Grafting watermelon (Citrullus lanatus) is a common practice in many parts of the world and has recently received increased interest in the United States. The present study was designed to evaluate early season growth, yield, and fruit quality of watermelon in response to grafting and in the absence of known disease pressure in a fumigated system. Field experiments were conducted using standard and mini watermelons (cv. Exclamation and Extazy, respectively) grafted onto 20 commercially available cucurbit rootstocks representing four species: giant pumpkin (Cucurbita maxima), summer squash (Cucurbita pepo), bottle gourd (Lagenaria siceraria), and interspecific hybrid squash [ISH (C. maxima × Cucurbita moschata)]. Nongrafted ‘Exclamation’ and ‘Extazy’ were included as controls. To determine early season growth, leaf area was measured at 1, 2, and 3 weeks after transplant (WAT). At 1 WAT, nongrafted ‘Exclamation’ produced the smallest leaf area; however, at 3 WAT, nongrafted ‘Exclamation’ produced the largest leaf area in 2015, and no differences were observed in 2016. Leaf area was very similar among rootstocks in the ‘Extazy’ study, with minimal differences observed. Marketable yield included fruit weighing ≥9 and ≥3 lb for ‘Exclamation’ and ‘Extazy’, respectively. In the ‘Exclamation’ study, highest marketable yields were observed in nongrafted ‘Exclamation’, and ‘Exclamation’ grafted to ‘Pelops’, ‘TZ148’, and ‘Coloso’, and lowest marketable yields were observed when using ‘Marvel’ and ‘Kazako’ rootstocks, which produced 47% and 32% of nongrafted ‘Exclamation’ yield, respectively. In the ‘Extazy’ study, the highest marketable yield was observed in nongrafted ‘Extazy’, and ‘Kazako’ produced the lowest yields (48% of nongrafted ‘Extazy’). Fruit quality was determined by measuring fruit acidity (pH), soluble solids concentration (SSC), lycopene content, and flesh firmness from a sample of two fruit from each plot from the initial two harvests of each year. Across both studies, rootstock had no effect on SSC or lycopene content. As reported in previous studies, flesh firmness was increased as a result of grafting, and nongrafted ‘Exclamation’ and ‘Extazy’ had the lowest flesh firmness among standard and mini watermelons, respectively. The present study evaluated two scions with a selection of 20 cucurbit rootstocks and observed no benefits in early season growth, yield, or phytonutrient content. Only three of 20 rootstocks in each study produced marketable yields similar to the nongrafted treatments, and no grafted treatment produced higher yields than nongrafted ‘Exclamation’ or ‘Extazy’. Because grafted seedlings have an associated increase in cost and do not produce increased yields, grafting in these optimized farming systems and using fumigated soils does not offer an advantage in the absence of soilborne pathogens or other stressors that interfere with watermelon production.
Grafting watermelon is a common practice in many parts of the world, including China, Korea, Japan, Spain, Italy, and Israel, but has yet to be adopted widely in the United States (Kubota et al., 2008; Sakata et al., 2007). Reported benefits of grafting include resistance to diseases caused by soilborne pathogens, abiotic stress tolerance, and enhanced yield and fruit quality (Davis et al., 2008b; Lee and Oda, 2003; Louws et al., 2010). Modern vegetable grafting was first introduced in watermelon production in 1920, when Japanese growers grafted watermelon to squash rootstocks to provide resistance to fusarium wilt (Fusarium oxysporum f. sp. niveum) and other soilborne pests (Lee and Oda, 2003; Tateishi, 1927). In Turkey, higher yields were reported in grafted watermelon compared with nongrafted watermelon, although grafted watermelon exhibited a decrease in fruit quality (Turhan et al., 2012). In Japan and Korea, watermelon production has undergone an intensification on limited arable acreage, and many growers use permanent high tunnel structures or greenhouses for production. Permanent structures restrict the ability for crop rotation, which led to the adoption of grafting to address diseases incited by soilborne pathogens (Kubota et al., 2008; Lee and Oda, 2003; Sakata et al., 2007).
Vegetable grafting became an important pest management option following the removal of methyl bromide as a soil fumigant (Zagheni, 2003). The loss of methyl bromide and the success of grafting in Asian countries have caused an increased interest in vegetable grafting in western regions including Europe, North Africa, the Middle East, and Central and South America (Kubota et al., 2008; Miguel et al., 2004; Moreno et al., 2016). A study in Egypt reported only 67% survival of nongrafted watermelon that were planted in a fusarium wilt-infested field, whereas grafted plants had 83% to 100% survival and a corresponding increase in yield (Mohamed et al., 2012). In Spain and South America, economic analyses support the use of grafting where growers observe high levels of fusarium wilt in fields (Miguel et al., 2004; Moreno et al., 2016).
