic conditions and Received for publication 26 Jan. 2018. Accepted for publication 24 Apr. 2018. ‘Honeycrisp’ has been suggested as one of the old orchard soil conditions of northern This research was partially supported by the New the few cultivars which could challenge New York. In addition, the aim of this study York Apple Research Development Program. ‘McIntosh’ as the most widely grown in was to evaluate the orchard performance of We wish to thank Everett Orchards for all the New England (Greene and Weis, 2001). ‘Hon- the new cold-hardy cultivar, Honeycrisp, resources invested in this trial. We thank Peter eycrisp’ also has a much higher fruit price in compared with the standard cold hardy cul- Herzeelle for field and laboratory support. The the market than ‘McIntosh’, which has tivar, McIntosh, at a wide range of planting authors also acknowledge Kevin Iungerman, Cor- a large impact on orchard profitability. densities, training systems, and rootstocks for nell Cooperative Extension, Eastern NY Commer- Although several studies have reported cold areas. cial Horticulture Program for his contributions to a positive relationship between yield and this research. 1Currently at Plant Pathology and Plant-Microbe planting density (Elkins and Dejong, 2002; Materials and Methods Biology, Plant Science, Cornell University, Ithaca, Kappel and Brownlee, 2001; Robinson, NY 14850. 2008b; Sansavini and Musacchi, 2002; Trial site and design. In 2002 an on- 2Corresponding author. E-mail: jl3325@cornell. Vercammen, 1999), it is worth noting that farm field trial was planted in Peru, edu. there is a point where increasing planting Clinton County, NY (lat. 44.597223 , long.
968 HORTSCIENCE VOL. 53(7) JULY 2018 –73.545078 ), with two apple cultivars (Hon- a steel conduit pipe (3 m with 15 cm in the each tree, and fruit size was then calculated eycrisp and McIntosh). We compared the ground) at each tree (Robinson and Hoying, using measured data (Marini et al., 2006). following (12) production systems: Central 1999). TS systems were supported by a 3- Trunk circumference (cm) at 30 cm above Leader (CL) on ‘M.M.111’ rootstock; Slen- wire trellis (2.5 m). graft union was measured at the end of the der Pyramid (SP) on ‘M.26’ and ‘GenevaÒ CL trees were developed by initially trial (2016) (Reig et al., 2018). Trunk cross- 30’ (‘G.30’); Vertical Axis (VA) on ‘M.9 heading each tree at 70 cm and thereafter sectional area (TCSA, cm2), cumulative yield (NicÒ 29)’ (‘M.9’), ‘Budagovsky 9’ (‘B.9’) annually heading the leader and lower tier efficiency (kg·cm–2), and crop load (fruit and ‘G.16’; SolAxe (SA) on ‘M.9’, ‘B.9’, and scaffolds by one-third each year to produce number/cm2) were then calculated (Marini ‘G.16’; and Tall Spindle (TS) on ‘M.9’, a strong sturdy trunk and branch framework et al., 2006). Absolute value for biennial ‘B.9’, and ‘G.16’ (Table 1). CL was planted with permanent branches. Two permanent bearing index (BBI) was calculated as at 539 trees/ha, SP at 1097 trees/ha, VA and tiers of scaffold branches were developed, follows: SA at 1794 trees/ha, and TS at 3230 trees/ha spaced 1 m apart, with four to five branches ðÞ yield year n ðÞyield year n þ 1 (Robinson, 2003). Experimental design was per tier. Limb spreading was carried out in BBI ¼ ; a randomized complete block with a split- years 3 and 5 using wooden limb spreaders. ðÞþyield year n ðÞyield year n þ 1 split plot and three replications. Training Scaffold branches between the first and system was the main plot with each main second tiers were removed between the years where 0 indicates no alternate bearing and 1 plot consisting of two (CL), four (SP), or six 5 and 6. Tree height was limited to 5 m. complete alternate bearing. Annual BBI was (VA, SA, and TS) 40-m-long rows. Cultivar SP and VA trees were developed by calculated for each year and the average of all was the subplot, which consisted of one (CL), heading the leader at 120 cm above the graft years calculated. From 2007 onward, an 18- two (SP), or three (VA, SA, and TS) 40-m- union at planting and shortening each feather kg fruit sample from each subplot was col- long rows. Rootstock was the sub-subplot by one-third their length. In years 2 through lected for fruit quality assessment. Fruit with each sub-subplot consisting of a single 6, leaders were not headed. In year 3, four to color, as a percentage of skin surface colored 40-m row. There were 72 rows of 40 m each five lower scaffold branches were tied down red (according to USDA color standards) was with 24 rows in each replication containing to horizontal. Beginning in year 4, large measured with an electronic weight size/ randomized main plots (system), randomized diameter limbs (>5 cm) were removed back color sorter (MAF Industries, Travers, CA). subplots (cultivar) within the main plots, and to the trunk with an angled cut to develop Flesh firmness (Fruit Texture Analyzer; QA randomized sub-subplots (rootstock) within replacement limbs. Each year, two to three Supplies LLC, Norfolk, VA) and soluble the subplots. The treatment