Effects of rootstocks on fruit yield and some quality traits of watermelon (

In this study the effects of rootstocks on watermelon fruit quality and yield were investigated by comparing grafted plants with non-grafted and self-grafted ones in open field conditions in over two consecutive years (2013-2014). The watermelon cultivar Crimson Tide was grafted onto Ferro RZ, Maximus, Nun 9075, RS 841, No:3, Strong Tosa and TZ 148 rootstocks. Non-grafted and self-grafted plants were used as control. Grafting increased the average fruit yield between 17.45% and 52.56% compared to non-grafted treatment. It was concluded that among the tested rootstocks No:3 along with Nun 9075 and TZ 148 rootstocks can be advised due to the increase in fruit yield, total soluble solids and taste panel test values and that the use of grafted plants can be advantageous alternative in watermelon production for producers.


Introduction
Watermelon (Citrullus lanatus) originating from Africa is a popular vegetable in Turkey as well as in the world.The world watermelon production is 109 million tons and China provides 66.8% of the world production.Turkey is the second largest producer after China with 3.9 million tons in 2015 (FAO, 2016).
One of the major problems in the watermelon production is soil-borne pathogens.Fusarium spp., Pythium spp., Phytophthora spp., Rhizoctonia spp.and Verticillium spp.are widespread in watermelon production areas in the world and Turkey, which cause diseases such as root rot, wilting, collar blight and fruit rot (Yıldız and Delen, 1977;Sağır, 1988;Correll et al., 1991).Kırbağ and Turan (2005) reported that soils are infected with Pythium sp., F. solani, F. oxysporium and R. solani as a result of watermelon disease survey in the province of Malatya.Kurt et al. (2008) determined F. oxysporum f. sp.niveum in watermelons showing signs of wilt disease in the provinces of Adana, Mersin and Diyarbakir.
Grafting is the most popular method in watermelon production against the soil-borne pathogens and nematodes as an alternative to soil fumigants and other control methods (Crino et al., 2007).In addition to preventing soil borne diseases, the grafted plants show more tolerance to abiotic stress conditions and stimulate mineral nutrient uptake (Pulgar et al., 2000).
The first interspecific grafting of vegetables for disease management was grafting of watermelon onto a squash rootstock (Cucurbita moschata Duch.) in Japan in the 1920s (Tateishi, 1927).The cultivation of grafted vegetable plants began in Korea and Japan to prevent the reduction in yield due to disease and abiotic stress conditions in watermelon at the end of the 1920's.Grafted seedlings were commercially available for watermelon grafted onto gourd (Lagenaria siceraria (Mol.)Standl.) in the early 1930s in Japan.In the late 20th century grafting was introduced to Europe and Asia countries (Kubota, 2008).Currently, grafted seedlings are very widely used with more than 500 million grafted seedlings produced annually in Japan (Kobayashi, 2005), and almost 60% of vegetable production is done on these seedlings.This rate is stated to have reached up to 81 % in South Korea (Lee, 1994).Grafted seedling production in Turkey was initiated with a total of 70 000 grafted tomato seedlings for the first time in Antalya in 1998.This figure reached 120 million in 2014 and about half of them accounted for watermelon seedlings.Almost all of watermelon is produced from grafted seedlings both in open fields and under plastic tunnels in Turkey.
Watermelons are generally grafted on rootstocks of Cucurbita maxima x Cucurbita moschata hybrids, Cucurbita moschata and Lagenaria siceraria commercially.Rootstocks are effective on plant growth, yield and fruit quality (Turhan et al., 2012;Alvares-Hernandez et al., 2015), and they can thus cause variation in the fruit yield and quality associated with the combination of scion and rootstock (Alan et al., 2007).The performance of grafted plants is also affected by the environmental conditions and production methods (Cohen et al., 2007).Therefore, it is important to select watermelon compatible rootstocks and combinations of scion and rootstock showing good performance should be identified.
With respect to grafted watermelon in Turkey, studies were started by testing 10 rootstocks consisting of Lagenaria, landrace and cucurbit hybrids in 2000s (Yetişir and Sarı, 2003)

