===== PAGE 1 =====
Journal of the American Pomological Society 62(3):89-97 2008
# Flowering and Fruiting Characteristics of Eight Pawpaw [Asimina triloba (L.) Dunal] Selections in Kentucky
KIRK W. POMPER1, SHERI B. CRABTREE2, DESMOND R. LAYNE3, AND R. NEAL PETERSON4
## Abstract
The North American pawpaw [Asimina triloba (L.) Dunal] is in the initial stages of commercialization; however, little information has been published concerning flowering and ripening periods of commercially available cultivars. Cultivars with late flowering or extended bloom period could be useful to growers for avoiding crop losses to late spring frosts. The objectives of this study were to determine if mature trees of commercially available pawpaw cultivars display: 1) concentrated bloom periods, 2) late flowering characteristics, 3) high fruit set, and 4) concentrated fruit ripening characteristics that were related to concentrated bloom periods. An orchard was planted in the spring of 1998 in a randomized block experimental design with 8 replicates. In 2004 to 2006, the cultivars ‘Middletown’, ‘Overleese’, ‘PA-Golden’, ‘Sunflower’, ‘Wells’, ‘Wilson’, ‘NC-1’ and the advanced selection 2-54 were evaluated for flower number per tree, density, peak, and duration, as well as trunk cross-sectional area, fruit set, days between flower and harvest peak, duration of harvest, fruit weight, yield, and number of fruit per cluster. An extended flowering period or larger flower number did not improve fruit set. Mature trees of commercially available pawpaw cultivars did not display concentrated bloom periods or ripening periods; however, ‘Wells’ and ‘Middletown’ did have late flowering peaks (maximum number of flowers) that could allow a partial crop if early spring frosts destroy flowers around or at the flowering peak. However, the average fruit size of ‘Wells’ and ‘Middletown’ is small and undesirable for commercial production.
The North American pawpaw [Asimina triloba (L.) Dunal] is in the initial stages of commercial production across the United States (6). The pawpaw fruit has both fresh market and processing appeal, with an intense flavor that resembles a combination of banana, mango, and pineapple (1). Little information has been published concerning flowering, bloom period or fruit set in commercially available pawpaw cultivars. This information would be valuable to growers for selection of cultivars for regional suitability and microclimates.
Pawpaw flowers are strongly protogynous and are likely self-incompatible (8), although some cultivars, such as ‘Sunflower’, may be self-fruitful. Pollination is by flies (Diptera) and beetles (Nitidulidae), and possibly other nocturnal insects (2, 3). Each fruit cluster develops from an individual flower. Seedlings normally flower when trees reach a height of about 1.8 m; cropping is achieved at five to eight years after planting. Grafted trees usually set fruit five to six years after planting, although some cultivars such as ‘PA-Golden (#1)’ may crop in the fourth year (7). The bloom period for a pawpaw tree usually occurs over 3 to 4 weeks; however, bloom periods for specific cultivars have not been evaluated. Late flowering or extended bloom period could be useful traits to growers for avoiding crop losses due to late spring frosts and microclimates that are frost prone. Harvest from trees is labor-intensive, occurring over several weeks, and has been thought to reflect an extended bloom period for a particular cultivar.
In the wild, pawpaw trees are usually
===== PAGE 2 =====
found in the understory of hardwood forests and low fruit set has been reported in wild patches (4, 8). Low light levels in the understory may limit flower bud formation during the previous summer. If flowers are formed and successfully pollinated, low light levels may also reduce photosynthate partitioning to fruit and reduce fruit set. Pawpaws in the wild often produce many root suckers, forming large clonal patches, thus leading to poor fruit set within a patch due to flower self-incompatibility. Pollinator limitation may also lead to low fruit set in wild patches (8). Because the pawpaw flowers are protogynous (8), lack of pollen availability from other pawpaw genotypes may also limit pollination. Pollinizer relationships among pawpaw cultivars have not been examined. Fruit set can be enhanced by hand cross-pollination (5).
We tested a number of hypotheses during this study concerning pawpaw fruit production and flowering, including: 1) an extended flowering period on the same tree will lead to greater fruit set, 2) a larger number of flowers on a tree will result in a greater number of fruit on that tree, and 3) fruit set is low in pawpaw. The objectives of this study were to determine if mature trees of commercially available pawpaw cultivars display: 1) concentrated bloom periods, 2) late flowering characteristics, 3) high fruit set, and 4) concentrated fruit ripening characteristics that were related to concentrated bloom periods.
## Materials and Methods
In 2004, 2005, and 2006, flowering and fruit set characteristics were evaluated for the commercially available cultivars ‘Middletown’, ‘Overleese’, ‘PA-Golden’, ‘Sunflower’, ‘Wells’, and ‘Wilson’ (Table 1). In 2005 and 2006 the cultivar ‘NC-1’ and the Paw-Paw Foundation (PPF) advanced selection 2-54 were also evaluated. The trees used in this study were part of a pawpaw regional variety trial (PRVT) that was planted in late March, 1998 in a Lowell silt loam soil (pH 6.9) at the KSU Research Farm in Frankfort, KY (7). The PRVT planting consists of about 300 trees, with five to eight replications (blocks) of 28 grafted scion varieties per block in a randomized complete block design (10 commercially available cultivars and 18 clones selected in the PPF orchards at the University of Maryland Experiment Stations at Queenstown, Md. and Keedysville, Md.). Seedlings from native Kentucky trees serve as border row trees. Trees were fertilized with Peters 20-20-20 (20 N-8.7 P-16.6 K) water-soluble fertilizer four times each year, in May, June, July, and August, for a total of 12.1 kg N •ha-1 (10.8 lbs N/ac). Supplemental irrigation was provided as needed via drip irrigation.
