Soils

Interchangeable bases within the soil (Ca⁺², Mg⁺² and Na⁺) showed very high contents (table 1), and sufficient K+. In tropical dry forests, high values in these bases, such as Na⁺, are promoted by annual precipitations that are insufficient to satisfy potential evapotranspiration.

Table 1 Physical and chemical characteristics of the soil in the Chochó district (tropical dry forest) (2015) 

Methods for texture analysis: Bouyoucos; pH: potentiometer, soil-water ratio 1:25; OM: organic matter, modified Walkley-Black; P: Bray II; S and B: monocalcium phosphate; Al: extraction with 1N KCl; exchangable bases: ammonium acetate; microelements: modified Olsen; som: soil organic matter; Dap: apparent soil density; CE: electrical ground conductivity; CEC: soil cation exchange capacity.

Source: Elaborated by the authors

Climate

Climatic variables were recorded using a meteoro logical station located 15 m from the cultivation plot. Environmental variables were quantified as precipitation (resolution of0.2 mm) and maximum and minimum temperatures (resolution of 1 °C).

As described in figure 1, maximum temperatures (T °C max) showed a higher variability than the minimum ones, since the latter fluctuated between 22 and 25 °C, with an average of 24.5 °C, while the first ones varied from 31 to 37 °C, with an average of 35 °C. Moreover, diurnal maximum temperature stabilized at 35 °C from 51 DAP (days after planting), and reached 37.3 °C at 87 DAP; likewise, the night minimum temperature (21 °C) were observed at 30 and 33 DAP and stabilized at 22 °C after 60 DAP.

Source: Elaborated by the authors

Figure 1 Evolution of maximum and minimum daily temperatures in the district of Chochó (Sincelejo, Colombia). Data recorded between October 1 and December 30, 2015. 

Maximum precipitations (greater than 5 mm) were concentrated between 13 and 47 days after crop establishment (figure 2), with an interval of five days between these; in addition, in this period maximum daily precipitation (34 mm) was showed after 30 days. Subsequently, between 48 and 71 days, maximum rainfall decreased drastically.

Source: Elaborated by the authors

Figure 2 Daily and accumulated rainfall in Chochó district (Sincelejo, Colombia). Data recorded between October 1 and December 30, 2015. 

Plant material

Sesame cultivars that were assessed in this study were introduced in Colombia in the fifties, when the former Instituto de Fomento Algodonero (IFA) was responsible for their promotion and dispersion, distributing various sesame cultivars in the eastern plains and in the Caribbean cost of Colombia (^{Caicedo et al., 1998}). Among these, we find the following cultivars, which are the most common ones found in the study area:

Planting density was 80,000 plants/ha, distributed in rows at 0.5 m, with 0.25 m between plants. Five to seven seeds were planted per site, from which only three to six of these germinated. Thinning was carried out ten days after planting, leaving only one plant per site. Fifteen experimental plots of 5 m long and 4 m wide were arranged in the field, each of which consisted of nine rows of 25 plants.

Experimental design

The three sesame cultivars described above were used in the treatments considered. Randomization was performed based on an experimental design in randomized complete blocks with five replicates, in which the blocking factor was the terrain slope.

Evaluation of variables

During the physiological maturity of the crop (90 days after planting) the following measure ments were made:

In relation to seed yield per hectare (SYH), this measure was established from the average seed weight per plant, found when harvesting ten central plants per plot, and multiplied by the number of plants within the hectare (80,000).

Likewise, oil content (OC) in the seed was obtained through solid-liquid extraction with a crushed seed mass (5 g) through Soxhlet, with ethanol as solvent (100 ml) for eight hours (^{Carvalho, Galvão, Barros, Conceição, & Sousa, 2012}).

Statistical analysis

The data was analyzed based on the experimental design section using the SAS 9.3 statistical program. To establish the magnitude of the variation between cultivars, comparison of means was made with the Tukey test, with a level of significance of 5 %.

Yield component analysis was performed for each sesame cultivar by a simple linear regression between seed yield per plant (R) and each of the performance components described above (R = β1 χ CP + β0, where CP is the specific yield component, and β1 and β0 are the regression coefficients of the model or the slope of the regression line and the intercept, respectively), considering in each of them the significance of the linear model represented by the p-value, which evaluates the hypothesis β1 ≠ 0 and the R ² value, which indicates the percentage of the variation of the R explained by the model.

Number of branches, fertile nodes in stem and average fertile nodes in branches

According to the analysis of variances there was a significant effect of the cultivar on the number of fertile nodes in the main stem (FNMS) (p < 0.05), while the average fertile nodes per branch (AFNB) and the number of branches (NB) did not exhibit this effect. It should be noted that in the Chino Rojo cultivar an average number of six branches was observed, and in the others, it was close to four.

