Introduction

In recent years, grafted watermelon production has become widespread in several regions of the world. The progressive increase of this technology is due to the desirable characteristics of rootstocks, such as tolerance to soil diseases, low and high temperatures, and salinity (^{Boughalleb, Tarchoun, El Mbarki, & El Mahjoub, 2007}; ^{Schwarz, Rouphael, Colla, & Venema, 2010}; ^{Yetisir & Uygur, 2010}), and the efficient use of water and nutrients (^{Rouphael, Cardarelli, Colla, & Rea, 2008}; ^{Colla, Rouphael, Marabelli, & Cardarelli, 2011}). The rootstocks commonly used are interspecific hybrids (Cucurbita maxima × Cucurbita moschata) and Lagenaria siceraria (^{Myung-Lee et al., 2010}), which favor fruit growth and yield (^{Yetisir, Kurt, Sari, & Tok, 2007}; ^{Yetisir & Uygur, 2009}; ^{Islam, Bashar, Howlader, Sarker, & Al-Mamun, 2013}).

Internal fruit quality is a determining factor for the commercialization of watermelon. Physico-chemical characteristics such as soluble solids, firmness and color components vary depending on the variety (^{Pardo, Gómez, Tardáguila, Amo, & Varón, 1997}). In grafted plants, quality has been found to be related to rootstock and variety (^{Petropoulos, Khah, & Passam, 2012}; ^{Petropoulos et al., 2014}). Other researchers argue that there are no significant effects on the expression of internal quality attributes in watermelon; rather, the differences are given by the variety and sequence in fruit cutting (^{Camacho & Fernández, 2000}). Also, the postharvest quality of watermelon varies according to handling procedures, temperature conditions and relative humidity during storage (^{Perkins-Veazie & Collins, 2006}; ^{Yau, Rosnah, Noraziah, Chin, & Osman, 2010}). In general, the storage conditions are: 10 to 15 °C and relative humidity of 85 to 90 % (^{Risse et al., 1990}). Some adverse effects on postharvest watermelon quality are accelerated ripening, reduced firmness, fruit discoloration, and reduced soluble solids content (^{Davis & Perkins-Veazie, 2005}).

The use of Lagenaria and squash as rootstocks in watermelon production has been widely documented in several studies (^{Myung-Lee et al., 2010}; ^{Yetisir & Uygur, 2010}); however, in terms of post-harvest quality, it is necessary to generate more information, especially on germplasm in Mexico since it has been little studied. The present work was carried out with the objective of evaluating the effect of eight rootstocks, six native Lagenaria siceraria varieties (L43, L46, L48, L50, L54 and L56) and two commercial squash varieties (Super shintosa and TZ 148), in addition to the plant (without grafting), on the post-harvest quality of watermelon variety Tri-X 313.

Materials and methods

This research was carried out in the Institute of Agricultural Sciences’ Experimental Field at the Autonomous University of Baja California (ICA-UABC), Mexico, during the 2014 spring-summer growing season. Planting was made in an open field in clay soil with a pH of 7. Water and fertilization were supplied via drip irrigation. Fertilizer doses were applied according to the recommendations of ^{Fernández-Cara (1998)}, 95-73-106 NPK, supplemented with 33 kg of CaO and 11 kg of MgO.

Watermelon (Citrullus lanatus [Thunb.] Matsum. & Nakai) seedlings of the hybrid Tri-X 313 were grafted onto native Lagenaria siceraria materials (L43, L46, L48, L50, L54 and L56), collected in different regions of Mexico (Table 1), and two commercial squash rootstocks (TZ 148 and Super shintosa). The cleft grafting technique was used (^{Maroto, Borrego, Miguel-Gómez, & Pomares- García, 2002}). Ungrafted watermelon plants were used as the control. Watermelon variety 2800 was used as the pollinator of Tri-X 313, at a 3:1 ratio. Two bee (Apis mellifera) hives were established from the beginning of flowering to promote flower pollination. The harvest was based on maturity indicators: dried stipule and tendril near the fruit peduncle (^{Miguel et al., 2004}). Twelve fruits were randomly taken per experimental unit for each treatment.