Flowers were counted on individual trees for the selected varieties three times a week during April and May in 2004, 2005, and 2006. Fruit clusters were counted in June and late July. Fruit were harvested from individual trees three times a week as they ripened during the harvest season of mid-August to late September. Trunk diameters were measured at 30 cm from ground level in March of each year prior to bloom. The graft union
Table 1. Genetic background of pawpaw clones examined in this study.
| Clone | Genetic background |
| Middletown | Wild seedling from Middletown, Ohio |
| NC-1 | ‘Davis’ × ‘Overleese’ |
| Overleese | Open-pollinated seedling from Rushville, Ind. |
| PA-Golden | Second-generation seedling from G.A. Zimmerman collection selected by John Gordon in Amherst, NY. |
| Sunflower | Wild seedling from Chanute, Kans. |
| Wells | Open-pollinated seedling from Salem, Ind. |
| Wilson | Wild seedling from Cumberland, Ky. |
| 2-54 | Open-pollinated seedling from George A. Zimmerman of Linglestown, Pa. |
===== PAGE 3 =====
was at a height of 15 cm from the soil line. Values for trunk diameters were converted to trunk cross-sectional area [TCA (cm2)]. The number of fruit on each tree was counted in late July each year following the normal fruit drop period. Growing degree days (GDDs) were calculated using a base temperature of 10circC (University of Kentucky Agricultural Weather Center calculator, http://wwwagwx.ca.uky.edu/calculators.html). Temperature data were obtained from the University of Kentucky Agricultural Weather Center's monthly climate summary (http://wwwagwx.ca.uky.edu/cgi-public/climsum2.ehtml), from temperatures recorded at Lexington.
Data on trunk cross-sectional area, flower number per tree, flower density (number of flowers per tree/TCA), flowering peak (date at which the greatest number of flowers were fully open; when flowers are maroon in color), fruit set, days between flower and harvest peak, duration of flowering, duration of harvest, and yield by cultivar or advanced selection were subjected to GLM analysis of variance, LSD mean separation, and regression analysis using the statistical program Costat (CoHort Software, Monterey, Calif.).
## Results
In 2004, 2005, and 2006, TCA differed significantly among cultivars, reflecting variation in vigor among the cultivars (Table 2). Flower number per tree varied among cultivars in each of the three years. There was variation among cultivars in flower density, with ‘Middletown’ tending to produce the most flowers per unit TCA in 2005 and 2006. Significant cultivar differences in fruit set were observed each year. For example, ‘Middletown’ displayed the highest flower density in 2005 and 2006, but the highest fruit set was not observed in this cultivar in either year (Table 2). Yield varied by cultivar in each year. No crop occurred in 2003 due to an April frost event. This likely resulted in a heavy crop for each cultivar in 2004, lower yields in 2005, and high yields again in 2006. ‘PA-Golden’ was the highest yielding cultivar in each year of the study.
Peak flowering date, bloom duration period, days between flower peak and harvest peak varied among cultivars during each year of the study (Table 3). Late-flowering cultivars were ‘Middletown’ and ‘Wells’, which were about 8 to 10 days later than the early flowering cultivars ‘NC-1’ and ‘Overleese’, and the advanced selection 2-54 (Table 3). Bloom duration varied among cultivars in all years ranging from 23-36 days (Table 3 and Fig. 1). Significant differences in harvest duration were only found in 2006, and harvest duration varied from 15 to 31 days across all years (Table 3). ‘Middletown’, ‘Wilson’, ‘Overleese’, ‘NC-1’ and ‘Sunflower’, and the advanced selection 2-54 had the longest number of days between flowering peak and harvest peak. ‘PA-Golden’ (New York), ‘Taytwo’ (Michigan), and ‘Taylor’ (Michigan) were selected in the most northern regions and had the fewest days between flowering peak and harvest peak. Fruit weight varied (65 to 189 g) with cultivar and year (Table 3). ‘NC-1’, ‘Overleese’ and ‘Sunflower’ had the largest fruit, averaging over 150 g, and ‘Middletown’, ‘Wilson’, ‘PA-Golden’, ‘Wells’ and the advanced selection 2-54 had fruit weights that averaged less than 150 g.