structure was an large branches were removed. Tree height solids content (Atago USA, Inc., Bellevue, incomplete factorial of four systems, two was limited to 4 m. With the VA, lateral WA) were also assessed (Torres et al., 2017). cultivars, and six rootstocks with only 24 branches were kept simple by removing sub- Fruit were harvested in two picks at com- out a possible 48 combinations of cultivar · lateral branches to create a single axis for mercial maturity when red color was at least training system · rootstock. The number of each branch. With the SP, sublateral branches 50% and a starch index of 5 for ‘McIntosh’ trees in each sub-subplot depended on tree were allowed to remain but were removed if and 7 for ‘Honeycrisp’ using the Cornell spacing and varied from 13 trees/row for CL, they became as large as the main axis of the generic starch chart (scale 1–9). All the 18 for SP, 25 for SA and VA, and 38 for TS. lateral branch. systems and rootstocks were harvested on Soil type was a Bombay loam with good SA trees were developed by heading the the same date. Cumulative yield was then water–holding capacity, moderately well leader at 120 cm above the graft union at calculated, whereas fruit quality was assessed drained and fertile (USDA, 2018). The site planting, removing one to three of the largest at the main harvest date, usually the first pick. had previously been planted with apple trees feathers, and leaving the remaining feathers Statistical analysis. Response variables on seedling rootstock spaced 6 · 12 m for 40 unpruned. In years 2 through 4, the leaders were modeled using linear mixed effect years and was not fumigated. The replant were not headed. Beginning year 4 and models. Mixed models including each com- disease severity at the trial site was not continuing in years 6 and 8, scaffold branches bination of cultivar · training system · evaluated before planting, but field fumiga- longer than 1 m and originating above 120 cm rootstock as fixed factors and block as a ran- tion trials in the same county by Merwin et al. height on the trunk were tied down below dom factor were built to compare treatment (2001) showed low to moderate replant horizontal ( 120 from vertical), and one to effects between cultivars. Mixed models in- disease pressure for similar soils. The soil two scaffold branches originating below 120 cluding each combination of training system · was tilled (to remove old roots) and then cm on the trunk were removed each year until rootstock as fixed factor and block as a ran- cropped with sudan grass [Sorghum drum- no branches were left below 120 cm. Tree dom factor were built to separate treatment mondii (Steud.) Millsp. & Chas] for 2 years height was limited to 4 m by bending the top effects for TCSA, average fruit size, cumu- after removal of the old orchard before of the tree horizontal at 4 m height in year 8. lative yield, cumulative yield efficiency, planting the experiment. The plot was laid Sublateral branches on the lateral branches cumulative crop load, and BBI for each out perpendicular to the old orchard rows, were allowed to develop producing a highly cultivar. Crop load was included as a covar- thus, giving a balanced effect of the location branched scaffold with ‘‘fingers.’’ iate to adjust fruit size. Mixed models in- of the old trees on each rootstock and system. TS trees were developed by heading the cluding each combination of training system The trees were trickle irrigated as needed leader at 150 cm above the graft union at · rootstock as fixed factor, and block nested during the growing season using the Cornell planting, removing one to two of the largest to year as a random factor were built to Apple Irrigation Model based on a modified feathers, and leaving the remaining feathers separate treatment effects for firmness, solu- Penman–Monteith equation (NEWA.org) unpruned. In years 2 through 4, the leaders ble solids, and color for each cultivar. Ana- (Robinson et al., 2017) and received from were left unheaded. Beginning in year 3, lyses for the systems (SA, VA, and TS) that 200 to 1000 mm of water depending on the large-diameter limbs (>2 cm) were removed had the same common rootstocks (‘B.9’, year. Average annual precipitation for Peru, back to the trunk with an angled cut to ‘G.16’, and ‘M.9’) were also performed. NY, is 812 mm. The trial was managed with develop replacement limbs. Each year one Therefore, mixed models including system conventional pesticides and fertilizers accord- to two branches larger than 2 cm were or rootstock as fixed factor, and block nested ingtoindustrystandards. removed. Tree height was limited to 3.5 m. to system or rootstock as a random factor Tree management. All of the trees had Only small lateral branches (<2 cm) were were built to separate treatment effects for three to five feathers (lateral branches pro- allowed to remain in the tree and they were TCA, average fruit size, cumulative yield, duced the previous year) at planting. CL each kept simple by removing sublateral cumulative yield efficiency, cumulative crop system was supported by a steel conduit pipe branches to create a single axis for each load, and BBI for each cultivar. Crop load (3 m with 0.6 m in the ground) at each tree. branch. was included as a covariate to adjust fruit SP, VA, and SA trees were supported by Agronomic assessments. Yield (kg) and size. Mixed models including each combina- a single-wire trellis (2.3 m) connected to fruit number were recorded annually from tion of system or rootstock as a fixed factor