Despite reported benefits of grafting, U.S. watermelon growers are slow to adopt the practice because of the associated increases in labor costs, seed costs, greenhouse space, and management time required to oversee the grafting process (Kubota et al., 2008). If grafted seedlings are purchased, growers can expect to pay $0.75–$1.00 per seedling compared with $0.28 per nongrafted seedling, so grafting must provide substantial benefits to offset the expense (Taylor et al., 2008). Research in Oklahoma demonstrates that if a grower expects high yields (50,000 kg·ha−1), grafted watermelon fruit would have to sell for $0.22/kg to break even compared with a price of $0.15/kg for nongrafted watermelon fruit (Taylor et al., 2008). Furthermore, researchers in Oklahoma emphasize the importance of the history of the field being considered for watermelon production. If a field is historically free of fusarium wilt and disease is not anticipated to become a problem, a grower would benefit by avoiding the increased cost associated with grafted seedlings (Taylor et al., 2008). In Washington, researchers developed cost estimate analyses that favor the implementation of grafting for growers in the Pacific northwestern United States. The results were based on a series of assumptions, particularly an increase in yield of grafted plants and the effectiveness of grafting as an alternative to fumigation for disease control (Galinato et al., 2016). However, the authors note that growers should review the assumptions to check whether the expenses and yields in the models match those of the grower.
Several studies report no difference or even reduced yields in grafted watermelon as compared with nongrafted watermelon, particularly in the absence of disease pressure (Bertucci et al., 2017; Kokalis-Burelle et al., 2016). Under disease-free conditions in Florida, grafted watermelon yielded nearly 50% less total fruit weight than nongrafted watermelon (Kokalis-Burelle et al., 2016). Yields from seeded watermelon scions grafted to bottle gourd rootstocks were increased and yields were decreased when grafting to ISH rootstocks (Yetisir et al., 2003). This disparity emphasizes the need for further investigation of yield response to grafting watermelon in the southeastern United States.
It is important that grafted plants are tested and evaluated in the environment where they will be grown. Rootstock performance is dependent on edaphic factors and climatic conditions, which will define how each rootstock performs in a given environment (Kumar et al., 2017). Thus, one of the major difficulties of grafting is finding the appropriate rootstock–scion combination. The objectives of this study were to graft both standard and mini watermelons (cultivars Exclamation and Extazy, respectively) on 20 commercially available rootstocks to determine the effect of grafting on early season growth, time to maturity, fruit yield, and fruit quality of standard and mini watermelons in the southeastern United States.
Materials and methods
Field studies were conducted in separate fields in 2015 and 2016 at the Cunningham Research Station (lat. 35.297°N, long. 78.275°W) in Kinston, NC. Soil at this site was Norfolk loamy sand (fine-loamy, kaolinitic, thermic Typic Kandiudults) with pH 5.8 and 6.7, and 1.2% and 1.6% organic matter in 2015 and 2016, respectively.
Scions for the experiment were the two triploid (seedless) watermelon cultivars: Exclamation (Syngenta Seeds, Greensboro, NC), an oblong, standard-sized watermelon and Extazy (Hazera Seeds Ltd., Berurim, Israel), a round mini watermelon. ‘Exclamation’ and ‘Extazy’ were selected based on availability to North Carolina growers and to evaluate the utility of grafting for both standard and mini watermelons. Scions were grafted onto one of 20 commercially available rootstocks representing four species: giant pumpkin, summer squash, bottle gourd, and ISH rootstocks (Table 1). Nongrafted ‘Exclamation’ or ‘Extazy’ were included as control treatments. Rootstocks were selected based on previous literature or personal communication that indicated a potential benefit for watermelon production. An up-to-date list of cucurbit rootstock options for growers is available online (Kleinhenz, 2015).
Comprehensive list of rootstock, scion, and pollinizer cultivars used in replicated field experiments conducted in Kinston, NC, to determine the effect of grafting on early season growth, yield, and fruit quality of grafted and nongrafted standard watermelons and mini watermelons.
Rootstock and scion seeds were sown in 72-cell planting trays (T.O. Plastics, Clearwater, MN) filled with soilless propagation media (CC Tobacco Mix; Carolina Soil Co., Kinston, NC). Twenty-four hours before sowing, planting trays were filled with propagation media, watered thoroughly, brought into a greenhouse adjusted to 84 °F, and placed on heating pads set to 89 °F. Seeds of triploid watermelon were oriented with the radicle oriented upward to facilitate release of the seed coat (Maynard, 1989). After seeds were sown, planting trays were placed on heating pads and covered with clear polyethylene film overnight for insulation. During daylight, polyethylene was removed and seedling trays were watered overhead as needed to keep substrate moist. On seedling emergence, heating pads were removed from beneath trays and polyethylene was no longer used to cover trays. When rootstock cotyledons had fully opened and the apical meristem was first exposed, rootstocks were treated with 25 μL of a 6.25% dilution of fatty alcohol solution (Fair 85; Fair Products, Cary, NC) to prevent regrowth of the rootstock following grafting, according to methods developed by Daley and Hassell (2014). Preliminary germination tests were conducted using 24 seeds of each rootstock and scion to determine the appropriate dates to sow each cultivar to conduct all grafting in a single operation (data not shown). Seeds for each rootstock were sown at staggered timings over a 4-d span to ensure all could be treated with fatty alcohol within 1–2 d of each other. Twenty-four hours before grafting, all plants were taken indoors to slow photosynthesis and reduce metabolic activity. Grafting was conducted using the one-cotyledon method (Hassell et al., 2008).