In the post-harvest fruit quality analysis, a completely randomized design with a factorial arrangement was used, considering the factors rootstocks and storage time. Eighteen treatments (Table 2) resulting from 8 rootstocks, one control and two storage periods (0 and 14 days) were evaluated. Each treatment was established in triplicate, considering three fruits per replicate. Storage conditions were 15 to 17 °C and relative humidity of 80 %. The evaluated variables were fruit weight and color, pulp firmness and color, and total soluble solids.

Firmness (Newton) was quantified with a Chatillon DFE-100 digital force gauge (AMETEK Inc., USA). The content of total soluble solids (°Brix) was determined with a Digital Reichert AR200 refractometer (Reichert Inc., New York, USA). Color was obtained using an X-Rite SP62 sphere spectrophotometer (^{X-Rite Inc., USA}), which expresses values as L*, C* and °h, where: L* (lightness) defines color clarity, C* (chroma) indicates color saturation and °h (hue) indicates hue angle (^{X-Rite, 2002}). The data were subjected to analysis of variance and comparison of means in the ^{Statistical Analysis System program (2009)}, using the Tukey test (P ≤ 0.05) for the comparison of means.

Results and discussion

The postharvest quality of watermelons evaluated at the beginning and end of the period (14 days) was not altered by the storage time and grafted plant condition interaction (P ≤ 0.05). However, when independently considering the species of rootstocks employed and the storage time in relation to the control plant, different behaviors were observed (Table 3).

Fruit weight increased by 44 % (Table 4); this variation is associated with grafting (P ≤ 0.001). The superiority of the grafted plant was mostly expressed with Cucurbita rootstocks, where the increase was 55 % (P ≤ 0.001), while when using Lagenaria the increases were 40 % (P ≤ 0.001). However, the fruit weight response in both rootstock species was not significant (P = 0.053). Regarding the specific behavior of the rootstocks, the ones with the greatest increase in fruit weight were Super shintosa, L54 and L56 (Table 5). Fruit weight was a parameter that did not vary significantly (P ≤ 0.05) due to the storage effect, although the tendency was to fall by up to 1.2 % (Table 5).

These results show that grafting is an alternative to increase fruit weight, but it depends on the rootstock used. Similar results were reported by ^{Alan, Ozdemir, and Gunen (2007)}, who recorded higher fruit weight in plants grafted onto Lagenaria and Cucurbita hybrids under microtunnel conditions. ^{Perkins-Veazie and Collins (2006)} and ^{Yau et al. (2010)} documented 1-2 % weight loss in watermelon fruit due to the effect of the storage period. The weight reduction is attributed to moisture loss by evaporation through the fruit epidermis (^{Yau et al., 2010}), which is related to the presence of a wax layer (n-paraffin) in the fruit epicarp that acts as a moisture barrier (^{Panchev, Pashova, Radev, Petrov, & Kovacheva, 2014}).

Pulp firmness showed significant statistical differences (P ≤ 0.001) due to grafting and storage time (Table 3). The combination of watermelon plus Cucurbita rootstock generated greater firmness compared to the control (P = 0.001), while the Lagenaria rootstocks did not show significant variations (P = 0.345) (Table 4). This behavior was also evident when comparing both species, where Cucurbita was 30 % firmer than Lagenaria (P ≤ 0.001). The rootstock with the greatest increase in firmness was Super shintosa (Table 5), which surpassed the control plant by 29 %.