HORTSCIENCE VOL. 53(7) JULY 2018 969 and block nested to system or rootstock and package (Version 12; SAS Institute, Inc., trees (105 cm2), whereas all the systems that year as random factors were built to separate Cary, NC). had ‘B.9’ and ‘M.9’ rootstocks were the the treatment effects for firmness, soluble smallest (40–56 cm2). solids, and color for each cultivar. A mixed Results When comparing within systems that had model including each combination of train- common rootstocks, the smallest trees were ing system · rootstock, year, and training Tree size. Larger trees were observed with with TS whereas there was no difference system · rootstock · year as fixed factors, ‘McIntosh’ compared with ‘Honeycrisp’ between SA and VA. Among rootstocks and block as a random factor was built to (Table 2). For both cultivars, the biggest trees common across the three systems, ‘G.16’ separate the treatment effects for yield for of the trial were with CL systems on was bigger than ‘B.9’ and ‘M.9’ for both each cultivar. Data were square root trans- ‘M.M.111’, which were 200% to 300% larger cultivars (Table 2). formed to normalize data distribution. All than the rest of the treatments averaged. With Yield and fruit size. Yield was higher for mean separations were made by Tukey’s ‘Honeycrisp’, SP with ‘G.30’ trees were ‘McIntosh’ compared with ‘Honeycrisp’ honestly significant difference (P = 0.05). largest (70 cm2), whereas the smallest (29 (Table 2; Fig. 1). For ‘Honeycrisp’, the Residual analysis was performed to insure cm2) ones were with SA with ‘B.9’, and TS highest yields were observed for TS with that model assumptions were met. Data were with ‘B.9’ and ‘M.9’. With ‘McIntosh’, SP ‘B.9’ and SP with ‘G.30’; however, signifi- analyzed using the JMP statistical software with ‘G.30’ trees were also the second largest cant differences were only observed for SP with ‘M.26’ and CL with ‘M.M.111’, which had the lowest yields of the trial. For ‘McIn- tosh’, the highest yield was observed for TS Table 1. Spacing, tree planting density, training system, and rootstock evaluated at Peru, Clinton County, with ‘M.9’ (680 t·ha–1) and ‘B.9’ (608 t·ha–1). NY, 2002–16. The lowest yields were for CL (284 t·ha–1) Spacing (m) Planting density (trees/ha) Training system Rootstock and SP ( 364 t·ha–1). 3.05 · 6.08 539 Central Leader (CL) M.M.111 · When comparing within systems with 2.13 4.28 1,097 Slender Pyramid (SP) G.30 common rootstocks, VA and TS had the M.26 · greatest yields with ‘Honeycrisp’ (306– 1.52 3.67 1,794 SolAxe (SA) B.9 –1 G.16 341 t·ha ), and there were no significant M.9 (NicÒ 29) differences among the three rootstocks in Vertical Axis (VA) B.9 common across systems (Table 2). With G.16 ‘McIntosh’, higher yields were observed for M.9 (NicÒ 29) TS (593 t·ha–1) compared with VA and SA 1.01 · 3.06 3,230 Tall Spindle (TS) B.9 ( 475 t·ha–1), whereas among rootstocks, G.16 –1 Ò ‘M.9’ had the greatest yields (586 t·ha ), M.9 (Nic 29) followed by ‘B.9’ (511 t·ha–1), and finally
Table 2. Final trunk cross sectional area (TCA, cm2), cumulative yield (t·ha–1), average fruit size (g), cumulative yield efficiency (kg·cm–2 TCA), cumulative crop load (fruit number/cm2 TCA), and biennial bearing index for each combination of training system (Central Leader—CL, SolAxe—SA, Slender Pyramid—SP, Tall Spindle—TS, and Vertical Axis—VA) and rootstock (‘M.M.111’, ‘B.9’, ‘G.16’, ‘M.9’, ‘G.30’, and ‘M.26’) for ‘Honeycrisp’ and ‘McIntosh’ at Peru, Clinton County, NY over 15 years (2002–16). Yield data from 2017 (spring frost) was not used to estimate the biennial bearing index. Means followed by different letters denotes significant differences (Tukey’s honestly significant difference, P # 0.05). <0.0001-H or <0.0001-M, significant with higher values for ‘Honeycrisp’ or ‘McIntosh’, respectively. Gray bars represent variable value.
970 HORTSCIENCE VOL. 53(7) JULY 2018 Fig. 1. Annual yields (t·ha–1) for each combination of training system (Central Leader—CL, SolAxe—SA, Slender Pyramid—SP, Tall Spindle—TS, and Vertical Axis—VA) and rootstock (‘M.M.111’, ‘B.9’, ‘G.16’, ‘M.9’, ‘G.30’, and ‘M.26’) for ‘Honeycrisp’ and ‘McIntosh’ at Peru, Clinton County, NY, 2002–16.
HORTSCIENCE VOL. 53(7) JULY 2018 971 Fig. 2. Regressions for annual yield (t·ha–1) to tree density (trees/ha) for each cultivar (Honeycrisp and McIntosh) at Peru, Clinton County, NY, 2002–16.
972 HORTSCIENCE VOL. 53(7) JULY 2018 Fig. 3. Regressions for annual cumulative yield (t·ha–1) to tree density (trees/ha) for each cultivar (Honeycrisp and McIntosh) at Peru, Clinton County, NY, 2002–16.