Field site preparation.
In one operation, beds (6 inches high × 30 inches wide) were formed on 10-ft centers, drip tape was laid (3 inches deep), and beds were covered with black polyethylene mulch on 29 Apr. 2015 and 8 Apr. 2016. For soilborne pest control, fumigant (Pic-Clor 60; TriEst Ag Group, Greenville, NC) was applied during bed formation, delivering chloropicrin and 1,3-dichloropropene at 174 and 114 kg·ha−1 a.i., respectively. Watermelon transplanting occurred on 1 June 2015 and 18 May 2016. Planting holes were punched in polyethylene mulch using a tractor-mounted fixed wheel hole puncher (Berry Hill Irrigation, Buffalo Junction, VA) 24 h before transplanting to allow fumigants, if present, to dissipate. The experimental unit for both watermelon types was a plot with six plants. ‘Exclamation’ seedlings were transplanted in 15-ft plots with each seedling transplanted at 2.5-ft in-row spacing on 10-ft centers and 10-ft alleys to separate plots on the same bed. ‘Extazy’ seedlings were transplanted in 9-ft plots at 1.5-ft in-row spacing on 10-ft centers, with 10-ft alleys. To minimize border effects within plots, competitor seedlings were transplanted on each side of a plot at the same spacing as the triploid plants. In 2015, ‘Little Deuce Coup’ (Syngenta Seeds) and ‘Orange Crisp’ (Harris Moran Seed Co., Rochester, NY) were transplanted on the ends of each plot at 2.5- and 1.5-ft spacing for ‘Exclamation’ and ‘Extazy’ plots, respectively. In 2016, cultivars Palomar and Petite Perfection (Syngenta Seeds) were transplanted as competitor plants for ‘Exclamation’ and ‘Extazy’ plots, respectively. Pollenizer seedlings [SP-6 (‘Super Pollenizer 6’; Syngenta Seeds)] per plot were transplanted within plots between every second triploid plant, to ensure sufficient pollen production and proper fruit set for the triploid watermelons (Dittmar et al., 2010).
Watermelon leaf area.
To determine early season growth characteristics, leaf area was measured from two triploid plants per plot at 1, 2, and 3 WAT. Leaf area was measured by capturing an overhead image of the two center plants of a plot with a digital SLR camera (Rebel EOS T3; Canon USA, Huntington, NY) from a height of 7 ft above the beds. Images were then analyzed using the Easy Leaf Area software (Easlon and Bloom, 2014). Easy Leaf Area software compared green pixels in an image with contrasting red pixels from a scale of known dimensions (3 × 3 cm, in this case) and then calculated the total leaf area from the two values. The final leaf area measurement was recorded at 3 WAT, which was the last opportunity to record individual watermelon plant growth before substantial canopy overlap.
Watermelons were harvested at weekly intervals when ripe, beginning at 65 and 72 d after transplanting in 2015 and 2016, respectively. Ripeness was determined by scouting for a dried tendril where the fruit meets the vine, a well-developed ground spot where the fruit contacts the soil surface, ridges extending along the rind of the melon, and a breakdown of epicuticular wax and hairs on the fruit surface. Ripeness was more difficult to identify with ‘Extazy’, so harvest was initiated 10 d later in 2016 (82 d after transplanting), and fruit were scouted for an additional senescent tendril proximal to the crown from the fruit (Vinson et al., 2010). Fruit of ‘Exclamation’ were harvested four times in each year, whereas fruit of ‘Extazy’ were harvested five times and three times in 2015 and 2016, respectively. Marketable yields and marketable fruit counts were summed across all harvests within each year. Standard and mini watermelons were considered marketable if they weighed ≥9 and ≥3 lb, respectively. Standard watermelons are commonly divided into “bin counts” referring to the number of fruit that would fill one 800-lb watermelon bin. For standard watermelons, watermelon yield and fruit count were categorized in accordance with the National Watermelon Research and Development Group standards: culls (<9 lb), 60-count (≥9–13.5 lb), 45-count (>13.5–17.5 lb), 36-count (>17.5–21.5 lb), and 30-count (>21.5 lb) (Coolong, 2015). No national standard for grading has been reported for mini watermelons, so fruit weighing ≥3 lb were considered marketable, which is similar to previous studies with mini watermelon (Colla et al., 2011; Rouphael et al., 2008).
