The cultivation of grafted plants gradually increased in Tunisia the last years. This technique was intended to reduce damage caused by soilborne pathogens, to boost plant growth and development, to control wilt caused pathogens and to increase the tolerance to biotic stresses such as viral, fungal and bacteria infection (Rivero et al., 2003; Edelstein et al., 2004; Cohen et al., 2007). In melons and watermelons, resistant Cucurbita rootstocks were used for different purposes. Intraspecific grafting is mainly used to ovoid damage due to wilt pathogens, such as Fusarium oxysporum f. sp. melonis and F. oxysporum f. sp. niveum for which resistance genes exists in certain melon varieties. This approach provides complete protection from the disease with no reduction in fruit quality and quantity (Cohen et al., 2002). Mechanisms of diseases tolerance in grafted plants may be due to the resistance of the rootstocks as it is accepted that the root system synthesizes substances resistant to pathogen attack and these are transported to the shoot through the xylem (Biles et al., 1989). The activity of these substances, related to disease resistance can vary during the development stages of grafted plants (Heo, 1991).

Fusarium wilt and Fusarium crown and root rot of watermelon induced by Fusarium oxysporum Schlechtend.: Fr. f. sp. niveum (E.F.Sm.) W.C. Snyder and H.N. Hans. (Fon) and Fusarium solani (Mart.) Sacc. f. sp. cucurbitae W.C. Snyder and H.N. Hans. (Fsc) race1, respectively, were probably the most damaging soilborne diseases of this cucurbit causing heavy economic losses (Bruton et al., 1998; Martyn and Bruton, 1989). These diseases were serious in many areas of the world such as Northern Africa, Italy, United States and Israel (Messiaen et al., 1991). In Tunisia, the problem becomes more and more important and the races of these fungi were frequently identified in the majority of watermelon cropping areas (Boughalleb, 2003; Boughalleb and El Mahjoub, 2005; Boughalleb et al., 2005).

The watermelon grafting onto cucurbit rootstocks is a best alternative to control soilborne diseases and an agronomic interest for plant vigour and production (Blancard et al., 1991; Messiaen et al., 1991; Gignoux, 1993; Jebari, 1994; Aounallah-Chouka and Jebari, 1999; Aounallah et al., 2002; Tarchoun et al., 2005). Benincasa cerifera, Cucurbita maxima, Lagenaria vulgaris and Lagenaria leucantha were reported by many authors as best watermelon rootstocks (Messiaen et al., 1991; Gignoux, 1993). In Tunisia, the hybrid RS 841 has often been used for melon grafting (Aounallah-Chouka and Jebari, 1999). This rootstock is known resistant to Fusarium wilt and to nematodes (Blancard et al., 1991). Since several years, many rootstocks such RS841, Emphasis, Strong Toza, originated from the hybrid Cucurbita moschata x Cucurbita maxima were used for cucurbit grafting (Blancard et al., 1991). Currently, several rootstocks were subscribed and recommended for watermelon grafting. Aounallah et al. (2002) showed that, among the three watermelon rootstocks tested (Lagenaria siceraria, local of Mahdia and RS 841), L. siceraria is resistant to Fon isolated but susceptible to Fsc. However, local of Mahdia is resistant to Fsc and RS 841 proved to be resistant to Fon and moderately resistant to Fsc. In the same way, the variety Sugar Baby used like control revealed to susceptible Fon and Fsc.

Recently, Boughalleb et al. (2007) showed that the use of grafted seedlings of sugar pack cultivar on Strong toza, TZ148, Emphasis, Achille and Ercole rootstocks were resistant to Fsc and Fon isolates. Authors concluded that these rootstocks could be interesting in grafting of watermelon to resolve the problem of two Fusarium species affecting watermelon.

Rootstocks-scion interaction showed against soilborne pathogens tolerance was indicated (Cohen et al., 2007). It is interesting to enlarge the game of available rootstocks tested in our conditions for each cropping watermelon cultivar.