‘G.16’ had the lowest cumulative yield (451 The relationship between annual yield annual yield curves could be flat, negative, or t·ha–1) among the three rootstocks. and tree planting density was explored. In quadratic positive depending on years. This Yield was reduced in 2011 because of the early years (2–5) the relationship was was especially true for ‘Honeycrisp’, where a spring frost, leading to the highest annual nearly linear (Figs. 2 and 3). For ‘McIntosh’ biennial bearing would create situations yields the following year (2012) (Fig. 1). The the relationship in years 5–10 also had where the high-density system would have very large crop in 2012 affected return bloom a strong linear tendency but with a slight a low-crop year, whereas the low-density in 2013, especially for ‘McIntosh’, reducing quadratic shape, where the highest planting system would have a high-crop year. For the yield down to the level of the frost density had the highest annual yield. How- ‘McIntosh’ in most years after the first 5 affected year (2011). ever, as the orchard aged, the shapes of the years, the relationship had a strong quadratic
HORTSCIENCE VOL. 53(7) JULY 2018 973 shape with a maximum yield in the mid ‘G.16’ had lower crop load values than ‘M.9’ with ‘Gala’ (Robinson et al., 2003), confirm- densities. and ‘B.9’. ing the importance of testing rootstock A similar trend for each cultivar was Biennial bearing was more apparent on performance for each particular cultivar. observed when studying the relationship of ‘Honeycrisp’ than on ‘McIntosh’ (Table 2). Training system and planting density also cumulative yield to tree density for the entire For ‘Honeycrisp’, biennial bearing decreased affected tree vigor. The smallest trees were 15 years (Fig. 3). Cumulative yield was with increasing planting density, TS (3230 on TS (3230 trees/ha) or SA and VA (1794 highly correlated to planting density, with trees/ha) had the lowest BBI, followed by VA trees/ha). This agrees with previous studies higher R2 values for ‘McIntosh’ than for and SA (1794 trees/ha), SP (1097 trees/ha), (Lordan et al., 2018a; Robinson, 2008b), ‘Honeycrisp’. Cumulative yield of ‘McIn- and CL (539 trees/ha) the highest. Among which found the highest density plantings tosh’ after 15 years was double when planting rootstocks common to the same systems (TS, had smaller trees. This was likely because of density was 3230 trees/ha compared with 500 SA, and VA), ‘G.16’ induced higher biennial the limb removal pruning to contain trees in trees/ha. For ‘Honeycrisp’ cumulative yield bearing than ‘B.9’ or ‘M.9’ (0.19 vs. 0.15, a smaller planting space of the TS system was increased 1.75 times when planted at the respectively). that has a dwarfing effect (Robinson, 2007). highest density compared with the lowest For ‘McIntosh’, differences in biennial The dwarfing associated with the high- (539 trees/ha). The doubling of annual yield bearing were less clear, although the highest density TS system could also have been at the highest density compared with the values were still for CL, and tended to partially due to greater root competition lowest density was less apparent in later years decrease with increasing planting density. for water and nutrients, and greater parti- than at the early orchard stage. For ‘Honey- Biennial bearing was lower for TS (0.09) tioning of carbon into fruit. crisp’, there was a substantial increase in compared with SA (0.13), and lower on ‘M.9’ Yield, fruit size, yield efficiency, crop cumulative yield in 2009 and 2012 due to (0.1) than ‘G.16’ (0.13) (Table 2). load, and biennial bearing. Our study strong biennial bearing (low crops in 2008 Fruit quality. There were no significant showed higher yields of ‘Honeycrisp’ on and 2011), whereas for ‘McIntosh’ a large differences in fruit firmness or soluble solids ‘M.9’ and ‘B.9’ rootstocks vs. ‘G.16’ when increase in yield was observed in 2012 due to for ‘Honeycrisp’ (Table 3). When comparing trained on TS, VA, or SA in high planting a spring frost in 2011, which damaged rootstocks (‘M.9’, ‘B.9’, and ‘G.16’) com- densities (1794–3230 trees/ha). However, ‘McIntosh’ flowers. mon to the same three systems (TS, SA, and ‘B.9’ performance with ‘McIntosh’ was in- Larger fruit sizes were observed for ‘Hon- VA), fruit from ‘G.16’ trees had a slightly ferior to ‘M.9’. In previous studies, for eycrisp’ (222 g on average), compared with higher firmness compared with ‘B.9’ fruit different soil types and climatic conditions, ‘McIntosh’ (141 g) (Table 2). There were few (6.2 vs. 6 kg, respectively). Fruit red color ‘B.9’ has been reported to produce lower significant differences among treatments in was highest (52%) from CL trees with yield than ‘M.9’ (Autio et al., 2013; Bonany fruit size, especially for ‘McIntosh’. For ‘M.M.111’ and TS with ‘G.16’, and lowest et al., 2004; Lordan et al., 2016; Marini et al., ‘Honeycrisp’, fruit size was smaller for SP (42%) from SA and VA trees with ‘B.9’. 