Correlation analyses performed on all cultivars combined (Table 4) showed a positive linear relationship between TCA and number of clusters, TCA and number of flowers, flower peak and harvest peak, number of flowers and number of fruit clusters, number of flowers and yield, and a negative linear relationship between number of flowers and fruit set in 2004; TCA and number of clusters, TCA and number of flowers, fruit set and yield, and number of flowers and yield in 2005; and TCA and number of flowers, number of flowers and number of clusters, fruit set and yield, and number of flowers and yield in 2006. Using regression analysis, a positive binomial relationship was also found between flower peak and harvest peak in 2006. When combining all data and using regression analysis on the averages of all years, there was a positive linear relationship between TCA and number of clusters, TCA
===== PAGE 4 =====
Table 2. Trunk diameter, flower number, flower density, and fruit set for 7-9 year old trees of eight pawpaw selections in 2004, 2005 and 2006.
| Clone | Trunk cross-sectional area (cm2) | Flower density (flowers/cm2 TCA) | Fruit set (%) | Yield (kg) per tree |
| 2004 | 2005 | 2006 | Number of flowers per tree | 2004 | 2005 | 2006 | 2004 | 2005 | 2006 | 2004 | 2005 | 2006 |
| 2-54 | 28.7 bcy | 51.2 ab | 54.6 ab | NA | 938 a | 597 bc | NA | 20.0 ab | 11.0 bc | NA | 0.7 c | 6.5 b | 10.8 abc | 3.1 c | 12.8 abc |
| Middletown | 24.0 c | 31.6 cd | 36.6 cd | 480 bc | 592 d | 609 bc | 18.8 | 20.5 a | 17.6 a | 9.2 b | 1.9 b | 6.6 b | 7.0 c | 3.1 c | 6.2 de |
| NC-1 | 36.3 ab | 53.3 ab | 60.5 ab | NA | 918 a | 690 ab | NA | 18.4 ab | 11.6 bc | NA | 1.0 bc | 3.0 c | 7.9 c | 3.7 bc | 11.3 bcd |
| Overleese | 32.8 b | 44.4 abc | 52.0 b | 676 ab | 662 bcd | 460 cd | 20.6 | 15.1 bc | 9.0 c | 4.4 c | 1.5 bc | 4.6 bc | 9.8 bc | 4.9 abc | 10.1 cde |
| PA-Golden | 41.1 a | 51.3 ab | 59.6 ab | 758 a | 781 abc | 817 a | 18.6 | 15.2 bc | 13.2 b | 8.8 b | 3.1 a | 9.4 a | 13.8 ab | 7.7 a | 17.2 a |
| Sunflower | 33.0 b | 43.1 bc | 48.3 bc | 545 bc | 828 ab | 663 ab | 16.6 | 19.6 ab | 13.8 b | 6.8 bc | 1.3 bc | 7.0 b | 14.0 ab | 3.8 bc | 16.7 ab |
| Wells | 23.9 c | 28.5 d | 32.8 d | 349 c | 660 cd | 417 d | 13.8 | 19.1 ab | 10.7 bc | 13.8 a | 3.0 a | 6.6 b | 9.5 bc | 4.9 abc | 5.0 e |
| Wilson | 42.9 a | 58.9 a | 68.1 a | 806 a | 699 bcd | 831 a | 19.0 | 11.9 c | 11.8 bc | 7.7 bc | 3.4 a | 5.2 bc | 15.7 a | 6.0 ab | 13.0 abc |
| Main effectz | *** | *** | *** | ** | ** | *** | NS | * | *** | *** | *** | *** | ** | * | *** |
| Block effectz | NS | NS | NS | * | NS | NS | NS | NS | NS | NS | NS | ** | NS | NS | NS |
| LSD | 8.0 | 14.8 | 15.3 | 225 | 219 | 187 | 8.1 | 5.9 | 3.8 | 3.9 | 1.3 | 2.6 | 5.3 | 3.1 | 6.3 |
| Mean | 32.7 | 44.9 | 51.0 | 601 | 743 | 635 | 17.8 | 17.5 | 12.5 | 8.5 | 2.1 | 6.1 | 11.0 | 4.6 | 11.5 |
2 NS = not significant,* significant at P=0.05,** at P=0.01, and*** at P=0.001.y Any two means within a column not followed by the same letter are significantly different at P ≤ 0.05 using Fisher's protected LSD mean separation.
===== PAGE 5 =====
Table 3. Peak flowering date, bloom duration and harvest duration periods, and days between flowering peak and harvest peak for mature trees of eight pawpaw selections for 2004, 2005 and 2006.
