Introduction

Several countries in the world have supported the use of biofuels producing species in order to reduce emissions of greenhouse gases (^{Zamarripa et al., 2012}). One alternative is the higuerilla (Ricinus communis L.), oilseed species that is widely distributed in Mexico, and has high potential for seed production to obtain oil (^{Martinez et al., 2012}). More than 95% of the production of higuerilla in the world is concentrated in India, China and Brazil (^{Sailaja et al., 2008}). India 840 000 ha^-1, harvested; China 210 000 ha^-1; Brazil 159 205 ha^-1 with a grain yield 1.09, 0.190 and 0.09 million t respectively (^{FAOSTAT, 2011}). The higuerilla crop development can cause a chain of new businesses, technologies and products (^{Mazzani, 1983}, ^{Brigham, 1993}; ^{Stachetti et al., 2007}). Several authors mention that the progressive development of aviation fuels and the use of high-revving engines generate high demand for higuerilla oil, this creates a need for technical and scientific information on this crop in various ecological niches (^{Tobar, 1981}; ^{Scholz and Da Silva, 2007}; ^{Sepúlveda, 2012}).

The higuerilla is native to Africa, being in the Republic of Yemen to North Africa, and the Near and Middle East (^{Purseglove, 1974}, ^{Vavilov, 1992}). It is a plant annual or perennial habit according to environmental conditions (^{Vibrans, 2009}), plant high canopy, sometimes a little shrub, from light to gray-green color, sometimes reddish, with erect stems up to 6 m. (^{Rzedowski and Rzedowski, 2001}). The seed is oval, spherical or elongated rarely, from 0.8 to 3 cm long, 0.6 to 1.5 cm wide and 0.4 to 1.0 cm thick (^{Moshkin, 1986}).

Variable presented with reddish brownish stains and has a hard and brittle outer shell and a very fine whitish underside. Both protect the seed, which consists of a tiny embryo with two cotyledons thin and soft, compact and oily albumen (^{Weiss, 1983}; ^{Rzedowski and Rzedowski, 2001}). The seed contains toxins such as ricin (which is an albumin belonging to the family of ribosome inactivating protein, they stop protein synthesis, causing cell death by apoptosis) and ricinina very toxic alkaloid that attack the liver and pancreas (^{Rzedowski and Rzedowski, 2001}; ^{Fanan et al., 2009}). Higuerilla oil meets physicochemical characteristics that position it as an option on the production of biofuels (^{Martinez et al., 2012}). Besides its use in the manufacture of surfactants, coatings, greases, fungistatic, pharmaceuticals, cosmetics and many other products (^{Lima et al., 2013}).

Studies on yield components provide guidance for optimal production. These are interdependent and change in response to environmental conditions. Negative correlations between yield components are very common (^{Kumar et al., 1997}; ^{Rajala et al., 2009}; ^{Soratto et al., 2011}, ^{Sadras and Slafer, 2012}). Some authors point out that for the selection of highly productive variants higuerilla emphasis should be placed on the number of clusters and fruit weight; as well as the length of the bunches and seed weight (^{Sarwar and Boota, 2008}). In some studies, the number of seeds per cluster breaks down the number of fruits per cluster and the number of seeds per fruit; however, these two components are preferably analyzed together because there is negligible variation in the number of seeds per fruit (^{Fanan et al., 2009}; ^{Machado et al., 2009}). Individual seed weight plays an important role (^{Severino, 2012}).

Seed weight is determined by the genetic potential of the plant (^{Egli et al., 1987}), but the variability in seed weight may be caused by many physiological processes that are sensitive to environmental changes. In cross-pollinated species, the origin of diploid embryo and endosperm triploid nature is a potential variability in seed weight (^{Hay et al., 2010}) source. Given the above, the objective of this work was to study the variability and adaptation of a group of varieties of higuerilla in Montecillo, State of Mexico, from April to December 2013.