Watermelon fruit quality.
A two-fruit sample from the initial two harvests in each year was tested for flesh firmness, soluble solids content, pH, and lycopene content. Flesh firmness was measured using a digital force gauge (Force One FDIX; Wagner Instruments, Greenwich CT) mounted to a drill press and fitted with a 0.45-mm-diameter flat probe. Fruit were prepared for flesh firmness measurements by cutting each fruit longitudinally, from stem to blossom, and ensuring that the cut passed through the ground spot. One half of each fruit was then positioned on a staging platform with a rounded recession to secure the fruit. The force gauge was pressed into the fruit and peak resistance (Newtons) was recorded and averaged for five positions within the fruit: top, bottom, stem, blossom, and heart.
For SSC, pH, and lycopene content, 100-g samples were scooped from the center of each watermelon, bagged individually, and immediately placed on ice. Samples were kept on ice for no longer than 6 h before storage in a freezer (0 °F) until transport to the Plants for Human Health Institute (Kannapolis, NC) for analysis. Before analysis, samples were thawed and blended for 45 s until homogenized using a laboratory blender (model 7010S 1L; Waring Commercial, Torrington, CT). Samples were analyzed for SSC and pH using a digital refractometer (PAL-1; Atago USA, Bellevue WA) and digital pH meter (model H260G; Hach, Loveland, CO), respectively. Lycopene concentration was determined using 5-mL aliquots diluted in 15 mL deionized water. Sample absorbance at 560 and 700 nm were measured with a spectrophotometer (UltraScan PRO; HunterLab, Reston, VA). Lycopene content (micrograms per gram fresh weight) was calculated using the following formula: (Abs560-Abs700) × [dilution factor (DF)] × slope, where Abs560 and Abs700 are sample absorbance at 560 and 700 nm, respectively; DF is the ratio of sample weight to deionized water; and slope is a calculated correction (28) for the spectrophotometer using an external lycopene standard (Davis et al., 2003).
Treatments for both ‘Exclamation’ and ‘Extazy’ studies were arranged in a randomized complete block design with three replications of all treatment combinations in 2015 and four replications in 2016. Analysis of variance (ANOVA) was conducted separately for ‘Exclamation’ and ‘Extazy’ studies using PROC MIXED in SAS (version 9.4; SAS Institute, Cary, NC) with year, rootstock, and the interaction of year with rootstock treated as fixed effects, and rep nested within year treated as a random effect. All dependent variables were checked for signs of heteroscedasticity using residual plots output from PROC MIXED. Marketable fruit number, total fruit number, and leaf area required a square root transformation to meet the ANOVA assumptions of normality and homoscedasticity. Following ANOVA and means separation, these data were back-transformed for presentation in tables for ease of interpretation. If no significant interaction of year with rootstock was observed at a 0.05 significance level, then least square means for rootstocks were averaged over years. Means separation was achieved using the post hoc Tukey’s honestly significant difference (hsd) adjustment for multiple comparisons with the significance level of P = 0.05.
Watermelon leaf area.
Watermelon leaf area at 1, 2, and 3 WAT were recorded separately for ‘Exclamation’ and ‘Extazy’ studies. In the ‘Exclamation’ study, no significant interaction of year with rootstock was observed for leaf area measurements at 1 and 2 WAT; thus, means separation was conducted using combined means across both years. However, a significant interaction of year with rootstock was observed at 3 WAT, and means from 2015 and 2016 are reported separately (Table 2). At 1 WAT, nongrafted ‘Exclamation’ had the smallest leaf area and measured only 53% of ‘Marvel’ rootstock, which had the largest leaf area at that time. At 2 WAT, nongrafted ‘Exclamation’ had an intermediate leaf area relative to grafted treatments, and it did not differ statistically from ‘Camelforce’, which exhibited the largest leaf area. At 3 WAT, nongrafted ‘Exclamation’ produced the largest leaf area in 2015, but it was only statistically different from two rootstocks (‘Emphasis’ and ‘Kazako’). In 2016, no statistical differences were observed in leaf area at 3 WAT.
Comparison of the effect of rootstock selection on ‘Exclamation’ watermelon leaf area using image analysis outputs from Easy Leaf Area software (Easlon and Bloom, 2014) of overhead images captured at a height of 7 ft (2.1 m) of two standard watermelon plants at 1, 2, and 3 weeks after transplanting (WAT) from field trials in Kinston, NC, in 2015 and 2016. For all images, a 3 × 3-cm (1.2 inch) red scale was used as a reference to calibrate the Easy Leaf Area software and determine watermelon leaf area.