Materials and Methods
This experiment was carried out in fields located at Batı Akdeniz Agricultural Research Institute in Antalya between 2013 and 2014.A total of seven commercial hybrid rootstocks namely Ferro RZ (Rijk Zwaan), Maximus (Antalya Tarim), Nun 9075 (Nunhems), RS 841 (Monsanto), No:3 (Sakata), Strong Tosa (Syngenta) and TZ 148 (HM.Clause) were used in the experiment.All were Cucurbita maxima x Cucurbita moschata commercial hybrids.
The watermelon hybrid variety Crimson Tide (CT) as a scion was used and grafted onto Ferro RZ, Maximus, Nun 9075, RS 841, No:3, Strong Tosa, TZ 148 commercial hybrid rootstocks of C. maxima x C. moschata.Nongrafted and self-grafted plants were used as a control.Grafting was made by Fidesan Seedling Company when both rootstocks and scion had their first true leaves completely developed by using the tongue-approach technique (Oda, 1999).
Grafted seedlings were planted to the field on the 12 th of April in 2013 and on the 15 th April in 2014.The field soil was a sandy clay loam (59.0%sand, 14.0% silt and 27.0% clay) with pH 7.6, EC 0.30 dSm -1 and organic matter 1.3%.Rows were 2 m apart and the distance between plants within rows was 2 m and drip irrigation system was used.The experiment was designed in a randomized complete block design (RCBD) with three replications, eight plants per replication.While 5 kg da -1 N, 4 kg da -1 P 2 O 5 and 6 kg da -1 K 2 O were applied to soil from planting to flowering, 12 kg da -1 N, 3.5 kg da -1 P 2 O 5 and 17 kg da -1 K 2 O were delivered from flowering to the end of harvest.The experiment plots were irrigated with dripping system two times a week and fertilizer applications were made through drip irrigation.Mites and aphids were controlled by chemicals according to plant protection technical instructions.During the growing season, necessary cultural practices were carried out.Harvest time was determined based on the withering of the tendril closest to the fruit on the vine as proposed by Robinson (1997).Marketable fruits were harvested in each parcel and parcel yields were found and converted to kg da -1 . Fruits from plots were chosen to determine fruit rind thickness (mm), fruit flesh firmness (N), soluble solids (%), lycopene, taste panel test and fruit flesh color.Fruit firmness was measured in kg force mounting 8 mm probe on FT 327 penetrometer and converted to Newtons (N).A digital refractometer from Kruess and a digital caliper were used for determination of total soluble solids and fruit rind thickness, respectively.The nine-point Hedonic scale for evaluating fruit flesh taste was used.The taste of fruits was evaluated using a 1-9 scale by 10 panelists each replicate.1: dislike extremely, 2: dislike very much, 3: dislike moderately, 4: dislike slightly, 5: neither like nor dislike, 6: like slightly, 7: like moderately, 8: like very much and 9: like extremely (Naes et al., 1995;Arslan, 2010). 1/2 .Hue angle (h 0 ) represents a color wheel with red-purple at an angle of 0°, yellow at 90°, bluishgreen at 180°, and blue at 270°, and it was calculated by h 0 = tan −1 (b*/a*) (McGuire, 1992).
Data for all measurements were subjected to analyses of variance and the differences between the means were compared by using LSD test (p<0.05).

