Introduction

Jatropha curcas L., is considered an important plant for the biodiesel production by the content and quality of oil in its seeds (^{Achten et al., 2008}). It is a shrub that belongs to the Euphorbiaceae family (^{CATIE, 2003}; ^{Muñoz and Jiménez, 2009}), originally from Mexico and Central America, grows preferentially in tropics and subtropics, and it is susceptible to flooding (^{Berger, 2010}). It grows in different soils and agroclimatic conditions (^{Heller, 1996}; ^{Srivastava et al., 2011}); although it usually grows on stony or sandy soils with low nutrient content and soil moisture (^{Li et al., 2008}; ^{Wiebe et al., 2008}). This plant can be used to recover degraded and poorly fertile areas and is relatively resistant to pests and diseases (^{Francis et al., 2005}; ^{Zahawi 2005}). The genus Jatropha is morphologically diverse and comprises about 188 species, 48 of which are found in Mexico, of which 81% are endemic (^{Steinmann, 2002}; ^{Martínez et al., 2006}). Of these, J. curcas stands out because of the high oil content in its seeds and for this reason it has been considered as a promising bioenergetic source by the Ministry of Energy (SENER) in its plan to “introduce bioenergy sources” (^{SENER, 2009}).

The global interest in J. curcas arises because of the need to find a plant with a high oil content that is not edible (^{King et al., 2009}; ^{Lafargue et al., 2012}). The conversion of vegetable oil into biodiesel, is a process developed since the twentieth century, by Rudolf Diesel, when extracting peanut oil to run an engine. Recently, with the increase in oil prices associated with the possible decrease of fossil fuel reserves, there is a need to seek alternative energy sources (^{Salaet and Roca, 2009}, ^{SENER, 2012}), in order to reduce the use of fossil fuels and thereby reduce CO₂ emissions, and thus contribute to the mitigation of climate change (^{Mergier, 2007}).

Worldwide, there are different oil plants used for oil extraction and biodiesel production, such as sunflower, soybean, oil palm and oilseed rape. However, these oils are essentially for human consumption, which generated great criticism and discussions at a global level, since it can affect the food supply and impact on its price (^{Martínez et al., 2006}; ^{Achten et al., 2008}; ^{Muñoz y Jiménez, 2009}). Therefore, new sources of raw materials have been sought, including J. curcas, an option that is not a food product, and which has a high oil content in its seeds (^{Arruda et al., 2004}; ^{Nunes, 2007}). In addition, the remains after the extraction of oil can be used as fertilizer; the branches as energy and glycerin for soap making (^{Achten et al., 2009}). J. curcas, is a plant that has great properties for industrial use (^{Johannes and Hirata, 2007}), in medicine (^{Mujumdar and Misar, 2004}), as a coagulant or blood anticoagulant (^{Osoniyi and Onajobi, 2003}), for treating infections and sexually transmitted diseases in humans (^{Aiyelaagbe et al., 2007}). In Mexico there are genotypes of non-toxic J. curcas, whose seeds can be edible (^{Schmook and Sánchez, 2000}; ^{Martínez et al., 2010}). In the Totonacapan region in the state of Veracruz, the seeds of non-toxic J. curcas are consumed in typical dishes (^{Martínez et al., 2006}).

For the extraction of oil there are two methods, one of them is by pressing and the other with the use of solvents (^{Forson et al., 2004}). The extraction with solvents is better, since it presents a recovery of 95% of the oil with high quality and purity (^{Achten et al., 2008}).

The extraction by the solvent method, originated in Europe in the year 1870 with the Batch process. Processes based on this method, use solvents such as hexane, acetone or ether. The seed of J. curcas can have an oil content between 21% and 40%, depending on the environmental conditions and crop management (^{Gübitz et al., 1999}; ^{Shah et al., 2005}; ^{Henning, 2009}). Therefore, the objective of this paper was to determine the oil content and yield of different accessions of non-toxic J. curcas collected in different regions of the state of Veracruz, sexually and asexually propagated, and established in the central region of Veracruz.

Materials and methods

Study area. Plants of J. curcas were used from the Germplasm Bank, located in the Campus Veracruz of the College of Postgraduates located at the coordinates 19° 11’38.62” north latitude and 96° 20’ 31.26” west longitude, an altitude of 24 m. The climate is Aw (w) (i’) g, which corresponds to warm sub-humid with summer rains, an average annual precipitation of 1 100 mm and an average temperature of 26 °C and a temperature fluctuation of 5-7 °C with 5% precipitation in winter (^{García, 1988}).