2006; Robinson et al., 2003). Its high perfor- on ‘M.26’. Among rootstocks, there were When comparing systems with common mance with ‘Honeycrisp’ in our study could larger fruit of ‘Honeycrisp’ on ‘M.9’ com- rootstocks, TS had higher (47%) color than be indicative of its relatively low biennial pared with ‘B.9’ (215 vs. 223 g, respectively) either SA (44%) or VA (45%). Color was best bearing index in our study, its adaptation to (Table 2). from trees grafted on ‘G.16’ (49%), followed the cold winters and cool growing climate of Yield efficiency, crop load, and biennial by ‘M.9’ (45%), and poorest on ‘B.9’ (43%). northern New York State or both. Although bearing. ‘Honeycrisp’ yield efficiency was the For ‘McIntosh’, few differences in fruit ‘B.9’ had similar tree growth as ‘M.9’ with highest for VA and SA with ‘B.9’ and ‘M.9’, quality among treatments were observed ‘McIntosh’, its yield was inferior to that of and for SP with ‘G.30’ (Table 2). The lowest (Table 3). Lower overall values for both fruit ‘M.9’, indicating a better adaptability of values were for CL with ‘M.M.111’ and TS with firmness and SS were observed for CL with ‘M.9’ with ‘McIntosh’. The lowest planting ‘G.16’. Similarly for ‘McIntosh’, the highest ‘M.M.111’, whereas higher values were mea- density (CL/‘M.M.111’, 539 trees/ha) had yield efficiencies were for SA, VA, and TS sured on ‘G.16’. Color was higher (64% and the lowest yields, followed by SP (‘G.30’ with ‘B.9’ and ‘M.9’, and the lowest values 60%) for SA with ‘G.16’ and VA with ‘G.16’, and ‘M.26’, 1097 trees/ha). Similar to tree were with ‘G.16’ and ‘M.M.111’. and lowest (45%) for CL with ‘M.M.111’. size, yield was related to rootstock and For ‘Honeycrisp’ within systems with When comparing systems with common root- planting density. ‘G.30’ and ‘M.26’, have common rootstocks, yield efficiency was stocks, color was the best for SA (61%), also been reported to give less yield than higher for SA and VA compared with TS followed by VA (55%), and TS (52%). The ‘M.9’ in previous studies (Autio et al., 2013; ( 4 vs. 3.4 kg·cm–2, respectively), whereas highest values were observed on ‘G.16’ (59%), Lordan et al., 2016; Marini et al., 2006; among rootstocks across all systems, ‘B.9’ followed by ‘M.9’ (56%), then ‘B.9’ (54%). Robinson et al., 2003). Although ‘M.26’ has had the highest value, followed by ‘M.9’, and relatively good winterhardiness (Robinson then ‘G.16’ (Table 2). For ‘McIntosh’, there Discussion et al., 2003), its poor performance in our were no differences among systems, whereas study could have been due to replant disease ‘M.9’ and ‘B.9’ had similar yield efficiencies, Tree size. In our trial, ‘Honeycrisp’ trees susceptibility (Kviklys et al., 2016) and to which were significantly higher than ‘G.16’ were smaller than ‘McIntosh’. Greene and less than optimum planting density. ( 5.6 vs. 2.8 kg·cm–2, respectively). Weis (2001) observed a weak growing habit Cumulative yield was highly correlated ‘Honeycrisp’ average crop load was of ‘Honeycrisp’ in New England, whereas with planting density in our study. For both higher for VA with ‘B.9’, followed by SA ‘McIntosh’ has been described as a moderately cultivars, the highest cumulative yields were with ‘B.9’, VA with ‘M.9’, and SP with vigorous cultivar (Ferree and Warrington, 2003), observed at 3230 trees/ha. Greater differ- ‘G.30’ (Table 2). The lowest values were supporting the differences in tree size that ences among planting densities were ob- observed for CL with ‘M.M.111’, VA with we observed. served for ‘McIntosh’ than ‘Honeycrisp’. ‘G.16’, and TS with ‘G.16’. For ‘McIntosh’, The biggest trees in this trial were ‘Mc- An interesting exception to the relationship the SA, VA, and TS with ‘B.9’ and ‘M.9’ had Intosh’ on ‘M.M.111’ and ‘G.30’. Both of planting density and cumulative yield was the highest crop loads and ‘G.16’, ‘G.30’, and rootstocks have been reported to be more with ‘Honeycrisp’ on ‘G.30’ rootstock. This ‘M.M.111’ the lowest. vigorous than ‘M.9’ and ‘B.9’ in previous combination had essentially the same yield as TS had a lower crop load than SA and VA studies (Robinson et al., 2003, 2011b; Russo the best performing system of TS on ‘B.9’ for ‘Honeycrisp’, whereas among rootstocks et al., 2007). ‘G.16’ was significantly more but with much lower planting density. This crop load were the highest for ‘B.9’, followed vigorous than ‘M.9’ and ‘B.9’ within the may indicate a good match between scion by ‘M.9’ and then ‘G.16’ (Table 2). No same training systems for both ‘Honeycrisp’ productivity and tree architecture, which differences among systems (TS, SA, and and ‘McIntosh’. ‘G.16’ tree size has been produced sufficient fruit-filled canopy. In VA) with common rootstocks were observed previously reported to be similar to ‘M.9’ size addition, ‘G.30’ was suggested to have an for ‘McIntosh’, whereas among rootstocks with ‘Jonagold’ but slightly more vigorous excellent adaptation to cold winters and cool
974 HORTSCIENCE VOL. 53(7) JULY 2018 Table 3. Average flesh firmness (kg), soluble solids ( Brix), and color (%) at harvest for each combination of training system (Central Leader—CL, SolAxe—SA, Slender Pyramid—SP, Tall Spindle—TS, and Vertical Axis—VA) and rootstock (‘M.M.111’, ‘B.9’, ‘G.16’, ‘M.9’, ‘G.30’, and ‘M.26’) for ‘Honeycrisp’ and ‘McIntosh’ at Peru, Clinton County, NY over 15 years (2002–16). Means followed by different letters denotes significant differences (Tukey’s honestly significant difference, P # 0.05). <0.0001-H or <0.0001-M, significant with higher values for ‘Honeycrisp’ or ‘McIntosh’, respectively. Gray bars represent variable value.