| Clone | Peak flowering date (month/day) | Bloom duration period (days) | Harvest duration period (days) | Days between flower peak and harvest peak | Average fruit weight (g) |
| 2004 | 2005 | 2006 | 2004 | 2005 | 2006 | 2004 | 2005 | 2006 | 2004 | 2005 | 2006 | 2004 | 2005 | 2006 |
| 2-54 | NAz | 4/17 d | 4/20 d | NA | 29 ab | 25 f | NA | 19 | 23 ab | NA | 139 ab | 145 a | 134 c | 117 c | 112 bc |
| Middletown | 4/27 ay | 4/27 a | 4/28 a | 23 c | 25 c | 32 bc | 31 | 18 | 19 bc | 141 ab | 135 bc | 140 abc | 76 e | 85 d | 65 d |
| NC-1 | NA | 4/17 d | 4/21 cd | NA | 30 a | 28 e | NA | 18 | 20 ab | NA | 138 abc | 144 a | 176 a | 189 a | 171 a |
| Overleese | 4/21 b | 4/20 cd | 4/21 cd | 28 a | 29 ab | 31 cd | 23 | 18 | 19 abc | 130 c | 136 abc | 142 abc | 145 bc | 180 ab | 185 a |
| PA-Golden | 4/22 b | 4/24 ab | 4/23 bc | 28 a | 28 abc | 31 cd | 24 | 27 | 23 ab | 123 d | 127 d | 136 c | 112 d | 116 cd | 103 c |
| Sunflower | 4/22 b | 4/19 cd | 4/24 b | 27 ab | 27 bc | 29 de | 25 | 18 | 25 a | 136 abc | 140 a | 143 ab | 169 ab | 156 b | 140 b |
| Wells | 4/28 a | 4/26 ab | 4/30 a | 26 b | 29 ab | 34 ab | 26 | 23 | 15 c | 134 bc | 133 bc | 138 bc | 111 d | 105 cd | 90 cd |
| Wilson | 4/23 b | 4/22 bc | 4/23 bc | 27 ab | 28 ab | 36 a | 28 | 21 | 25 a | 143 a | 132 c | 140 abc | 98 de | 84 d | 85 cd |
| Main effectz | *** | *** | *** | *** | * | *** | NS | NS | ** | *** | *** | * | *** | *** | *** |
| Block effectz | NS | NS | NS | * | NS | ** | * | NS | NS | NS | NS | NS | NS | NS | NS |
| LSD | 4 | 5 | 3 | 3 | 4 | 3 | 9 | 10 | 6 | 9 | 6 | 6 | 31 | 33 | 33 |
| Mean | 4/24 | 4/22 | 4/24 | 26 | 28 | 31 | 26 | 20 | 21 | 134 | 135 | 141 | 128 | 130 | 118 |
2 NA = not available, NS = not significant, * significant at P=0.05, ** at P=0.01, and * * * at P=0.001.y Any two means within a column not followed by the same letter are significantly different at P ≤ 0.05 using Fisher's protected LSD mean separation
===== PAGE 6 =====
Table 4. Correlation coefficients for trunk cross-sectional area, flowering characteristics, and harvest data for mature trees of eight pawpaw selections in 2004, 2005, and 2006.
| Independent variable | Dependent variable | 2004 | 2005 | 2006 | Average 2004-2006 |
| Significance | r | Significance | r | Significance | r | Significance | r |
| TCAz | Fruit set | 0.63 NS y | | 0.23 NS | | 0.34 NS | | 0.88 NS | |
| TCA | Number of clusters | 0.0000*** | 0.60 | 0.02* | 0.32 | 0.18 NS | | 0.0003*** | 0.52 |
| TCA | Number of flowers | 0.0000*** | 0.72 | 0.0000*** | 0.58 | 0.0000*** | 0.59 | 0.0000*** | 0.76 |
| Flower duration | Harvest duration | 0.49 NS | | 0.46 NS | | 0.82 NS | | 0.62 NS | |
| Flower duration | Fruit set | 0.50 NS | | 0.31 NS | | 0.61 NS | | 0.64 NS | |
| Flower peak | Harvest peak | 0.003** | 0.44 | 0.06 NS | | 0.09 NS | | 0.001** | 0.41 |
| Number of flowers | Fruit set | 0.002** | -0.49 | 0.75 NS | | 0.74 NS | | 0.31 NS | |
| Number of flowers | Number of clusters | 0.0002*** | 0.55 | 0.08 NS | | 0.0000*** | 0.51 | 0.0000*** | 0.51 |
| Fruit set | Yield | 0.23 NS | | 0.0000*** | 0.67 | 0.0000*** | 0.51 | 0.005** | 0.36 |
| Number of flowers | Yield | 0.004** | 0.44 | 0.01* | 0.35 | 0.0000*** | 0.56 | 0.0000*** | 0.69 |
| Fruit weight | Yield | 0.90 NS | | 0.65 NS | | 0.09 NS | | 0.25 NS | |
2 TCA: trunk cross-sectional area in cm2y NS = not significant, * significant at P=0.05, ** at P=0.01, and * * * at P=0.001. r = the sample correlation coefficient and is a measure of extent to which samples of two variables are linearly related.
===== PAGE 7 =====
and number of flowers, flowering peak and harvest peak, number of flowers and number of clusters, fruit set and yield, and number of flowers and yield. Due to year-to-year inconsistencies and variation among cultivars, no definitive conclusions could be reached concerning correlations between flower and fruiting characteristics in individual pawpaw cultivars.