Vegetative material. In 2011, fruits and vegetative material of J. curcas from different localities of the state of Veracruz were collected. These were propagated in seed and stake, to integrate the germplasm bank of Campus Veracruz (Table 1). The criterion for selecting the accesses to be established was for the oil content to be equal to or greater than 40%.

Table 1 Original geographical location of the 16 accesses of J. curcas L., which are part of the Germoplasm Bank, Veracruz Campus. 

The experimental design is in randomized complete blocks, with three replicates, in three-year-old plants. Each type of propagation was established in separate plots and the plants were transplanted at a distance of 3 m between groove for 2 m between plants. The fruit harvest was carried out during the months of May to October of 2014. The seeds were extracted and dried at room temperature, the oil content determinations were performed from November 2014 to February 2015 in the Laboratory of Water, Soil and Plant from the School of Postgraduate Campus Veracruz.

Sampling. A sample of 20 seeds of J. curcas, of three plants per access, was taken randomly. The initial weight of the sample was recorded, then dried in a Riossa stove (model H-33) at a temperature of 55 °C for five hours, allowing to stand for 20 min in a desiccator with silica.

Preparation of sample. Removal of the forehead was done manually. Almond grinding was performed with a mortar, then on an analytical balance 8 g of sample was weighed in an extraction thimble (alundum), sealed with cotton and placed inside the tube (sample).

Oil extraction. For the extraction of oil, the Goldfish ^® equipment was used. The sample was placed in the extraction equipment, closed with a ring and packing, 30 ml of hexane were placed in extraction vessels previously stabilized at constant weight and dried in the Riossa stove (model H-33), at a temperature of 100 °C for 1 hour. The extraction process lasted five hours, then the tubes were removed (sample) and the solvent recovery tubes were inserted into the equipment, once the solvent was recovered the glasses with the obtained oil were removed, they were put in the drying oven for four hours to completely remove the solvent. Finally, the oil was filtered with filter paper (Wattman) number 44, in 16 ml test tubes. The oil content was obtained according to the following formula: oil content (%)= oil yield (g)/sample weight (g)*100.

Statistic analysis. The obtained data were captured in an Excel Version 2010^® spreadsheet, statistical analysis was performed using the Statistical Analysis System (SAS)program, v. 9.4, with Anova and Tukey test of means (p≤ 0.05) by the procedure PROC GLM. In addition, a Pearson correlation analysis was performed to determine the relationship between the morpho-productive variables and the oil content.

Results and discussion

Table 2 shows the averages of the main morpho-productive variables, associated to the production and final yield of fruits, seeds and oil content in plants of three-year-old J. curcas established in seed propagation.

Table 2 Morphological characteristics of 16 accessions of J. curcas propagated by seed, coming from the state of Veracruz. 

Medias Tukey p≤ 0.05. ACC= acceso; APL= altura de planta; NR1= número de ramas primarias; NR2= número de ramas secundarias; NR= número de racimos; NF= número de frutos; PS= peso de semillas.

The analysis of variance detected significant statistical differences for the number of fruits (p= 0.021) and the weight in seed production (p= 0.007). According to the Tukey test (p≤ 0.05), I-34 access was the most outstanding in number of fruits (693) and seed weight (817.8 g plant^-1). In contrast to the I-14 access, which presented lower number of fruits (69) and I-31 access had lower seed weight (80.9 g plant^-1). These results are superior to what ^{Machado (2011)} found in the one-year-old Cabo Verde variety, which reports a production of 102 fruits. However, ^{Srivastava et al. (2011)} reports, 210 fruits in three-year seed-propagated plants. ^{Sosa-Segura et al. (2012)}, reports from 30 to 36 fruits as a maximum number per plant, in germplasm from the states of Puebla and Morelos and a seed production of 39 to 50 g plant^-1 year old. What is reported in this paper is superior to that found by ^{Sosa-Segura et al. (2012)}.