The aim of this study is to evaluate resistance level of some new cucurbit rootstocks against Fon and Fsc isolates under controlled conditions as potential sources for grafting of commercial watermelon varieties.

Plant Material
The cultivar Delta grafted on the four rootstocks was tested in this research: TZ 148 as a positive control, Ferro, GV 100 and Just as new testing rootstocks. Non-grafted cv. Delta inoculated and none-inoculated were used as negative control. This study was conducted at the laboratory of Higher Institute of Agronomy in Chott Mariem, Tunisia in 2007.

Growth Conditions and Testing Rootstocks Resistance to Fusarium Wilt and Crown and Root Rot of Watermelon
Grafted plants were transplanted into pots containing previously sterilized peat. Plants were kept at growth chamber for 30 days at 23°C with a 16 h day length. A completely randomised design with 3 replicates was used. For each treatment, 30 plants were evaluated. Plants were watered daily and no fertilizers were applied.

Two Fusarium species were used for evaluation for resistance to F. solani f. sp. cucurbitae with two isolates (Fsc) and F. oxysporum. f. sp. niveum (Fon1 and Fon2) collected from different regions in Tunisia and preserved in glycerol (50%) to 4°C.

The method of inoculation used for resistance evaluation of plants was similar to that developed by Latin and Snell (1986).

Spore suspensions were prepared from cultures grown on PDB on a rotary shaker at room temperature (22°C) for 14 days and adjusted at a concentration of 1x10⁶ conidia mL^-1 with a hematocymeter. When the first true leaf was evident, the plants were uprooted and the roots washed under a stream of gently flowing water. Seedlings were root-dipped into the respective inocula for 15-20 sec, swirled several times and transplanted into 7.5 cm diameter pots (three seedlings per pot containing vermiculite) and five pots per isolate. Thus, ten plants per isolate were tested. Controls were prepared by root-dipping the plants into sterile distilled water. All plants were maintained in the greenhouse. The average air temperature, during the experiment, was about 27°C. The experiment was conducted three times.

Plants were classified as very resistant, when the level of mortality is ranged between 0 to 15%, resistant, if the percent of infested plants varied from 15 and 30% and were considered as susceptible when presented more than 30% of diseased plants. The survey of inoculated plants was conducted regularly at three stages of plant development: vegetative growth, flowering and fructification.

Statistical Analysis
Variance analysis of the treatment effect was made using SPSS software logicial. Means were compared by Duncan multiple test at 5% level.

Evaluation of Plant Resistance at Vegetative Growth Stage
The results showed a meaningful difference between the rootstocks to the three isolates of Fusarium spp. The rootstocks Ferro and Just seemed to be very resistant to all isolates of Fusarium spp. tested in this study. Grafted plants on TZ 148 demonstrated low percent of wilted plants for Fsc and Fon isolates (6.66%). Besides, Ferro, very resistant to Fon1 and Fon2, proved to be moderately susceptible to Fsc (Table 1). It revealed that at vegetative growth stage, all rootstocks could be efficient to resolve the problem of wilting and crown and root rot of commercial varieties.

Evaluation Plant Resistance at Flowering Stage
The results showed a meaningful difference between the rootstocks to the tested isolates. It revealed that only rootstock GV 100 remained very resistant only to Fsc (Table 2). Ferro and GV 100 showed a partial resistance to some isolates of Fon with percent of infested plants of 12%. TZ148 proved to be resistant to Fon2 (6.66%) but moderately susceptible to Fon1 and Fsc (Table 2).