growing seasons (Robinson et al., 2003). that after 20 years, ‘McIntosh’ had higher the present study carried out in Peru, Clinton Training systems planted at mid-densities yields in high-density plantings (3500 trees/ County, NY. (1794 trees/ha), such as VA and SA, with ha). In the present study, average annual yield For both ‘Honeycrisp’ and ‘McIntosh’, either ‘B.9’ or ‘M.9’ had the highest yield for ‘McIntosh’ was 41 t·ha–1, whereas for biennial bearing was lower for TS with on efficiencies and crop loads with a low-vigor ‘Honeycrisp’ it was 25 t·ha–1.Theseyields either ‘M.9’ or ‘B.9’. These observations cultivar such as Honeycrisp; even SP (1097 are lower than reported by other studies (Autio coincide with our previous study with ‘Hon- trees/ha) on ‘G.30’ had one of the highest et al., 2011; Lordan et al., 2016; Robinson eycrisp’ (Lordan et al., 2017b), where we yield efficiencies with that cultivar. On the et al., 2011a). The shorter growing season in found no significant differences between other hand, for the moderate-vigor cultivar Northern New York, compared with other major ‘B.9’ and ‘M.9’. Furthermore, the more McIntosh, TS had similar yield efficiency as apple-producing areas such as Washington State vigorous the trees were in our study, the SA and VA with either ‘B.9’ or ‘M.9’, and or western New York, plus harsh climatic higher the biennial bearing they had: for higher cumulative yield. This is similar to conditions (very cold winters, late fall and spring instance, TS had both the lowest vigor and previous observations (Lordan et al., 2018a) frosts), might explain lower yields observed in biennial bearing. This suggests a better
HORTSCIENCE VOL. 53(7) JULY 2018 975 balance of vegetative growth and fruiting for winterhardiness but despite its good tree evaluation of five cultivars in two organic trees on TS rather than with CL systems. As growth and survival it was not as productive apple orchard systems in Vermont, USA, 2006– many authors suggest, equilibrium between as ‘M.9’ and ‘B.9’. These results are specific 2013. Acta Hort. 1137:315–321. fruiting and growth is key in reducing alter- to the cold climate in northern New York Bravin, E., A. Kilchenmann, and M. Leumann. nate bearing (Costes et al., 2006; Forshey and State. In climates that are warmer, the rela- 2009. Six hypotheses for profitable apple pro- duction based on the economic work-package Elfving, 1989). tively easy of management of very high- within the ISAFRUIT Project. J. Hort. Sci. When comparing fruit size, no important density systems such as the TS becomes Biotechnol. 84(6):164–167. effect of either training system or rootstock more difficult because of greater vegetative Costes, E., P. Lauri, and J. Regnard. 2006. Ana- was observed for either cultivar. Only ‘Hon- vigor. In those climates, weaker growing lyzing fruit tree architecture: Implications for eycrisp’ on SP with ‘M.26’ had smaller fruit. rootstocks would be preferred. tree management and fruit production. Hort. Other studies have observed no significant To sum up, although TS on ‘M.9’ would Rev. 32:1–61. differences in fruit size between ‘M.9’ and be the best option for new orchards, TS on Cummins, J.N. and H.S. Aldwinckle. 1983. Breed- ‘M.26’ for ‘Honeycrisp’ (Robinson et al., ‘B.9’ rather than ‘M.9’ might be the way to ing apple rootstocks, p. 294–394. In: J. Janick 2016; Russo et al., 2007), ‘McIntosh’ go when fire blight pressure is high. On the (ed.). Plant breeding reviews. Springer, Boston, (Autio et al., 2011), or ‘Gala’ (Autio et al., other hand, ‘G.30’ with SP (1097 trees/ha) MA. DeMaree, A. 1995. Critical factors in orchard 2013; Russo et al., 2007). might be the least risky option for weak- profitability. Compact Fruit Tree 28:92–94. Fruit quality. Internal quality was only growing cultivars that are planted in cold DeMarree, A., T. Robinson, and S. Hoying. 2003. slightly affected by training system and climates and short growing seasons, and Economics and the orchard system decision. rootstock, but red color was more affected when fumigation for ARD is not possible. Compact Fruit Tree 36:42–49. depending on the training system and root- Other newer GenevaÒ rootstocks that are not Elkins, R.B. and T.M. DeJong. 2002. Effect of stock combination. Smaller trees with less only similar in dwarfing as ‘M.9’, but also training system and rootstock on growth and vegetative growth, such as those on SA, VA, winterhardy and resistant to ARD and fire productivity of ‘Golden Russet Bosc’ pear and TS, had the greater overall red color. This blight could be even a better option. How- trees. Acta Hort. 596:603–607. coincides with other studies reporting sun- ever, these need long-term testing in cold Elkins, R.B., T.M. DeJong, K. Klonsky, and R. light distribution within the canopy as the climates. DeMoura. 