Fruit yield
During 2013 and 2014 growing seasons, the fruit yields for watermelon grafted onto the different C. moshata x C. maxima hybrid rootstocks are given in Table 1.Fruit yields were significantly affected by the rootstocks, but there was not a significant year*rootstock interaction.All the grafted plants produced significantly higher yield than ungrafted and self-grafted ones.Grafted plants onto No:3 rootstock gave the highest fruit yield in both years with the average of 4084 kg d -1 .No:3 rootstock was followed by Maksimus, Nun 9075, TZ 148, Strong Tosa, RS 841, Ferro rootstocks.The lowest fruit yields with the averages of 2677 kg da -1 and 2695 kg da -1 were obtained from ungrafted and self-grafted plants, respectively.
The harvested fruit size, yield and quality are evidently affected by the scion and environmental factors, but the rootstock may also have significant effects on plant growth and fruit quality (Davis et al., 2008).Many researchers found that grafting on hybrid rootstocks promoted fruit yield increase (Yetişir et al., 2003;Alexopoulos et al., 2007;Alan et al., 2007).In this study, an increase in fruit weight was determined to be consistent with previous studies.Since the interspecific hybrid rootstocks with vigorous root system are able to absorb water and nutrient elements more efficiently in addition to disease resistance, they are superior to non-grafted plants in terms of fruit yield (Huitron-Ramirez et al., 2009).

Content soluble solids (°Brix)
The content of soluble solids was significantly affected by grafting and significance in interaction of year*rootstock was also found (Table 1).showed that the content of soluble solids was affected by grafting.Alexopoulos et al. (2007) stated that grafted watermelons had less soluble solids content than the ungrafted ones.

Lycopene
The effects of watermelon rootstocks on lycopene contents were significant and also year*rootstock interaction was found significant (Table 1).The highest difference in lycopene content over years was in No:3 rootstocks, while the most stable rootstock was RS 841.In the present study, the average lycopene content varied between 63.1 and 84.5 µg/g.While the highest values of average lycopene contents were 84.5 µg/g and 80.7 µg/g for RS 841 and Strongthoza rootstocks, the lowest values of average lycopene contents were 63.13 µg/g and 64.9 µg/g for Ferro and Maximus rootstocks, respectively.Bruton et al. (2009) reported that lycopene content for watermelons ranged from 43.6 to 78.9 µg/g.Fekete et al. (2015) reported that interspecific hybrid rootstocks (C.maxima x C. moschata) increased the lycopene contents of watermelons.Contrary to the above-mentioned researcher, Turhan et al. ( 2012) didn't observe any differences in watermelon lycopene content for grafted and ungrafted plants.

Taste panel test
The effect of rootstock on the taste of watermelon fruits was found to be statistically significant and the year*rootstock interaction was also significant (Table 1).Crimson Tide CT/Maximus combination had lower taste scores than other scion/rootstock combinations.CT/Nun 9075 (7.77) and CT/No:3 (7.30) combinations had the highest taste panel test scores, whereas CT/Maximus (6.08) and CT/Strong Tosa (6.57) had the lowest taste scores.The greatest difference between years in tasting panel realized in Ferro RZ rootstock taking the lowest and highest values in two years, whereas Nun 9075 and No:3 rootstocks appeared to be higher values and more stable.On the contrary El-Wanis et al. ( 2013) didn't find any significant effect on the fruit taste panel test of grafted watermelons.It is known that scion/rootstock combination and environment can affect some fruit quality attributes of watermelon.Therefore, a suitable variety and rootstock combination should be carefully chosen for a particular watermelon production area.

Fruit rind thickness
Fruit rind thickness of watermelon is important in terms of resistance to transport and long shelf life.Fruit rind thickness values are presented in Table 1.These values ranged from 17.39 mm (Nun9075) to 11.73 mm (ungrafted CT).Grafting onto rootstocks affected significantly fruit rind thickness over self-grafted and ungrafted ones.The highest values were observed in watermelon fruits harvested from plants grafted onto Nun 9075 and Strong Tosa rootstocks, but the lowest values were in non-grafted watermelon fruits.Arslan (2010) found a higher rind thickness in fruits taken from grafted plants.The maximum rind thickness was measured 14.03 mm in the fruits of the watermelon variety Crisby grafted onto Ferro RZ rootstock, whereas the minimum rind thickness was measured 12.83 mm in the fruits of ungrafted Crisby.Alexopoulos et al. (2007) reported that grafting increased fruit size and yield and that the fruits of grafted plants had a thicker rind than the ungrafted.These results are in agreement with Tokgöz et al. (2015).Watermelon rind thickness varies depending on the selected rootstock.Fruit rind thicknesses of varieties grafted onto vigorous and late rootstocks increase.However, rind thickness is in the fruits of varieties grafted onto weak rootstocks (Alexopoulos et al., 2007;Edelstein et al., 2014).