Table 1:	Percentage of plants showing symptoms of wilting, crown and root rot of the grafted plants inoculated by F. solani f. sp. cucurbitae (Fsc) and F. oxysporum f. sp. niveum, at vegetative growth stage (Fon1 and Fon2)
The values followed by the same letter(s) are not significantly different at p<5%. The reading of the results has been done after 15 on 20 days of the inoculation

Table 2:	Percentage of plants showing symptoms of wilting, crown and root rot of the grafted plants inoculated by F. solani f. sp. cucurbitae (Fsc) and F. oxysporum f. sp. niveum, at flowering stage (Fon 1 and Fon 2)
The values followed by the same letter(s) are not significantly different at p<5%. The reading of the results has been done after 15 on 20 days of the inoculation

Table 3:	Percentage of plants showing symptoms of wilting, crown and root rot of the grafted plants inoculated by F. solani f. sp. cucurbitae (Fsc) and F. oxysporum f. sp. niveum, at fructification stage (Fon1 and Fon2)
The values followed by the same letter(s) are not significantly different at p<5%. The reading of the results has been done after 15 on 20 days of the inoculation

Evaluation Plant Resistance at Fructification Stage
The results of rootstocks behaviour, at fructification stage, against Fsc and Fon are shown in Table 3. At the end of the assay, it appeared that the rootstocks Ferro, GV 100 and Just were very good included to the two species of Fusarium spp. affecting watermelon and are considered like very resistant or resistant. However, Ferro, TZ148 and GV 100 showed an intermediate behaviour and were moderately resistant to Fon2 (Table 3).

Data obtained from this study suggest that grafting the susceptible watermelon cv. Delta is an effective control measure against of Fusarium wilt and Fusarium crown and root rot affecting watermelon. Similar results were reported by Trionfetti et al. (2002) and Miguel et al. (2004) on controlling Fusarium wilt by grafting two muskmelon cultivars and triploid watermelon respectively onto commercial rootstocks. Grafting was effective in controlling some other pathogens such as melon sudden wilt caused by Monosporascus cannonballus (Edelstein et al., 1999). Resistant rootstocks provide excellent control of many watermelon soilborne pathogens and particularly F. oxysporum f. sp. niveum, F. solani f. sp. cucurbitae, Monosporascus cannoballus and nematodes. In addition, watermelon grafting gave others advantages such as plant growth promotion, yield increase, extension of yield period and improvement of fruit quality (Tarchoun et al., 2005). Many rootstocks were evaluated for their resistance level to be introduced in Tunisia and used for resolving the problem of Fusarium spp. affecting watermelon (Boughalleb et al., 2007). However, it is recognized that grafting watermelon cultivars onto resistant rootstocks are more expensive, since both scions and rootstocks are expensive hybrids. In addition, the development of grafted plants requires more time, materials, space, a high level of expertise, improved cultivation methods and expensive post graft handling. But, actually in Tunisia, grafting is expected to increase significantly despite the high cost of labor and supplies, since it is one of the best alternative effective control methods found up to now for Fusarium spp. Lee (1994) reported that the tolerance disease exhibited by grafted plants could be explained by their vigorous roots. In fact, high and significant positive correlations were found between plant fresh weights and mortality rates in grafted plants developed during two different cropping seasons (Edelstein et al., 2004).

This mechanism has not been intensively investigated. The disease tolerance in grafted seedlings may be entirely due to the tolerance of stock plant roots to such diseases. However, in actual plantings, adventitious rooting from the scion is very common (Lee and Oda, 2003). Plants having the root systems of the scion and rootstock are expected to be easily infected by soilborne diseases. However, seedlings having dual root systems often exhibit excellent disease resistance, almost comparable to those having only rootstock roots. This observation partially supports the previous report that substances associated with Fusarium tolerance are synthesized in the root and translocated to the scion through the xylem (Biles et al., 1989). The activity of the substances related to disease resistance may vary during the development stages of the grafted plants (Heo, 1991).

The watermelon grafting proved to be an effective method to attenuate the impact of Fusarium wilt and Fusarium crown and root rot of watermelon caused by Fon and Fsc, respectively. The present study permitted to enlarge the game of rootstocks used for watermelon grafting. In fact, the three new rootstocks could be kept from this work.