Analysis of ‘Extazy’ leaf area detected a significant interaction of year with rootstock at 1 WAT; thus, means from 2015 and 2016 are reported separately (Table 3). At 2 and 3 WAT, no significant interactions of year with rootstock existed; thus, means separation was conducted using combined means across both years. At 1 WAT, no differences in leaf area were observed among rootstocks in 2015, and little separation was observed in 2016 with 18 of 21 treatments producing leaf area that was not significantly different from any other treatment. No differences were observed in leaf area at 2 WAT. At 3 WAT, ‘Kazako’ rootstock produced a very small leaf area, only 53% of the nongrafted ‘Extazy’. Means separation of leaf area according to Tukey’s hsd indicated that only nongrafted ‘Extazy’ and ‘Kazako’ differed from one another at 3 WAT, and all other treatments shared a common statistical group (Table 3).
Comparison of the effect of rootstock selection on ‘Extazy’ watermelon leaf area using image analysis outputs from Easy Leaf Area software (Easlon and Bloom, 2014) of overhead images captured at a height of 7 ft (2.1 m) of two mini watermelon plants at 1, 2, and 3 weeks after transplanting (WAT) from field trials in Kinston, NC, in 2015 and 2016. For all images, a 3 × 3-cm (1.2 inch) red scale was used as a reference to calibrate the Easy Leaf Area software and determine watermelon leaf area.
Leaf area measurements at 1 WAT were indicative of initial seedling size rather than the rate of seedling growth. Watermelon seedlings did not have rapid growth between 0 and 1 WAT; furthermore, the one-cotyledon grafting method affected initial leaf area due to the large rootstock cotyledons, particularly the ISH rootstocks. The leaf area results suggest that, aside from two poor-performing rootstocks (‘Kazako’ and ‘Emphasis’), grafting does not significantly affect early season growth of watermelon plants, and most treatments produce similar leaf area at the end of the initial 3 WAT.
Watermelon yield and fruit number.
Means separation was conducted according to Tukey’s hsd to compare the effects of rootstock on watermelon marketable yield, total yield, marketable fruit number, average marketable fruit size, and overall average fruit size. For most of the yield responses, no significant interaction of year with rootstock was observed, and least square means were averaged across both years (Table 4). However, a significant interaction (P = 0.026) of year with rootstock was observed in total yield of ‘Extazy’ mini watermelon; thus, means were reported separately by year (Table 5).
Comparison of the effect of rootstock selection on yield parameters of ‘Exclamation’ standard watermelon in Kinston, NC, in 2015 and 2016. Ripe fruit were harvested four times each season at weekly intervals, weighed individually, and then summed within each year to determine yield parameters.
Comparison of the effect of rootstock selection on yield parameters of ‘Extazy’ mini watermelon in Kinston, NC, in 2015 and 2016. Ripe fruit were harvested five times and three times in 2015 and 2016, respectively. Fruit were weighed individually and then summed within each year to determine yield parameters.
In the standard fruit size ‘Exclamation’ study, nongrafted plants produced the highest marketable yield, total yield, marketable fruit number, total fruit number, and average marketable fruit size (Table 4). ‘Exclamation’ grafted onto ‘Pelops’, ‘TZ148’, and ‘Coloso’ produced similar marketable yield to nongrafted ‘Exlcamation’, and the next highest marketable yield was ‘Root Power’ rootstock which produced only 69% of the nongrafted ‘Exclamation’. The lowest marketable yields were observed when ‘Exclamation’ was grafted onto ‘Marvel’ and ‘Kazako’ rootstocks which produced 47% and 32% of nongrafted ‘Exclamation’, respectively. Rootstock had a relatively small effect on average fruit size andaverage marketable fruit size. Nongrafted ‘Exclamation’ produced the largest marketable fruit, but the average marketable fruit size was similar to 15 of the remaining 20 rootstocks (Table 4). And the smallest marketable fruit (‘Kazako’) were 82% of the nongrafted ‘Exclamation’ treatment.
To further describe watermelon fruit size distribution, ‘Exclamation’ watermelons were graded according to bin counts (Coolong, 2015). Because total yield differed between rootstocks, the percentages (by weight) of total yield in each category were reported for each rootstock treatment (Table 6). A significant effect of rootstock was observed for culls, 60-count, and marketable fruit. With 34% of its yield graded as culls, ‘Kazako’ rootstock produced the highest percentage of nonmarketable fruit (and correspondingly, the lowest percentage of marketable fruit). A significant difference between the percentages of 60-count fruit was only observed between ‘AQ’ and ‘BS1’ rootstocks, which accounted for 26% and 51% of the total yield, respectively (Table 6). No differences were observed in 45-count, 36-count, and 30-count categories.
Comparison of the effect of rootstock selection on fruit size distribution (by weight) of ‘Exclamation’ standard watermelon at Kinston, NC, in 2015 and 2016. Ripe fruit were harvested four times each season at weekly intervals, weighed individually, and then summed within each year, according to bin counts.