The average temperature in April was similar in 2004 and 2005, and was warmer in April 2006 (Table 5). May 2005 and 2006 temperatures were similar, with May 2004 being warmer than the following years. GDDs from January 1 to first flower and from January 1 to flower peak, were higher each year of
Table 5. Monthly temperature summary during pawpaw flowering for 2004, 2005 and 2006.
| Month/year | Average monthly temperature (°C) |
| April 2004 | 13 |
| May 2004 | 21 |
| April 2005 | 13 |
| May 2005 | 17 |
| April 2006 | 15 |
| May 2006 | 17 |
the study (Table 6). GDDs from first flower to peak flower were similar in 2004 and 2006, and were lower in 2005. Flowering peak and duration were not correlated with GDD; in 2006, which had the largest GDDs before flowering and by flowering peak, did not display an earlier flowering peak or shorter duration for most cultivars. The GDDs from
Table 6. Growing degree days (GDDs) during pawpaw flowering across all cultivars calculated using a base temperature of 10circC for 2004, 2005 and 2006.
| Growing degree day accumulation | 2004 | 2005 | 2006 |
| January 1 to first flower | 174 | 216 | 244 |
| January 1 to flower peak | 299 | 308 | 366 |
| first flower to peak flower | 125 | 92 | 122 |
| flower peak to harvest peak | 2755 | 2932 | 2954 |
flower peak to harvest peak were fewer for 2004, than for 2005 and 2006, indicating that these later years had warmer temperatures than 2004. However, the number of days between flowering peak and harvest peak are similar for the tested cultivars.
## Discussion
This is the first report of combined flowering and ripening characteristics of commercially available pawpaw cultivars. There was significant variation in year to year flowering and ripening characteristics among the cultivars examined. Mature trees of currently available commercial pawpaw cultivars did not display concentrated bloom periods or ripening periods; however, ‘Wells’ and ‘Middletown’ did have late flowering peaks that could allow a partial crop if early spring frosts destroy flowers around or at the flowering peak. However, the average fruit size of ‘Wells’ and ‘Middletown’ is small (about 100 g; Table 3) and undesirable for commercial production. Based on the data collected in this study, we reject the hypotheses that 1) an extended flowering period on the same tree will lead to greater fruit set and we do not reject hypothesis 2) that a larger number of flowers on a tree will result in a greater number of fruit on that tree. Pollinator limitation has often been suggested as an explanation for low fruit set (<0.5%) in wild patches (8). Fruit set here was greater (5.6%) than that reported in wild pawpaw patches (around 0.5%) (4, 8). Pollinator limitation may reduce fruit set in the wild compared to that in the orchard. Strategies to attract flies to flowering pawpaw orchards could increase fruit set further. Alternatively, pawpaw trees may be able to support a limited number of fruit per tree both in the wild and in orchards and pollinators may not be the limiting factor. Additionally, year to year fruit set may have been influenced by a 2003 frost event that destroyed the pawpaw crop. This likely has resulted in biennial bearing which led to high fruit set in 2004, low fruit set in 2005, and once again high fruit set in 2006. Fruit ripening duration was not related to bloom.
===== PAGE 8 =====
clone
Figure 1. Flowering duration and flowering peak for eight pawpaw selections in 2004, 2005, and 2006. FP = flowering peak
duration. Identification of unique wild germ-plasm and selective breeding will be required to improve many of the flowering and fruiting characteristics of pawpaw.
## Conclusions
Annual variation in flowering and ripening characteristics among the cultivars examined was great. An extended flowering period or larger flower number did not improve fruit set. Mature trees of currently available commercially pappaw cultivars did not display concentrated bloom periods or ripening periods; however, ‘Wells’ and ‘Middletown’ did have late flowering peaks that could allow a partial crop if early spring frosts destroy flowers around or at the flowering peak. Unfortunately, the average fruit size of ‘Wells’ and ‘Middletown’ is small and undesirable for commercial production.
## Acknowledgments
This research was supported by U.S. Dept. of Agriculture, Cooperative State Research, Education, and Extension Service Agreement No. KYX-10-05-40P with Kentucky State University. The technical assistance of S. Jones, E. Reed, and J. Lowe is also gratefully acknowledged by the authors.
===== PAGE 9 =====
## Literature Cited
1. Duffrin M.W. and K.W. Pomper. 2006. Development of flavor descriptors for pawpaw fruit puree: A step toward the establishment of a native tree fruit industry. Family and Consumer Sci. Research J. 35:118-130.
2. Faegri, K. and L. van der Pijl. 1971. The principles of pollination ecology. Pergammon Press, New York.
3. Kral, R. 1960. A revision of Asimina and Deeringothamnus (Annonaceae). Brittonia 12:233–278.
4. Lagrange, R.L. and E.J. Tramer. 1985. Geographic variation in size and reproductive success in the paw paw (Asimina triloba). Ohio J. Sci. 85:40-45.
5. Peterson, N. 1997. How to hand-pollinate paw-paws. Fruit Gardener 29:10-11.
6. Pomper, K.W. and D.R. Layne. 2005. The North American pawpaw: botany and horticulture. Hort. Reviews 31:351-384.
7. Pomper, K.W., D.R. Layne, R.N. Peterson and D. Wolfe. 2003. The pawpaw regional variety trial: background and early data. HortTechnology 13: 412-417.
8. Willson, M.F. and D.W. Schemske. 1980. Pollinator limitation, fruit production, and floral display in pawpaw (Asimina triloba). Bull. Torrey Bot. Club 107:401–408.