Materials and methods

The experiment was conducted in the period from april to december 2013 in the experimental field of the Graduate College, located in Montecillo, Texcoco, State of Mexico, with geographic coordinates 19° 28' 02.9'' north latitude and 98° 54' 02.3'' west longitude and an altitude of 2 240 meters. The climate according to Koppen climate classification modified by ^{García (2005)} is C (Wo) (W) b (i'), corresponding to temperate humid with summer rains. The average annual rainfall and temperature are 625 mm and 16 °C respectively. Precipitation of the year into the experiment in april and lasted to mid-october with a bimodal regime with intraestival drought (SI) in the month of august and the great winter drought (GSI) in mid-october to mid-march. The cumulative rainfall during the crop growing season was 639 mm (Figure 1). frost early december to mid-february (Figure 2) were presented. The data of meteorological variables were obtained from the stations in the Graduate College and the University of Chapingo.

Figure 1 Rainfall during the growing season of the crop in Montecillo, cycle from april to december 2013. 

Figure 2 Minimum daily at the end of the crop cycle in Montecillo, cycle from april to december 2013 temperatures. 

The soil of the Experimental Station (Table 1) was analyzed, it is a sandy loam, moderately basic and slightly saline soil, with percentages of organic matter (M.O.) of 2-3%

Was evaluated 11 varieties of higuerilla: 10 improved and CP13 collection "unimproved" (Table 2). The INIFAP varieties were selected according to previous results obtained by ^{Zamarripa et al. (2010)}. seed collected was used in 2013 in the Graduate College in Montecillo in unimproved. The experiment was established on April 6, 2013. The experimental plot was 6 meters long and 1.5 meters wide (9 m2), with five plants per row, in square 1.5 x 1.5 m. The population density was 4 444 plants ha^-1. The five plants formed the useful plot. design used randomized block with four replications. In those variables that readings on several dates were conducted it generated a factorial design with two factors, varieties and readings under a split plot arrangement where readings were for large plots and subplots varieties. The model is as follows:

Where: Y= total variation of the response variable; Rep= variation of repetitions; Lec= change in lectures; error= variation associated with large plots; Var= variation of the varieties; Var x= interaction Lec varieties readings; Error b= variation associated with subplots.

The variables evaluated with more than one reading date were: plant height (AP) measured from the ground surface to the upper end, stem diameter (DT) 40 cm above the ground, leaf number (NH) counted from the base of the stem to the high end, number of branches (NR) of the base of the stem to the first fork, plant canopy area (ABS) for this width (ADS) and took over (LDS) canopy.

The variables are recorded once a day to flowering (DF), number of clusters per plant (NRP), cluster length (LR), number of fruits per bunch (NF), equatorial diameter (OD) and polar (DP) of fruit and fruit weight (PF). Likewise characteristics were measured in seed: length (LS), width (AS), thickness (GS), volume (VL), length / width ratio (RLA), 100 seed weight (PCS) and seed production (REN). The model for the analysis of these variables experiment corresponded to a simple blocks, and error varieties as variation factors as described immediately.

Where: Y= total variation of the response variable; rep= variation of repetitions; var= change in varieties; error = variation associated with the error.

The analyzes were performed using the statistical program SAS 9.0 (SAS Insitute Inc., Cary, NC, USA) and the averages compared by Duncan test, with a significance level of 5%. To develop a selection index (IS), which involve all response variables were assigned a "1" for each "a" in the line of each variety. The sum of "1" was the IS. Thus the index integrates the adaptive effect of all the features that the variety showed good response in the environment of Montecillo. Finally stockings varieties of all variables in the study were subjected to multivariate analysis of main components.

Results and discussion

Variation of the characteristics of seed and selection index between varieties

The varieties 5(Riric10), 6 (Riric19) and 10 (Riric1) were the best selection index values 6 (Table 8). The results for the variable weight of 100 seeds show an amplitude in the means obtained for varieties which gives an idea of the variability between these grown materials. This same variability was observed in Morelos by ^{Barrios et al. (2013)}. Likewise, ^{Mazzani (1983)} reports similar results to the research in Maracay, Venezuela found 100 seed weight between 19 and 91.8 g.

Table 8 Comparison of means of the characteristics of seeds and IS higuerilla varieties in Montecillo, cycle from april to december 2013. 