As for the storage effect, in general, firmness losses of 14 % (P ≤ 0.05) were obtained, with TZ 148 fruits having the greatest deterioration. Similar results were highlighted by ^{Huitrón-Ramírez, Ricárdez-Salinas, and Camacho-Ferre (2009)}, who recorded increases of up to 23 % in pulp firmness when using the Shintosa Camelforce (Cucurbita hybrid) rootstock with the Tri-X 313 variety. ^{Álvarez-Hernández, Castellanos-Ramos, Aguirre-Mancilla, Huitrón-Ramírez, and Camacho-Ferre (2015)} found increases of 4.3 to 18.8 % in pulp firmness in triploid watermelon variety Crunchy Red grafted onto Super shintosa. Likewise, it has been determined that Lagenaria rootstocks do not alter the consistency of watermelon pulp (^{Yetisir, Sari, & Yucel, 2003}). ^{Bruton, Fish, Roberts, and Popham (2009)} agree in pointing out that fruit firmness in triploid watermelon grafted onto Lagenaria "RS1332" is not significantly modified.

The content of total soluble solids was modified with the rootstocks (P ≤ 0.05), while the storage period had no significant effect (P ≥ 0.05) (Table 3). The grafted condition with Lagenaria rootstocks (Table 4) was 4 and 5 % higher in °Brix compared to Cucurbita (P = 0.040) and ungrafted plants (P = 0.043). The L54 and L56 rootstocks had a greater content of total soluble solids (12.55 and 12.44 °Brix, respectively), while Super shintosa and the control showed values of 11.24 and 11.63 °Brix, respectively (Table 5).

Consistent with these results, ^{Karaca et al. (2012}) and ^{Candir, Yetisir, Karaka, and Ustun (2013}) showed that there are Lagenaria genotypes that can favor °Brix content in watermelon. ^{Yetisir and Sari (2003)} and ^{Yetisir et al. (2003)} recorded statistically similar values in commercial hybrid rootstocks of Lagenaria and the normal condition. Other studies agree that the soluble solids content is not affected in triploid watermelons on hybrid rootstocks of Cucurbita in the combinations “Reina de Corazones/Shintosa”, “Tri-X 313/Shintosa Camelforce” and “Crunchy Red/Super Shintosa” (^{Miguel et al., 2004}; ^{Huitrón-Ramírez et al., 2009}; Álvarez- Hernández et al., 2015).

Watermelon pulp color varied with the rootstock and storage period (Table 3). Among the components that make up this parameter, pulp hue angle varied between grafted and ungrafted plants (P = 0.043), with the fruits of plants grafted with Lagenaria exhibiting a less-intense red color (P = 0.014), while with Cucurbita (Super shintosa and TZ 148) there were no significant statistical differences (P = 0.701) in pulp hue. The species Lagenaria siceraria was identified as having the lowest hue (P = 0.007) and greatest lightness (P = 0.001) in pulp compared to Cucurbita (Table 4). However, when comparing the different rootstocks, Lagenaria L43 was identified with similar characteristics in pulp color to the ungrafted plant (Table 6).

Fruits under storage conditions presented a pulp color transition from bright red to orange-red, opaque and clear. This variation was not related to the normal and grafted condition of the plants. The changes observed in hue (Δ°h = 4.40), lightness (ΔL* = 1.95) and chroma (ΔC* = -2.53) resulted from the storage days (15 to 17 °C, Table 6). These results coincide with those reported by ^{Karaca et al. (2012)}, who found a change in color from bright red to orange-red, arguing that this change is associated with the progressive level of senescence as a result of the days spent during storage, in addition to the temperature conditions present (^{Gil, Aguayo, & Kader, 2006}; ^{Perkins-Veazie & Collins, 2006}).

Conclusion

The postharvest quality of the watermelon fruits was not modified by the grafted condition. Grafting induces greater fruit weight, pulp firmness and °Brix. Pulp color was slightly affected by the Lagenaria rootstocks by presenting pulp with a less-intense red color in relation to the control plant fruits. Fruit quality due to the effect of storage (14 days, 15 to 17 °C) varied in firmness, °Brix and pulp color. The L43, L46 and L48 rootstocks of Lagenaria were identified as the most promising ones to reduce weight loss, retain firmness and maintain pulp color. For their part, L54 and L56 favor °Brix content, but they slightly diminish pulp color. The native L. siceraria variety rootstocks represent a potential genetic resource that can be integrated into breeding programs for the commercial production of grafted watermelon.