Number of branches (NB) is a varietal parameter that is intimately related to genetic and environ mental aspects (^{Langham, 2008}, ^{Peter, 2006}). In addition, in sesame, it is used to categorize the genotype. In this study, for example, the cultivars used fit into the category offew ramifications in the canopy (as described by Peter [2006]), since this parameter did not exceed three pairs of branches, concentrating a high number of fertile nodes, compared to the stem.

Likewise, the Average number of fertile nodes per branch (ANFNB) of the cultivars Criollo and ICA Matoso was sixteen, while Chino Rojo showed a lower value close to thirteen. Criollo and ICA Matoso cultivars had the highest number of nodes, with ca. 29 productive nodes in the stem, and there was no significant difference between them (table 2); on the contrary, Chino Rojo exhibited a significantly lower number of fertile nodes of ca. 31 % compared to the other cultivars.

In general, Criollo and ICA Matoso cultivars had a lower number of fertile nodes compared to Chino Rojo, and because it was more branched, it had more productive nodes in the stem and branch set. However, the first two showed a greater number of capsules per fertile node (table 2), a strategy that allows plants to compensate for their lower number of fertile nodes.

Table 2 Physiological attributes of yield in sesame cultivars (Sesamum indicum L.), in a locality of Sucre (Colombia) (2015) 

FNS: fertile nodes in the stem; CN: capsules per node; SF: seeds per fruit; ASW: average seed weight; NR: number of branches; PNFR: average of fertile nodes in branch; Y: seed yield per hectare. Averages with the same letter did not show significant differences according to the Tukey test at 5 %.

Source: elaborated by the authors.

Capsules per fertile node, seeds per capsule and weight of one thousand seeds

The analysis of variance for the number of capsules per node (CN), seeds per capsule (SC) and the average weight of one thousand seeds (AW1000S) found a significant effect of the cultivar factor in each of these variables (p < 0.05), which indicates that at least two cultivars differ in function of these.

The number of capsules per fertile nodes showed higher values in the cultivars Criollo and ICA Matoso, with an average of five capsules (table 2), meanwhile Chino Rojo showed significantly smaller quantities, with a difference of 25 % compared to the others. In sesame, it is common to find between four and six capsules per fertile node, that is, from two to three capsules (fruits) per axilla (^{Langham, 2008}), while the number per fertile nodes can vary from two to eight (^{Morris, 2009}).

Moreover, there was no significant difference in the number of seeds per capsule (SC) among the cultivars studied (table 2), and the general average was close to 68 seeds per fruit, an amount similar to what was published by ^{Sheahan (2014)} , who reports 70; however, these results are very different from the one found by ^{Morris (2009)} , who found between 80 and 131 in the evaluation of 192 sesame accessions from 38 countries. The above mentioned shows the wide variability of this parameter, in response to the interaction of genetics with the environment, that in relation to our study, did not show many changes in the cultivars studied.

The average seed weight (ASW) did not differ significantly between Criollo and ICA Matoso, with a value close to 3 g, but it significantly exceeded the value found in Chino Rojo by 26 %. However, according to the above mentioned, the latter had a greater number of fertile nodes and branches in plant, which would indicate a high competition for assimilated between stem and branch capsules, if we consider that in the other cultivars there were fewer ramifications and fertile nodes per plant.

In the evaluation of 17 sesame cultivars, ^{Pham et al. (2010)} found that values of weight of one hundred seeds ranged between 0.25 to 0.45 g, while in another study, considering five sesame cultivars, weights of one thousand seeds ranged between 2.4 and 3.9 g (^{Olowe & Adeoniregun, 2010}). Furthermore, ASW found in this study corresponds to what was indicated by these authors.

Yield and oil content

There was a significant effect of the sesame cultivar on seed yield and oil content (p < 0.05), which shows differences in these two aspects in at least two cultivars.

Yield was significantly higher in Criollo with about 1,290 kg/ha (table 2) followed by ICA Matoso, with ca. 95 kg less seeds than the first, while Chino Rojo had the lowest yield, that was significantly lower in 42 % and 32 %, in relation to Criollo and ICA Matoso, respectively.

Oil content in sesame varies according to environ mental and genetic factors (^{Peter, 2006}), which was evident in this study, in which statistically, cultivars ICA Matoso and Criollo produced the same content that was higher than in Chino Rojo (table 2). Regarding percentages, ICA Matoso contained around 5 % more oil in seeds than Criollo and Chino Rojo, which did not differ significantly, as did Criollo and ICA Matoso. This suggests some correspondence that was also found in seed production.

In studies carried out on eight genotypes of sesame under water stress conditions (^{Golestani & Pakniyat, 2015}), seed yields were 879.8 kg/ha, while in non-stressed plants, these reached average values of 1,404 kg/ha. This indicates a decrease in plant productivity when faced with water deficit, which could have occurred in this study; this because during the development stage of capsules, when rainfall decreased, environmental temperatures increased, and the study soil showed high content of interchangeable bases (table 1). All this favors water deficit in these environments (^{Gasca, Menjivar, & Trujillo, 2011}), which would affect Chino Rojo more than the other cultivars.