Variedades	LS (cm)	AS (cm)	GS (cm)	VL (ml)	RLA	PCS(g)	REN (g ρ1)	IS
1	1.53 e	1.21 ed	0.73 d	0.60 ef	1.27 cd	60.28 cd	1102 a	1
2	1.86 ab	1.50 abe	0.82 be	0.93 ab	1.24 def	88.22 a	663 b	4
3	1 81 bc	1.50 bc	0 83 be	0.85 ¿c	1.20f	83.96 ab	585 b	2
4	1 74 cd	1.44 c	0.81 с	0.78 cd	1 21 f	73 25 bc	868 ab	1
5	1.95 a	1 58 a	0.85 ab	0.93 ab	1.23 ef	93.88 a	771 ab	6
6	1.96 a	1.57 ab	0.86 ab	0.95 a	1.25 ede	93 25 a	794 ab	6
7	1.69 d	1.26 d	0.74 d	0.67 de	1.34b	67.88 cd	801 ab	1
8	1.77 bcd	1.44 c	0.80 c	0.82bc	1.23 ef	82 38 ab	1106 a	2
9	1.36 f	1 16 e	071 d	0.53 f	1.17 g	54.80 d	914 ab	1
10	1.96 a	1.54 ab	0.87 a	0.91 ab	1.27 c	88 52 a	767 ab	6
11	1.21 g	0.76 f	0.54 e	0.21g	1.60 a	22.85 e	297 b	1
Media	1.71	1.36	0.78	0.74	1.27	73.57	787.85
+Frec-pc	4	4	3	5	1	6	8

TLS= longitud de semilla (cm); AS= ancho de semilla (cm); GS= grosor de semilla (cm); VL= volumen de semilla (ml); RLA= relación largo- ancho; PCS= peso de 100 semillas (g); REN= rendimiento de semillas (g p^-1); IS= índice de selección o frecuencia de “aes” por variedad, ⁺frecuencia de “aes” por característica, Promedios con la misma letra dentro de columnas son estadísticamente iguales según la prueba de Duncan (p≤ 0.05).

The varieties had the highest frequencies in the selection indices: V6 (Riric19), V2 (Riric 265-1) and V10 (Riric1) conglomerated in the ACP group III with the difference that in the main components variables are confused and indices are explicit (Figure 3).

Figure 3 Distribution of higuerilla varieties in Montecillo, cycle april- december 2013: v1, v2, v3, v4, v5, v6, v7, v8, v9, v10 and v11. 

The variables that were correlated with REN: AP (r= 0.82, p< 0.01), NR (r= 0.70, p< 0.05), LDS (r= 0.74, p< 0.05) and NRP(r= 0.79, p< 0.01), other variables were also correlated positively. NRP/PCS, DP/PCS, DE/PCS, PF/ PCS and LS; AS and GS with PCS. negative correlations were found days to flowering; DF/NRP (r= -0.69, p< 0.05); DF/FR (r= 0.84, p< 0.01); DF/PF (r= -0.50, p< 0.05); DF/ PCS (r= -0.69, p< 0.05); DF/REN (r= -0.54, p< 0.05), LR/PF (r= -0.81, p< 0.01), LR/PCS (r= -0.84, p< 0.01), FR/PF (r= -0.83, p< 0.01) and FR/PCS (r= -0.91, p< 0.01);.

Conclusions

The variability was found in higuerilla varieties in terms of the morphological characteristics of the plant as well as features of clusters, especially fruits and seed size. This allowed us to differentiate groups which had different production potential, beating the unimproved variety CP13 used as a witness. Comparing with improved unimproved varieties included in this study, it follows that DT, NR, NH, LDS, bass and precocity (DF) flies are the characteristics of more emphasis on improving. The longer showed improved cluster and number of fruits per cluster that can be added to improved by breeding schemes. According to the best selection indices varieties V2 (RIRIC 265-1) Barranca Honda, Michoacan, V6 (Riric19) of Villa Corzo and V10 (Riric1) of Cacahoatán, Chiapas were the best adapted to the conditions of planting April Montecillo. However, the severe frost, seed production which suggests the possibility of generating improved material adapted to this altitudinal level was achieved. The method of selection indexes used was efficient to clarify the best varieties and clearly informs the characteristics that determine the index allowing the improvement and better target agreed with the groups determined by major components.