In this regard, ^{Francis et al. (2005)} report that using minimum inputs in the cultivation of uncultivated land in India a production of 0.370 kg plant^-1 in the first year was obtained, in the second year 0.925 kg plant^-1, without a doubt the age and the agronomic management favored the increase in the production. Also, ^{Wani et al. (2012)}, reported that of the seven accesses established in different localities, evaluated in 2007-2008, the J001 access reached a production of 0.635 kg plant^-1.

Regarding to the oil content in the seeds, there were no statistical differences in the evaluated accesses (Figure 1). The oil content was in the range of 54.15% to 60.98% with an overall average of 56.71%, therefore, all accesses present a good oil content, emphasizing that it was based on accesses with superior to 40% oil contents, without making any aggregation of inputs to the plants and this response is at the upper end of what was reported by ^{Martínez et al. (2011)}, who report variations in oil content from 18% to 60% in accesses collected in Mexico. On the other hand ^{Naresh et al. (2012)} reported that three-year-old plants in India, KM access had an oil content of 50% and ^{Cheng-Yuan et al. (2012)}, in plants from China report contents of up to 61%.

Figure 1 Oil content (%), in non-toxic J. curcas accesses in 3 year old seed propagated plants.  

This coincides with the results obtained in this study. ^{Silip et al. (2011)}, report contents from 59% to 63% and ^{Chen et al. (2012)} report values of 55.6%, which are within the ranges reported in this study. ^{Ovando-Medina et al. (2009)}, found that it is possible that there is a relationship in the variation of oil content (between 12.09% and 44.28%) in J. curcas, with the level of precipitation being higher in drier areas. That is, environmental conditions largely determine variables such as seed weight, protein content, and oil content (^{Heller, 1996}; ^{Sakaguchi and Somabhi, 1987}).

When making the correlations between the morpho-productive variables and the oil content (Table 3), for plants propagated by seed. A significant correlation was found between fruit numbers with seed weight (r= 0.79) and oil content (r= 0.74), as well as seed weight and oil content (r= 0.79). Sosa-Segura et al. (R= 0.935) and with the number of fruits per plant (r= 0.991).

Table 3 Correlation analysis between morpho-productive variables in J. curcas L. plants propagated by seed. 

^*= Todas las correlaciones marcadas son significativas p≤ 0.05; APL= altura de planta; NR1= número de ramas primarias; NR2= número de ramas secundarias; NR= número de racimos; NF= número de frutos; PS= peso de semillas; CA= contenido de aceite.

Regarding to propagation by stakes, there were significant differences only for the number of secondary branches (p= 0.003). According to Tukey’s test of means (p≤ 0.05), I-08 access excelled with an average of 41.3 secondary branches (Table 4). In contrast, the I-31 access showed the least number of secondary branches (12.7). ^{Sonnenholzner (2008)} mentions that stake-propagated plants have a poor root system in relation to foliage, which causes plant stress and slows development in the initial stage.

Table 4 Morphological characteristics of 16 accesses of J. curcas propagated by stake, coming from the state of Veracruz. 

Medias Tukey p≤ 0.05. ACC= acceso; APL= altura de planta; NR1= número de ramas primarias; NR2= número de ramas secundarias; NR= número de racimos; NF= número de frutos; PS= peso de semilla.

In this regard, ^{Laviola et al. (2009)}, mention that the number of secondary branches is a more interesting production component for a breeding program, since the inflorescences occur in the terminal sprouts of the branches, and also the production of fruits will depend on the greater number of branches.

The oil content in plants propagated by stake did not present statistical differences in the accesses evaluated (Figure 2). The oil content was in a range of 50.29% to 59.38%, which is much higher than was reported by ^{Machado (2011)}, who found in Cuba contents of 32.8% to 35%, in plants propagated by stakes, of one year old from SSCE-10 and Cabo Verde.

Figure 2 Oil content (%), in accesses of non-toxic J. curcas in 3 year old plants propagated by stake.  

The correlation between the morpho-productive variables and the oil content, for plants propagated by stake (Table 5). A significant correlation was found between height and number of fruits (r= 0.56). The number of clusters had a very high correlation with the number of fruits (r= 0.94) and seed weight (r= 0.67) and finally the number of fruits correlated with seed weight (r= 0.69). No significant correlation was found between morpho-productive variables and oil content, possibly due to the influence of plant size and low seed production. ^{Avelar (2005)} indicates that the productivity of the plants is related to the number of fruits, number of seeds and weight of fruits.