## When you pick ACN You pick the best!
Adams County Nursery, Inc. • Aspers, PA • (800) 377-3106 (717) 677-4124 fax • www.acnursery.com • Email acn@acnursery.com
===== PAGE 10 =====
Table 1. Average fruit weight of 'Gala' apple on 10 dwarfing rootstocks at five locations for three seasons. Values are least-squares means, adjusted for missing observations and crop density (CD).z
| Stock | Illinois | Maine | Michigan | New York - Geneva | Virginia |
| 2000 | 2001 | 2002 | 2001 | 2002 | 2003 | 2000 | 2001 | 2002 | 2001 | 2002 | 2003 | 2000 | 2001 | 2002 |
| M.9 EMLA | 185 b | 178 ab | 145 | 142 a | 121 ab | 144 a | 178 ab | 162 b | 149 b | 142 a | 121 ab | 144 a | 178 ab | 162 b | 149 b |
| M.26 EMLA | 187 b | 175 b | 147 | 136 ab | 121 ab | 136 ab | 176 bc | 159 b | 152 ab | 136 ab | 121 ab | 136 ab | 176 bc | 159 b | 152 ab |
| M.9RN29 | 192 ab | 177 ab | 147 | 135 ab | 118 abc | 144 a | 181 ab | 159 b | 149 b | 135 ab | 118 abc | 144 a | 181 ab | 159 b | 149 b |
| M.9 Pajam1 | 188 b | 180 ab | 143 | 138 ab | 118 abc | 141 a | 182 a | 164 b | 148 b | 138 ab | 118 abc | 141 a | 182 a | 164 b | 148 b |
| M.9 Pajam2 | 198 a | 186 ab | 145 | 128 bc | 116 bc | 134 ab | 181 ab | 163 b | 149 b | 128 bc | 116 bc | 134 ab | 181 ab | 163 b | 149 b |
| B.9 | 198 a | 172 b | 164 | 142 a | 124 ab | 144 a | 171 cd | 162 b | 150 b | 142 a | 124 ab | 144 a | 171 cd | 162 b | 150 b |
| O.3 | 187 b | 171 bc | 141 | 142 a | 126 a | 125 b | 181 ab | 158 b | 156 a | 142 a | 126 a | 125 b | 181 ab | 158 b | 156 a |
| V.1 | 183 b | 173 b | 157 | 135 a | 117 bc | 136 ab | 174 bc | 153 b | 154 ab | 135 a | 117 bc | 136 ab | 174 bc | 153 b | 154 ab |
| Mark | 185 b | 156 c | 137 | 125 c | 108 c | 125 b | 161 d | 148 c | 136 c | 125 c | 108 c | 125 b | 161 d | 148 c | 136 c |
| M.9T337 | 193 ab | 187 a | 146 | 140 a | 122 ab | 135 ab | 179ab | 173 a | 161 a | 140 a | 122 ab | 135 ab | 179ab | 173 a | 161 a |
| P-value from ANCOVA |
| Stock | 0.001 | 0.001 | 0.491 | 0.001 | 0.001 | 0.055 | 0.004 | 0.001 | 0.011 | 0.007 | 0.024 | 0.047 | 0.001 | 0.001 | 0.018 |
| CD | 0.007 | 0.003 | 0.001 | 0.003 | 0.019 | 0.021 | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 |
| CD*Stock | 0.544 | 0.281 | 0.347 | 0.136 | 0.261 | 0.313 | 0.519 | 0.431 | 0.304 | 0.329 | 0.225 | 0.873 | 0.938 | 0.223 | 0.169 |
2 LSmeans within location and year were compared with PDIFF, P=0.05.
===== PAGE 11 =====
Crop Density
Figure 1. Scatter plots for average fruit weight (g) on crop density (no. of fruit per cm² of trunk cross-sectional area) for two rootstocks at three locations. Scatter plots show the relationship is sometimes poor.
===== PAGE 12 =====
Table 2. Average fruit weight of ‘Gala’ apple on 10 dwarfing rootstocks at British Columbia for three seasons. Values are least-squares means, adjusted for missing observations and crop density (CD). In 2003 there was a significant rootstock by CD interaction, so least squares means were compared at three CDs (3.0, 5.6, and 7.0 fruit·cm⁻² TCA).