In the ‘Extazy’ mini watermelon study, nongrafted plants produced the highest marketable yield, total yield, marketable fruit number, and total fruit number (Table 5). ‘Extazy’ grafted onto ‘Coloso’, ‘Macis’, and ‘UG29A’ rootstocks produced similar marketable yields to nongrafted ‘Extazy’. The next highest marketable yield was ‘Camelforce’ rootstock, which produced 72% of the marketable yield of nongrafted ‘Extazy’. Similar to standard size ‘Exclamation’ watermelon results, ‘Kazako’ rootstock produced the lowest marketable yields (48% of nongrafted ‘Extazy’). Marketable and total fruit numbers followed similar patterns with highest fruit numbers produced by nongrafted ‘Extazy’, and lowest marketable and total fruit numbers produced by ‘Kazako’ and ‘Super Shintosa’ rootstocks, respectively. However, rootstock had no significant effect on average marketable fruit size (P = 0.073); and, despite a significant effect of rootstock on average fruit size (P = 0.033), Tukey’s hsd multiple comparisons adjustment detected no significant differences among rootstock treatments.
Although studies were designed to compare the performance of individual rootstocks, notable trends were observed among marketable watermelon yield when grafting to different rootstock species. Highest marketable yields in the ‘Exclamation’ study included the nongrafted control, two bottle gourd rootstocks (‘Coloso’ and ‘Pelops’), and an ISH rootstock (‘TZ148’) (Table 4). In the ‘Extazy’ study, highest marketable yields included the nongrafted control and three bottle gourd species [‘Coloso’, ‘Macis’, and ‘UG29A’ (Table 5)]. Thus, the only treatments to consistently produce the highest marketable yields across both ‘Exclamation’ and ‘Extazy’ studies were the nongrafted controls and ‘Coloso’ rootstock. Neither of the summer squash rootstocks (‘AQ’ and ‘BS1’) nor the giant pumpkin rootstock (‘Root Power’) produced marketable yields similar to nongrafted treatments when grafting with ‘Exclamation’ or ‘Extazy’. Although an ISH rootstock (‘TZ148’) produced a similar marketable yield to nongrafted ‘Exclamation’, no ISH rootstock produced marketable yields similar to nongrafted ‘Extazy’. ‘Kazako’ ISH rootstock produced among the lowest marketable yield, total yield, marketable fruit number, and total fruit number in both studies.
Watermelon fruit quality.
Analysis of fruit quality included watermelon flesh firmness (Newtons), acidity (pH), SSC, and lycopene content (micrograms per gram lycopene·fresh tissue). No interaction of year with rootstock was observed for any dependent variable; thus, means separation was conducted using the combined means across both years in each experiment (Tables 7 and 8). Rootstock selection had a significant effect on flesh firmness in both ‘Exclamation’ and ‘Extazy’ experiments (P < 0.001 and 0.001, respectively); and fruit from the nongrafted plants in both studies had the lowest flesh firmness. A significant effect of rootstock on fruit acidity existed in the ‘Exclamation’ study (P = 0.041), but means separation with post hoc Tukey’s hsd multiple comparisons adjustment only detected a difference between rootstocks with the lowest and highest fruit acidity—‘Pelops’ (pH 5.82) and ‘Cobalt’ (pH 5.51), respectively; and no statistical difference in acidity was observed among all other treatments (Table 7). By contrast, acidity was unaffected by rootstock selection (P = 0.591) in the ‘Extazy’ study (Table 8). Rootstock had no effect on SSC or lycopene content in ‘Exclamation’ or ‘Extazy’ studies.
Results of postharvest analyses to determine the effect of rootstock selection on ‘Exclamation’ watermelon fruit quality using two-fruit samples from each of two harvests from field trials in Kinston, NC, repeated in 2015 and 2016. Immediately following harvest, flesh firmness was measured by cutting watermelon longitudinally, from stem to blossom, and then pressing a force gauge into five positions on each fruit: top, bottom, stem, blossom, and heart. Flesh firmness was recorded as peak resistance and averaged across the five positions, two-fruit samples, harvests, and years. Watermelon acidity, soluble solids concentration (SSC), and lycopene content were determined using a 100-g (3.53 oz) sample scooped from center of each watermelon then bagged individually, placed on ice, and analyzed in the laboratory. Watermelon acidity, SSC, and lycopene content were averaged across two-fruit samples, harvests, and years.
Results of postharvest analyses to determine the effect of rootstock selection on ‘Extazy’ mini watermelon fruit quality using two-fruit samples from the initial two harvests from field trials in Kinston, NC, repeated in 2015 and 2016. Immediately following harvest, flesh firmness was measured by cutting watermelon longitudinally, from stem to blossom, and then pressing a force gauge into five positions on each fruit: top, bottom, stem, blossom, and heart. Flesh firmness was recorded as peak resistance and averaged across the five positions, two-fruit samples, and harvests. Watermelon acidity, soluble solids concentration (SSC), and lycopene content were determined using a 100-g (3.53 oz) sample scooped from the center of each watermelon, bagged individually, placed on ice, and analyzed in the laboratory.