In addition, there is a correspondence between higher yields in those cultivars that had higher average seed weights (ASW) (table 2), fertile nodes in the main stem (FNMS) (table 2) and capsules per fertile node (CFN), despite the fact that no effect of the cultivars on the number of branches (NB) or differences in the amount of seeds per capsule (SC) between them was found.

This indicates that these last variables do not influence seed yield too much, which allowed adequate yields to be produced in Criollo and ICA Matoso, given the positive association between these variables and seed yield per plant (^{Ismaila & Usman, 2012}).

Normally, the oil content in sesame varies between 37 % and 63 % (^{Peter, 2006}). Seeds can be qualified as premium if they have an approximate oil content of 50 % (^{Bennett, Imrie, Raymond, & Wood, 1998}); this implies that the ICA Matoso and Criollo cultivars could have this qualification (table 2).

Regarding the environment and according to studies carried out by ^{Were, Onkware, Gudu, Welander and Carlsson (2006)} , oil content is influenced by the amount of rainfall received during cultivation, especially in the capsule development stage. It should be noted that in this study rainfall decreased 70 days after planting.

On the other hand, oil content is associated with seed color. Dark seeds contain less oil than light ones (^{Akinoso, Aboaba, & Olayanju, 2010}, ^{Peter, 2006}). This could explain the particular case of the cultivar Chino Rojo, which has brown seeds with a low oil content, in comparison with other cultivars, in which light colors predominate.

Yield component analysis

Seed yield per plant (R) showed a significant association (p = 0.001) with the number of fertile nodes in the main stem (FNS) in Chino Rojo (table 3), while in the other two cultivars, this type of relationship was not observed. In this cultivar, for each FNS, there is an increase close to 0.08 g in R, which contrasts with the low amount of FNS found, which indicates that in this plant it is important to favor the development of fertile nodes, because an adequate emission of these in the stem is associated with increases in production.

Criollo cultivar showed a linear association ( p = 0.013) between R and the number of seeds per capsule (SC) (table 3), which makes it possible to explain 38 % of the variations in R compared to those in seeds per fruit (SF), with an increase close to 0.26 g for each SC of this cultivar.

Table 3 Linear regression models between seed yield per plant and yield components in sesame cultivars, in a locality of Sucre (Colombia) during 2015 

^*** Significance less than 1 %; ^** Significance less than 5 %; ^* Significance less than 10 %. FNS: fertile nodes per stem; CN: capsules per node; SF: seeds per fruit; AW1000S: average weight of one thousand seeds; NB: number of branches; ANFNB: average number of fertile nodes per branch; R: seed yield per plant.

Source: elaborated by the authors

Considering all the cultivars, no significant relation ship was found between the number of branches and the seed yield (table 3), which would indicate a greater influence of the stem capsules in R, especially in Chino Rojo, in which the increases of R, are associated in a greater degree (R ² = 0.54) in comparison to those of compared to those of FNS.

Moreover, table 3 shows the non-existence of a significant association (p > 0.05) between the average of fertile nodes in branch (AFNB) and R. However, in Chino Rojo, p = 0.056 was obtained, suggesting a slight association between these variables, because of the greater number of ramifications in this crop.

In all cultivars, the average weight of one thousand seeds (AW1000S) exhibited a linear association with R (table 3). In ICA Matoso and Chino Rojo the model manages to explain the variability of R, from the variations in the AW1000S, in a percentage higher than 70 %, which also indicates that for each increase per unit gram in the weight of a thousand seeds, R increases between 46.00 and 47.00 g in ICA Matoso, and 18.45 g and 7.60 g in Chino Rojo and Criollo, respectively. Furthermore, this indicates that, in the first cultivars, plant yield is more dependent of an adequate accumulation of biomass in seeds, an aspect that is subject, to a large extent, on the relationship between the genotype and the environment.

In general terms, the component with the greatest influence on seed production in sesame seeds is AW1000S, which on average increases seed yield by 24 g per gram of weight in a thousand seeds. On the other hand, elements such as FNS and SC showed influence on particular cultivars: the first one is associated with yield per plant in Chino Rojo, and the second one with those of Criollo.

Similar studies on yield components in sesame (^{Adebisi, Ajala, Ojo, & Salau, 2005}, ^{Ismaila & Usman, 2012}) show adequate associations of AW1000S and SC with seed yield, so these elements could be considered primary in seed production.

Similarly, ^{Morris (2009)} found a greater influence of the number and weight components of seeds in sesame production and plant height, which is closely related to the number of fertile nodes in the stem.