| 2001 | 2002 | | 2003 | |
| Stock | | | CD = 3.0 | CD = 5.6 | CD = 7.0 |
| M.9 EMLA | 190 a | 177 b | 142 b | 151 ab | 156 a |
| M.26 EMLA | 194 a | 180 ab | 169 ab | 143 b | 129 b |
| M.9RN29 | 202 a | 197 a | 156 ab | 159 a | 161 a |
| M.9 Pajam1 | 197 a | 187 ab | 180 a | 161 a | 151 a |
| M.9 Pajam2 | 195 a | 189 ab | 166 ab | 157 a | 153 a |
| B.9 | 194 a | 185 ab | 153 b | 154 ab | 155 a |
| O.3 | 196 a | 181 ab | 169 ab | 162 a | 158 a |
| V.1 | 196 a | 187 ab | 165 ab | 152 a | 144 ab |
| Mark | 161 b | 137 c | 122 c | 129 b | 134 b |
| M.9T337 | 203 a | 197 a | 165 ab | 161 a | 159 a |
2 LSmeans within columns were compared with PDIFF, P=0.05.
designs, scion cultivars, rootstocks and statistical methods may vary. This is the second trial with ‘Gala’ where trees on B.9 produced relatively large fruit. However, these results also contradict those of the previous trial (14), where trees on Mark produced intermediate sized fruit and trees on M.26 EMLA consistently produced small fruit. The positive relationships between FW and CD, as indicated by the positive slopes, were unexpected because there are many reports of a negative relationship between these two variables (1, 9, 12, 20). There are several possible explanations for these unexpected results. 1) Some cooperators may have thinned trees too late in the season to substantially improve fruit size. 2) The number of replications may have been too low to obtain the true relationship because unusual observations can be highly influential when there are few replicates. 3) The unexpected results most likely resulted from the narrow range of crop loads. In most thinning experiments, treatments are selected that will produce a wide range of crop loads. However in rootstock studies, cooperators use various fruit thinning techniques to obtain crop loads that would encourage good fruit size and adequate return bloom. For these reasons, typical rootstock trials and orchard observations may not be appropriate for evaluating the influence of rootstocks on fruit size. Perhaps the influence of crop load on FW is relatively minor and inconsistent when trees are thinned adequately. Ideally, experiments should be designed specifically to evaluate the influence of rootstock and cultural practices on fruit size. Such experiments would involve wide ranges in CD and overlapping CDs for all rootstocks or treatments. NC-140 cooperators are currently conducting a study to evaluate the effects of rootstock on FW over a wide range of CDs and results from that study may help explain previous inconsistent results.
## Literature Cited
1. Auchter, E.C. and H.B. Knapp. 1937. Orchard and small fruit culture. John Wiley and Sons, New York, NY.
2. Autio, W.R. 1991. Rootstocks affect ripening and other qualities of ‘Delicious’ apple. J. Amer. Soc. Hort. Sci. 116:378-382.
3. Barden, J.A. and R.P. Marini. 1997. Growth and fruiting of a spur-type and standard strain of ‘Golden Delicious’ on several rootstocks over eighteen years. Fruit Var. J. 51:165-175.
4. Barden, J.A. and R.P. Marini. 1999. Rootstock effects on growth and fruiting of a spur-type and a standard strain of 'Delicious' over eighteen years. Fruit Var. J. 53:115-125.
===== PAGE 13 =====
Table 3. Slopes of average fruit weight on crop density (CD) for ‘Gala’ apple on 10 rootstocks at six locations for three seasons. Slopes indicate the change in average fruit weight (g) for an increase of one fruit·cm⁻² in CD. Slopes were estimated from the solution vector provided by SAS’s MIXED procedure.
| Stock | IL | ME | MI | NY-G | VA | BC |
| 2000 | 2001 | 2002 | 2000 | 2002 | 2003 | 2001 | 2002 | 2003 | 2001 | 2002 | 2003 | 2000 | 2001 | 2002 | 2001 | 2002 | 2003 |
| M.9 EMLA | -3.6 | -3.5 | -3.8 | 1.4 | -4.6 | -1.3 | -4.4 | -4.5 | -4.5 | -4.0 | -3.6 | -3.3 | -2.9 | -2.0 | -2.9 | 1.9 | 0.1 | 3.5 |
| M.26 EMLA | -5.4 | -9.0 | -3.4 | 2.7 | -6.0 | 0.5 | -4.1 | -2.1 | -3.5 | -2.0 | -3.2 | -6.6 | -3.9 | -2.9 | -13.9 | -4.6 | -4.9 | -10.6 |
| M.9RN29 | -6.1 | 3.2 | -9.7 | 9.1 | -6.2 | 5.3 | -2.2 | -4.7 | -6.6 | -3.5 | -3.4 | -3.1 | -2.4 | -4.2 | -1.9 | -4.6 | -9.3 | 1.6 |
| M.9 Pajam1 | -4.0 | -3.4 | -8.3 | -0.2 | -1.2 | -4.2 | -1.4 | -2.4 | -0.5 | -6.5 | -2.4 | -6.2 | -3.8 | 3.4 | 3.8 | -0.4 | -3.8 | -6.8 |
| M.9 Pajam2 | -7.1 | -2.5 | -7.6 | 7.6 | 1.8 | -3.7 | -2.4 | -3.8 | -8.4 | -4.9 | -7.6 | -5.3 | -2.8 | -3.8 | -2.6 | -1.9 | -3.6 | -3.4 |
| B.9 | -11.2 | -2.9 | -10.6 | 8.6 | -2.7 | -6.6 | -1.6 | -1.2 | -8.4 | -0.9 | -2.2 | -6.2 | -7.7 | -4.1 | -8.1 | -2.7 | -5.5 | 0.7 |
| O.3 | -6.8 | -6.4 | -4.5 | 1.3 | -6.3 | -5.7 | -2.6 | -1.4 | -3.4 | -1.5 | -1.8 | -5.4 | -4.4 | -5.2 | -0.6 | -3.9 | -7.4 | -3.2 |
| V.1 | -3.6 | -1.9 | -2.9 | 0.1 | -8.4 | -0.3 | -4.3 | -2.7 | -10.6 | -6.1 | -8.3 | -8.5 | -1.2 | -4.0 | -3.6 | 2.6 | -5.9 | -5.3 |
| Mark | -7.6 | -5.6 | -6.3 | -4.4 | -8.6 | -1.6 | -2.2 | -0.9 | -5.1 | -1.9 | -5.0 | -8.5 | -9.7 | -4.0 | -21.7 | -0.3 | 12.1 | 2.7 |
| M.9T337 | -1.8 | -3.6 | -8.6 | 6.4 | 6.1 | -12.3 | 0.1 | -1.4 | -7.7 | -2.3 | -3.4 | -3.6 | -4.1 | -2.6 | -3.9 | -6.3 | -0.8 | -0.8 |
===== PAGE 14 =====
5. Barden, J.A. and R.P. Marini. 2001. Yield, fruit size, red color, and estimated crop value in the NC-140 1990 cultivar/rootstock trial in Virginia. J. Amer. Pom. Soc. 55:150-158.