Fruit flesh firmness
Watermelon fruit with less fibrous and more crispy tissue is a desirable feature.Fruit flesh firmness is closely related with the fibrousness.The effect of grafting on fruit flesh firmness was found to be statistically significant (P<0.05)(Table 1).Fruit flesh firmness ranged from 11.5 to 7.6 among rootstocks.The highest values were found in TZ 148 (11.5) and RS 841 (11.1) rootstocks, while the lowest values were found in Strong Tosa (8.8) and No:3 (7.6) rootstocks.Although firmness could be affected by different factors, grafting is one of the most important factors affecting fruit flesh firmness.
Rootstocks increased firmness by 38.5% with respect to the non-grafted control (Soteriou et al., 2014).In our study, fruit flesh firmness values were found mostly higher in grafted watermelons than ungrafted ones in agreement with Huitron-Ramirez et al. (2009), Petropoulos et al. (2014) and Bruton et al. (2009).Yetişir and Sarı (2003) stated an increase of fruit flesh firmness in watermelon plants grafted onto the rootstocks of C. maxima x C. moschata interspecific hybrids.On the other hand, Huitron et al. (2007) did not find any significant effect on fruit flesh firmness of grafted watermelons.

Pulp colorimetry
Flesh color of watermelon is the important quality feature attracting consumers.Flesh color, which consists of color lightness (L* value), colour intensity (C* value) and h 0 angle, is affected by the environment and scion/rootstock combination.Flesh color lightness, the intensity of flesh color and h 0 angle are seen to have affected by rootstocks (Table 2).As shown in Table 2, the CIE LCh L* C* and h values ranged from 37.10 to 44.18, from 28.92 to 33.52 and from 28.91 to 31.40, respectively.Grafting increased fruit flesh color lightness (L*) over control excluding Nun9075 and RS841 rootstocks.All rootstocks had more intense (higher C*) color than ungrafted ones.The effect of rootstock on the hue angle (h 0 ) of the fruit's flesh color was found to be significant.

Conclusion
Cultivation of grafted vegetables especially grafted watermelon is a common practice in Turkey as in the world.Grafted plants are preferred due to the resistance to soil-borne pathogens and higher fruit yield and quality.The properties such as fruit yield and quality are also influenced by environment and varieties (scions).Compatible and suitable rootstock/scion combinations with high fruit yield and quality should be carefully determined in each watermelon production region.With respect to this study, it was concluded that grafting onto rootstock could increase the average fruit yield between 17,45% and 52,56% compared to non-grafted treatments and among the tested rootstocks No:3 along with Nun 9075 and TZ 148 rootstocks can be advised due to the increase in fruit yield, total soluble solids and taste panel test values.
Fruit flesh color was determined by a Minolta CR-300 Chromameter.The CIE LCh color space values were used.The vertical L* axis represents lightness, ranging from 0 (black) to 100 (white).The other (horizontal) axes are represented by a* and b*.Chroma (C*) represents color saturation and was calculated using the formula C*=(a

Table 1 .
The values of fruit yield and quality traits for watermelon grafted onto various hybrid squash rootstocks in Antalya during[2013][2014]

Table 2 .
The values of fruit pulp colorimetry for watermelon grafted onto various hybrid squash rootstocks in ) was decreased by grafting except for Strong Tosa rootstock.Fruits from Crimson Tide (CT)/CT, CT/Nun9075, CT/Ferro, CT/No:3, CT/RS841 CT/TZ148 and CT/Maximus graft combinations had lower h 0 values compared to control fruits.Our study is compatible with the results of Petropoulos et al. (2014) reporting that fruit flesh color mostly depended on the rootstock-scion combination.Fekete et al. (2015) informed us that L* value varied from 35.17 to 42.71; C* value 31.59 to 38.41; h angle 33.42 to 36.13, and that while interspecific grafting decreased h angle value, it increased C* value.