Table 5 Correlation analysis between morpho-productive variables in plants of J. curcas L., propagated by stake. 

^*= Todas las correlaciones marcadas son significativas p≤ 0.05. APL= altura de planta; NR1= número de ramas primarias; NR2= número de ramas secundarias; NR= número de racimos; NF= número de frutos; PS= peso de semilla; CA= contenido de aceite.

Based on the average seed yield per plant, planting density (1 666 plants ha^-1) and oil content, seed production and liters of oil per hectare were estimated. For seed-propagated plants, the maximum yield was the I-34, under temporary conditions and without inputs and was 1 361.12 kg ha^-1, if the average oil content of this access is 55.96%, then the amount of oil obtainable is approximately 761.68 L ha^-1, for the I-32 access the production was 1 067.91 kg ha^-1, with an average oil content of 55.3%, the amount of oil to be obtained is 590 L ha^-1, the other accesses showed smaller productions (Table 6).

Table 6 Estimated yield of seed and oil per hectare in threeyear- old plants of J. curcas propagated by seed. 

ACC= acceso; ALP= altura de planta; PS= peso de semillas; REN= rendimiento; CA= contenido de aceite; PA= producción de aceite.

These seed yields are within the range reported by ^{Jones-Miller (1992)}, ^{Openshaw (2000)} and ^{Parsons (2005)}, who report yields of 0.4 to 12 t ha^-1 in five-year plants. On the other hand Zamarripa (2011), mentions that the first three harvests must be considered to determine the best genotype.

In plants propagated by seed, better results are obtained in the number of fruits, seed weight and oil contents, because in the propagation by cuttings these yields are inferior (^{Octagon, 2006}). However, stake propagation is faster, cheaper, and easier to establish, and it retains the same genetic information as the parent plant (^{Sonnenholzner, 2008}).

In the stake propagated plants the maximum production of seed was obtained in the access I-27 with 548.114 kg ha^-1, with a 51.22% oil content, an oil production of 280.74 L ha^-1 was estimated (Table 7). It should be noted that the outstanding accesses in propagation by seed, are not those of greater production in propagation by stake and vice versa.

Table 7 Estimated yield of seed and oil per hectare in threeyear- old plants of J. curcas, propagated by stake. 

ACC= acceso; APL= altura de planta; PS= peso de semillas; REN= rendimiento; CA= contenido de aceite; PA= producción de aceite.

^{Machado (2011)}, in a study with plants propagated by stake of one year of age, planted at a density of 2 500 plants ha^-1, reports a yield of 325.6 g plant^-1, for Cabo Verde origin, with which estimates a yield of 814 kg ha^-1, much higher than that obtained in this paper for propagation by stake, but lower than the production of seed propagated seedlings, even with the difference in density.

The importance of studying the non-toxic materials of J. curcas lies in its possible commercial use in human food, regardless of whether it is used for the production of biodiesel. Makkar et al. (1997) and ^{Martínez-Herrera et al. (2012)}, indicate the nutritional potential that J. curcas can have for humans and animals, for the quality of its oil, lipid content and protein value. The oil is composed of oleic acid (41.5-48.8%), linoleic (34.6-44.4%), palmitic (10.5-13%) and stearic acid (2.3-2.8%), according to ^{Martínez et al. (2006)}. Another important aspect of the morphological and productive characterization of J. curcas is the identification of promising material with potential for commercial cultivation or to integrate into a breeding program (^{Becker and Makkar, 2008}).

Finally, it should be pointed out that the desired elite genotypes must have a high seed yield, high oil and protein content, synchronization in fruit maturity, low plant size, tolerance to pests and diseases and to grow well even in drought conditions and low soil fertility.

Conclusions

The accesses evaluated had high values in oil content, the plants propagated by seed ranged from 54.13 to 60.98%, and those originated by stake from 51.75 to 58.48%.

Significant and positive correlations were found between the morpho-productive variables and the oil content in seed-propagated plants. The number of fruits and oil content (r= 0.74) and seed weight and oil content (r= 0.79).

The estimated yield of oil in liters per hectare, based on the number of fruits (157) (184) and seed weight (641.1 g plant) (817.8 g plant) and oil content, I-32 and I-34 were outstanding with 590.55 and 761.68 L ha^-1, respectively, in seed-propagated plants.