2008. Economic evaluation of high density versus standard orchard configura- main factor affecting fruit color and fruit Fruit price, labor (cost and availability), tions; case study using performance data for quality (Awad et al., 2001; Sinoquet et al., and other economical factors will also de- ‘Golden Russet Bosc’ pears. Acta Hort. 2008). We also saw significant differences termine profitability of each combination of 800:739–746. among rootstocks regarding fruit color; how- training system and rootstock. Therefore, Fazio, G., T.L. Robinson, and H.S. Aldwinckle. ever, these differences were relatively small, economic studies should complement this 2015. The Geneva apple rootstock breeding 2% to 5% for ‘Honeycrisp’ and ‘McIntosh’, kind of trial, since the agronomic and eco- program. Plant Breed. Rev. 39:379. respectively (Table 3). Rootstock effect on nomic optimum combinations may not be the Ferree, D.C. and I.J. Warrington. 2003. Apples: fruit color appears to be related to rootstock same. Botany, production and uses. CABI Publishing, vigor, affecting vegetative growth, and thus Cambridge, MA. light distribution within the canopy. To coun- Literature Cited Forshey, C. and D. Elfving. 1989. The relationship between vegetative growth and fruiting in teract this, some authors have suggested Autio, W., T. Robinson, D. Archbold, W. Cowgill, apple trees. Hort. Rev. 11:229–287. summer pruning to improve fruit color C. Hampson, R.P. Quezada, and D. Wolfe. Goedegebure, J. 1993. Economic aspects of super- (Palmer, 1999; Robinson et al., 1991). 2013. ‘Gala’ apple trees on supporter 4, P.14, intensive apple orchards. Acta Hort. 349:285– and different strains of B.9, M.9 and M.26 294. Conclusions rootstocks: Final 10-year report on the 2002 Greene, D.W. and S.A. Weis. 2001. Evaluation and NC-140 apple rootstock trial. J. Amer. Pomol. growing of Honeycrisp in New England. Com- In the cold climate of this study, both Soc. 67(2):62–71. pact Fruit Tree 34(4):100–103. ‘McIntosh’ and ‘Honeycrisp’ performed well Autio, W.R., T.L. Robinson, B. Black, T. Bradshaw, Heijerman, G., P. Roelofs, and M. Groot. 2015. in high-density plantings. TS (3230 trees/ha) J.A. Cline, R.M. Crassweller, C.G. Embree, Profitability of the Dutch growing system of with ‘M.9’ was the combination that gave the E.E. Hoover, S.A. Hoying, K.A. Iungerman, ‘Conference’. Acta Hort. 1094:233–238. R.S. Johnson, G. Lang, M.L. Parker, R.L. Perry, Kappel, F. and R. Brownlee. 2001. Early perfor- highest cumulative yields with higher crop G.L. Reighard, J.R. Schupp, M. Stasiak, M. mance of ‘Conference’ pear on four training loads and yield efficiency indices, reduced Warmund, and D. Wolfe. 2011. Performance systems. HortScience 36:69–71. biennial bearing and with no significant of ‘Fuji’ and ‘McIntosh’ apple trees after 10 Kviklys, D., T.L. Robinson, and G. Fazio. 2016. differences in fruit size and quality. How- years as affected by several dwarf rootstocks in Apple rootstock evaluation for apple replant ever, with the weak-growing cultivar Honey- the 1999 NC-140 apple rootstock trial. J. Amer. disease. Acta Hort. 1130:425–430. crisp, the largest fruit, and the highest yield Pomol. Soc. 65(2):2–20. Lordan, J., S. Alegre, R. Montserrat, and L. Asin. efficiency and crop load were with ‘B.9’ Awad, M.A., P.S. Wagenmakers, and A. de 2017a. Yield and profitability of ‘Conference’ instead of ‘M.9’, although cumulative was Jager. 2001. Effects of light on flavonoid pear in five training systems in North East of not different from ‘M.9’. In addition, a mid- and chlorogenic acid levels in the skin of Spain. Span. J. Agr. Res. 15(3):e0904. density system such as SP (1097 trees/ha) ‘Jonagold’ apples. Scientia Hort. 88(4):289– Lordan, J., G. Fazio, P. Francescatto, and T.L. 298. Robinson. 2016. Effects of apple (Malus · with ‘G.30’ provided similar results as the Badiu, D., F.H. Arion, I.C. Muresan, R. Lile, and V. domestica) rootstocks on vigor and yield scion high-density option. This may be due to the Mitre. 2015. Evaluation of economic efficiency response. XI Intl. Symp. Integrating Canopy greater cold-hardiness of ‘G.30’ compared of apple orchard investments. Sustainability Rootstock Environ. Physiol. Orchard Systems. with ‘M.9’ or ‘B.9’. Furthermore, although 7(8):10521–10533. ISHS. some growers may use ‘M.26’ (slightly more Balkhoven-Baart, J., P. Wagenmakers, J. Bootsma, Lordan, J., G. Fazio, P. Francescatto, and T.L. vigorous than ‘M.9’) for weak-growing cul- M. Groot, and S. Wertheim. 