Although watermelon grafting may provide benefits in the presence of biotic and abiotic stressors, the present study demonstrates that grafted plants do not increase yields nor increase fruit quality when production systems are optimized and fields are absent of environmental stress. Marketable yield, total yield, marketable fruit number, and total fruit number were highest in nongrafted ‘Exclamation’ and ‘Extazy’. Our findings demonstrated that several bottle gourd rootstocks (‘Emphasis’, ‘Coloso’, and ‘Pelops’) produced high marketable yields, comparable to nongrafted controls; however, all but one ISH rootstock (‘TZ148’) had reduced marketable yields. The present findings agree with several previous studies where grafted plants yielded similarly or worse than nongrafted plants (Bertucci et al., 2017; Kokalis-Burelle et al., 2016; Miller et al., 2013), but run contrary to others where yield increases were observed as a result of grafting (Miguel et al., 2004; Mohamed et al., 2012; Turhan et al., 2012). When comparing marketable yield based on rootstock species, our results agreed with previous research, which reported highest yields when grafting to bottle gourd rootstocks and decreased yields when grafting to ISH rootstocks (Yetisir et al., 2003). An increase in flesh firmness was associated with grafting, which was also reported in other studies (Davis and Perkins-Veazie, 2005; Kyriacou and Soteriou, 2015; Liu et al., 2017). Except for two rootstocks grafted with ‘Exclamation’ (‘Pelops’ and ‘Cobalt’), watermelon acidity was unaffected by grafting. And no differences were observed in SSC or the lycopene content for either watermelon type. A review article summarizes the effects of grafting on watermelon fruit quality characteristics, and it is concluded that rootstock–scion combinations may have differing effects on flesh firmness and SSC (Davis et al., 2008a). Specifically, watermelon grafting has been associated with reduced flesh firmness and decreased SSC (Lee and Oda, 2003), but one study found increased SSC, flesh firmness, and lycopene content as a result of grafting (Davis and Perkins-Veazie, 2005), whereas another study observed no difference in SSC between grafted and nongrafted watermelon (Miguel et al., 2004). Thus, variability exists in the literature regarding the effects of grafting on watermelon fruit quality. The results of the present study suggest that grafting ‘Exclamation’ and ‘Extazy’ to the selected rootstocks increases watermelon flesh firmness, but does not affect SSC or lycopene content of watermelon in a fumigated plasticulture production system.
Environmental conditions and production systems play a role in the performance of grafted plants, and some variability is to be expected when trials are conducted across differing geographies. Furthermore, specific rootstock–scion combinations may perform differently depending on environmental conditions. However, the present study evaluated two watermelon types (standard and mini-sized watermelon) with a diverse selection of 20 rootstocks. Authors observed that watermelon grafting with the studied rootstock–scion combinations did not improve yield and that only minimal fruit quality benefits occurred in fumigated North Carolina plasticulture production systems. This disparity between the present study and previous studies, which have reported an increase in yield as a result of grafting (Miguel et al., 2004; Mohamed et al., 2012; Turhan et al., 2012), is likely due to the fumigation and disease-free conditions of this study. Mohamed et al. (2012) evaluated grafted and nongrafted watermelons in fusarium wilt-infested fields and concluded that differences in yield were primarily explained by differences in the survival of plants: 67% among nongrafted plants and 83% to 100% among grafted plants. Thus, grafting watermelon offers utility for resistance to fusarium wilt or verticillium wilt (Verticillium dahliae), two diseases incited by soilborne pathogens (Guan et al., 2012; Keinath and Hassell, 2013; Kleinhenz, 2015). Two economic analyses indicate that watermelon grafting would be profitable in the United States, particularly when fields are infested with fusarium wilt or verticillium wilt (Taylor et al., 2008; Wimer et al., 2015). Therefore, it would be most beneficial to consider disease resistance characteristics of a rootstock and implement grafting only in situations where soilborne pathogens are anticipated to interfere with watermelon production. Our study highlights a selection of rootstocks that offer highest yield in the absence of soilborne disease or environmental stresses, and the data presented here may be used to compliment decisions for disease management.
BertucciM.B.JenningsK.MonksD.JordanD.LouwsF.SchultheisJ.R.SmithS.BasingerN.T.WaldschmidtM.2017Influence of grafting on the critical period for weed control in grafted watermelonHortScience52S74(abstr.)