6. Barrett, B.H., J.A. Barden, J. Cline, R.L. Granger, M.M. Kushad, R.P. Marini, M. Parker, R.L. Perry, T. Robinson, C.R. Unrath, and M.A. Dilley. 1997. Performance of ‘Gala’ at year 5 with eight apple root stocks in an 8-location North American NC-140 trial. Acta Hort. 451:129-135.
7. Barrett, B.H., B.S. Konishi, and M.A. Dilley. 1995. Performance of three apple cultivars with 23 dwarfing rootstocks during 8 seasons in Washington. Fruit Var. J. 49:158-170.
8. Barrett, B.H., B.S. Konishi and M.A. Dilley. 1996. Performance of three apple cultivars with 18 vigorous rootstocks during nine seasons in Washington. Fruit Var. J. 50:88-98.
9. Gourley, J.H. and F.S. Howlett. 1947. Modern fruit production. The Macmillan Company, New York, NY.
10. Hollander, M. and D.A. Wolfe. 1973. Nonparametric statistical methods. John Wiley and Sons, New York, NY.
11. Jackson, J.E. and A.B. Blasco. 1975. Effects of rootstock and crop load on fruit size and quality of Cox's Orange Pippin and Worcester Pearmain. Rpt. E. Malling Res. Sta. 1974, p. 45.
12. Lakso, A.N., T. L. Robinson, M.C. Goffinet and M.D. White. 2001. Apple fruit growth responses to varying thinning methods and timing. Acta Hort. 557:407-412.
13. Littell, R.C., G.A. Milliken, W.W. Stroup and R. Wolfinger. 1996. SAS System for mixed models. SAS Institute Inc., Cary, N.C.
14. Marini, R.P, J.A. Barden, J.Cline, R.L. Perry and T. Robinson. 2002. Effect of apple rootstocks on average fruit weight at four locations after adjusting for crop load. J. Amer. Soc. Hort. Sci. 127:749-753.
15. Marini, R.P., J.L. Anderson, W.R. Autio, B.H. Barritt, J. Cline, W.P. Cowgill, Jr., R.C. Crassweller, R.M. Garner, A. Gauss, R.Godin, G.M. Greene, C. Hampson, P. Hirst, M.M. Kushad, J.M. Masaabni, E. Mielke, R. Moran, C.A. Mullins, M. Parker, R.L. Perry, J.P. Privé, G.L. Reighard, T. Robinson, C.R. Rom, T. Roper, J.R. Schupp, E. Stover and R. Unrath. 2006. Performance of 'Gala' apple trees on 18 dwarfing rootstocks: ten-year summary of the 1994 NC-140 rootstock trial. J. Amer. Pomol. Soc. 60:69-83.
16. Marini, R.P., B.H. Barritt, J.A. Barden, J.Cline, R.L. Granger, M.M. Kushad, M. Parker, P.L. Perry, T. Robinson, S. Khanizadeh and C.R. Unrath. 2001. Performance of 'Gala' apple on eight dwarf rootstocks: ten-year summary of the 1990 NC-140 rootstock trial. J. Amer. Pomol. Soc. 55:197-204.
17. Milliken, G.A. and D.E. Johnson. 2002. Analysis of messy data. Vol. III: Analysis of covariance. Chapman and Hall/CRC, New York, NY.
18. NC-140. 1996. Performance of the NC-140 cooperative apple rootstock planting: I. Survival, tree size, yield and fruit size. Fruit Var. J. 50:6-11.
19. SAS Institute, Inc. 1990. SAS Procedures Guide, Version 6, 3rd Ed., Cary, NC.
20. Stover, E., M. Fargione, and R. Risio, X. Yang and T. Robinson. 2001. Fruit weight, crop load, and return bloom of 'Empire' apple following thinning with 6-benzyladenine and NAA at several phenological stages. HortScience 36:1077-1081. 