2000. Develop- Robinson. 2017b. Effects of apple (Malus · tivars such as Honeycrisp, others may use B.9 ments in Dutch apple plantings. Acta Hort. domestica) rootstocks on scion performance due to its higher tolerance to fire blight. 513:261–270. and hormone concentration. Scientia Hort. However, as pointed out in this study, both Bonany, J., J. Carbo, M. Casals, R. Montserrat, and 225:96–105. ‘M.26’ and ‘B.9’ might be compromised in I. Iglesias. 2004. Performance of some clonal Lordan, J., P. Francescatto, L. Dominguez, and replant sites, especially where fumigation is apple rootstocks in Girona and Lleida (Cata- T.L. Robinson. 2018a. Long-term effects of lonia, NE-Spain) with special reference to fruit tree density and tree shape on apple orchard not possible such as New York State. As this quality. Acta Hort. 658:333–339. performance, a 20 year study—Part 1, agro- study has shown, ‘G.30’ which has some Bradshaw, T.L., L.P. Berkett, R.E. Moran, M.E. nomic analysis. Scientia Hort. 238:303–317. genetic tolerance of ARD could be a good Garcia,H.M.Darby,R.L.Parsons,S.L. Lordan, J., M. Gomez, P. Francescatto, and T.L. option in these scenarios. The poorer perfor- Kingsley-Richards, M.C. Griffith, S.C. Bosworth, Robinson. 2018b. Long-term effects of tree mance of ‘G.16’ was unexpected. It has good and J.H. Gorres. 2016. Long-term economic density and tree shape on apple orchard
976 HORTSCIENCE VOL. 53(7) JULY 2018 performance, a 20 year study—Part 2, eco- Robinson, T.L. 2003. Apple orchard systems, p. Penman-Monteith model. Acta Hort. 1150:245– nomic analysis. Scientia Hort. https://doi.org/ 345–407. In: D.C. Ferree and I.J. Warrington 250. 10.1016/j.scienta.2018.03.031. (eds.). Apples: Physiology, production and Russo, N.L., T.L. Robinson, G. Fazio, and H.S. Marini, R.P., J.L. Anderson, W.R. Autio, B.H. uses. CABI, Wallingford, Oxon, UK. Aldwinckle. 2007. Field evaluation of 64 apple Barritt, J. Cline, W.P. Cowgill, R.C. Crassweller, Robinson, T.L. 2007. Effects of tree density and rootstocks for orchard performance and fire R.M.Garner,A.Gauss,R.Godin,G.M. tree shape on apple orchard performance. Acta blight resistance. HortScience 42:1517–1525. Greene,C.Hampson,P.Hirst,M.M.Kushad, Hort. 732:405–414. Sansavini, S. and S. Musacchi. 2002. European J. Masabni, E. Mielke, R. Moran, C.A. Mullins, Robinson, T.L. 2008a. The evolution towards more pear orchard design and HDP management: A M. Parker, R.L. Perry, J.P. Prive, G.L. Reighard, competitive apple orchard systems in the USA. review. Acta Hort. 596:589–601. T.L. Robinson, C.R. Rom, T. Roper, J.R. Acta Hort. 772:491–500. Sinoquet, H., G. Sonohat, A-M. Potel, P. Monney, Schupp, E. Stover, and R. Unrath. 2006. Perfor- Robinson, T.L. 2008b. Performance of pear and and P-E. Lauri. 2008. Using virtual plants to mance of ‘Gala’ apple trees on 18 dwarfing quince rootstocks with three cultivars in four estimate light distribution at intra-canopy scale rootstocks: Ten-year summary of the 1994 high density training systems in the northeast- in apple trees: Method presentation and assess- NC-140 rootstock trial. J. Amer. Pomol. Soc. ern United States. Acta Hort. 800:793–801. ment. Acta Hort. 803:225–234. 60(2):69–83. Robinson, T.L., H. Aldwinckle, G. Fazio, and T. Sojkova, Z. and I. Adamickova. 2011. Evaluation Merwin, I.A., R. Byard, T.L. Robinson, S. Carpenter, Holleran. 2003. The Geneva series of apple of economic efficiency of the orchards invest- S.A. Hoying, K.A. Iungerman, and M. Fargione. rootstocks from Cornell: Performance, disease ment project with respect to the risk. Agric. 2001. Developing and integrated program for resistance, and commercialization. Acta Hort. Econ. 57:600–608. diagnosis and control of replant problems in 622:513–520. Torres, E., I. Recasens, J. Lordan, and S. Alegre. New York apple orchards. New York Fruit Robinson, T.L., A.M. DeMarree, and S.A. Hoying. 2017. Combination of strategies to supply Qrty 9(1):11–15. 2007. An economic comparison of five high calcium and reduce bitter pit in ‘Golden De- Palmer, J.W. 1999. Light, canopies, fruit and dollars. density apple planting systems. Acta Hort. licious’ apples. Scientia Hort. 217:179–188. 42nd Annual IDFTA Conf., Ontario, Canada. 732:481–489. USDA. 2018. Natural Resources Conservation Reig,G.,J.Lordan,G.Fazio,M.A.Grusak,S. Robinson, T.L., G. Fazio, S.A. Hoying, M. Service, United States Department of Agricul- Hoying, L. Cheng, P. Francescatto, and T. Miranda, and K. Iungerman. 2011a. Geneva ture. Web Soil Survey. 19 Mar. 2018.
HORTSCIENCE VOL. 53(7) JULY 2018 977