CollaG.RouphaelY.MirabelliC.CardarelliM.2011Nitrogen-use efficiency traits of mini-watermelon in response to grafting and nitrogen-fertilization dosesJ. Plant Nutr. Soil Sci.1746933941
CoolongT.2015Trial report: Seedless watermelon variety evaluation. 5 May 2018. <http://blog.extension.uga.edu/colquittag/files/2016/01/2015-UGA-Tifton-Watermelon-Variety-Trial-Results.pdf>
DaleyS.HassellR.2014Fatty alcohol application to control meristematic regrowth in bottle gourd and interspecific hybrid squash rootstocks used for grafting watermelonHortScience49260264
DavisA.R.Perkins-VeazieP.SakataY.Lopez-GalarzaS.MarotoJ.V.LeeS.HuhY.SunZ.MiguelA.KingS.R.CohenR.LeeJ.2008bCucurbit graftingCrit. Rev. Plant Sci.275074
DittmarP.J.MonksD.W.SchultheisJ.R.2010Use of commercially available pollenizers for optimizing triploid watermelon productionHortScience45541545
EaslonH.M.BloomA.J.2014Easy leaf area: Automated digital image analysis for rapid and accurate measurement of leaf areaAppl. Plant Sci.271400033
GalinatoS.MilesC.WimerJ.2016Non-grafted and grafted seedless watermelon transplants: Comparative economic feasibility analysis. Washington State Univ. Ext. Publ. TB08
KeinathA.P.HassellR.L.2013Control of fusarium wilt of watermelon by grafting onto bottlegourd or interspecific hybrid squash despite colonization of the rootstocks by Fusarium oxysporumPlant Dis.98255266
KleinhenzM.D.2015Description of commercial cucurbit rootstocks. Common cucurbit diseases and pests and susceptibility characteristics. 5 Feb. 2015. <http://www.vegetablegrafting.org/wp/wp-content/uploads/2015/02/usda-scri-cucurbit-rootstock-table-feb-15.pdf>
Kokalis-BurelleN.ButlerD.M.HongJ.C.BausherM.G.McCollumG.RosskopfE.N.2016Grafting and paladin pic-21 for nematode and weed management in vegetable productionJ. Nematol.484231240
KubotaC.McClureM.A.Kokalis-BurelleN.BausherM.G.RosskopfE.N.2008Vegetable grafting: History, use, and current technology status in north AmericaHortScience4316641669
KumarP.RouphaelY.CardarelliM.CollaG.2017Vegetable grafting as a tool to improve drought resistance and water use efficiencyFront. Plant Sci.81130
KyriacouM.C.SoteriouG.2015Quality and postharvest performance of watermelon fruit in response to grafting on interspecific cucurbit rootstocksJ. Food Qual.3812129
LiuQ.ZhaoX.BrechtJ.K.SimsC.A.SanchezT.DufaultN.S.2017Fruit quality of seedless watermelon grafted onto squash rootstocks under different production systemsJ. Sci. Food Agr.971447044711
LouwsF.J.RivardC.L.KubotaC.2010Grafting fruiting vegetables to manage soilborne pathogens, foliar pathogens, arthropods and weedsScientia Hort.1272127146
MiguelA.MarotoJ.BautistaA.BaixauliC.CebollaV.PascualB.LopezS.GuardiolaJ.2004The grafting of triploid watermelon is an advantageous alternative to soil fumigation by methyl bromide for control of fusarium wiltScientia Hort.1031917
MillerG.KhalilianA.AdelbergJ.W.FarahaniH.J.HassellR.L.WellsC.E.2013Grafted watermelon root length density and distribution under different soil moisture treatmentsHortScience4810211026
MohamedF.H.El-HamedK.ElwanM.W.M.HussienM.N.E.2012Impact of grafting on watermelon growth, fruit yield and qualityVeg. Crops Res. Bul.7699118
MorenoB.JacobC.RosalesM.KrarupC.ContrerasS.2016Yield and quality of grafted watermelon grown in a field naturally infested with fusarium wiltHortTechnology26453459
RouphaelY.CardarelliM.CollaG.ReaE.2008Yield, mineral composition, water relations, and water use efficiency of grafted mini-watermelon plants under deficit irrigationHortScience43730736
TaylorM.BrutonB.FishW.RobertsW.2008Cost benefit analyses of using grafted watermelon transplants for fusarium wilt disease controlActa Hort.782343350
TurhanA.OzmenN.KuscuH.SerbeciM.S.SenizV.2012Influence of rootstocks on yield and fruit characteristics and quality of watermelonHort. Environ. Biotechnol.534336341
VinsonE.WoodsF.KembleJ.Perkins-VeazieP.DavisA.KesslerJ.2010Use of external indicators to predict maturity of mini-watermelon fruitHortScience4510341037
WimerJ.InglisD.MilesC.2015Evaluating grafted watermelon for verticillium wilt severity, yield, and fruit quality in Washington stateHortScience5013321337
YetisirH.SariN.YucelS.2003Rootstock resistance to fusarium wilt and effect on watermelon fruit yield and qualityPhytoparasitica312163169