Yield and physical-chemical characteristics of the grain

The yield of the dry weight of grain (RG) was calculated (t ha^-1), adjusted to 14% of humidity and extrapolated by one hectare. In the laboratory, a flotation index (IF, indirect measure of hardness) (SE, 2002), weight of 100 grains (PCG) (indirect measure of size) and hectoliter weight (PH) (^{Salinas and Vazquez, 2006}) were measured. Oil (ACE), protein (PRO) and starch (ALM) were quantified according to methods 30-25.01, 46-16.01 and 76-13.01 of ^{AACCI (2000)}.

Nixtamalization and tortilla making

The nixtamalization process consisted of cooking a 500 g sample of corn, with 3.5 g of Ca(OH)₂ and 1 L of water. The cooking time was assigned according to its hardness: IF= 0-12 (Very hard), 45 min; 13-37 (Hard), 40 min; 38-62 (Medium), 35 min; 63-87 (Soft), 30 min; 87-100 (Very soft), 25 min (SE, 2002). The nixtamal rested in the cooking solution (nejayote) for 18 h, and was subsequently washed with water (0.5 L). In the nejayote, loss of dry matter (PMS) was quantified. In a Fudough^® Mod. MN-80 Universal stone mill (Fudough S.A., Mex.), the nixtamal was ground with water (5% volume/weight), to obtain the dough (^{Vázquez et al., 2012}). The dough was conditioned until a 10 g ±0.02 sphere compressed for 5 s with a 2.5 kg capacity scale reached a diameter of 5 cm, that is, 58% ±0.05 humidity. The dough was die-cut to obtain tortillas 12 cm in diameter and 1.25 mm thick, which were cooked for 1.5 min in a Villamex^® Mod. V-14 C/R tortilla maker (Villamex, Mex.) with 3 comales (average temperature = 265 °C).

Once the tortillas were cold (25 °C), they were packed in hermetically sealed polyethylene bags. With the registered weights of nixtamal, conditioned dough and cold tortillas, the yields were calculated. In the freshly made tortillas (30 min) (RE), of 24 (HT24) and 48 (HT48) hours of storage at 4 °C, humidity (^{AACC, 2000}), breaking force (FZA) and maximum elongation (EL) with a Brookfield^® CT3 (Middleboro, EU) texturometer (^{Vázquez et al., 2015}) at a feed rate of 1 mm s^-1. With FZA and EL, Young’s modulus of elasticity (Y) (^{Mao et al., 2002}) was calculated with the equation Y= E/D, where the stress E= FZA*cross-sectional area^-1 of the tortilla disk; D= diameter of the tortilla used for evaluation*EL^-1.

Statistical analysis

An analysis of variance (Anova) was performed, where HIB= fixed effect, AMB= random and the HIB × AMB interaction; tests (p< 0.05) of comparison of means (Tukey) and correlation analysis (Pearson). The variables with significant interaction (HIB × AMB), the SREG biplot was obtained in which the main effects of hybrid (CP1) and those of interaction (CP2) were plotted (^{Vázquez et al., 2016}; ^{Crossa et al., 2015}; ^{Lozano et al., 2015}) with the use of the statistical package Statistical Analysis System (SAS) version 9.0 (^{SAS Inst., 2002}).

Physical characteristics

The IF was explained in 78.8% by the first two CP. Figure 2B shows in ascending order, from left to right, the number of floating grains; that is, inverse of hardness. According to the hardness classification by the IF (SE, 2002), the hardness of the hybrids was from intermediate to hard (IF). The dough and tortilla industry (IMT) prefer corn with IF ≤ 40 and that of nixtamalized flour (IHN) with less than 20 floating grains (^{Gaytan et al., 2013}). The preference grows to the left of the biplot, with lower IF (Figure 1B). The hybrids evaluated presented IF between 15-50, which are suitable for the dough and tortilla industry IMT, absorb more water during nixtamalization and rest, the above, is related to a higher yield of dough and tortilla (^{Vázquez et al., 2015}).

Within the group of hard grains are Tsiri Puma and Albatross (Table 1), the rest correspond to grains of intermediate hardness. Hard grain hybrids (Tsiri Puma and Albatros) are associated with higher yields, which coincides with what was reported by ^{Vázquez et al. (2012}), that the harder the grain, the higher its RG, PH and lower IF. These hybrids can be recommended for the nixtamalized flours industry (IHN), as they achieve less hydration during cooking, facilitating their grinding (^{Salinas et al., 2012}).

The hybrid H-51 AE corresponds to the one with the highest stability in the IF (Figure 1B) considering the criteria mentioned by ^{Crossa et al. (2013}) in the PC and in Jocotitlan presented the softest grains. The hybrids Albatros and Tsiri Puma responded positively to adaptability in ATL and JOC, while the hybrid H-49 AE in TEM.

In the IF biplot it is observed that in TEM and ATL there was greater variability between the HIB, so the results obtained in the two environments are more reliable in relation to the other production environments, although they present a greater effect on the IGA than in JOC, where they appear near the X axis. ^{Yan et al. (2007}) explain that, the greater the relative dispersion of the AMB, with respect to the horizontal axis, the representativeness of this in the analysis is reduced, since there is a greater effect of IGA in the variable. JIL and IXT did not discriminate between hybrids, given their proximity to the origin (shorter vector). ^{Yan et al. (2007)} point out that environments that present an angle less than 90° classify hybrids in a similar way, a similar case is shown in these two environments (Figure 1B).

The PH had a behavior similar to IF, registering the best response in the hybrids Atziri Puma and Albatros (79.3 kg hL^-1), followed by the hybrid H-49 AE (79.1 kg hL^-1). The hybrids that presented the lowest PH were H-47 AE, Atziri Puma and H-66 (Table 1). In all the hybrids evaluated, a value higher than that required by the NMX-032/1 (^{SE, 2002}) of 74 kg hL^-1 was found. These results exceed those reported by ^{Vázquez et al. (2016}) when evaluating the stability of the yield and the physical variables of the grain, as well as for nixtamal and tortilla of eleven corn hybrids in six environments in the central region of High Valleys of México.

In the grain size, represented by the weight of one hundred grains (PCG) is a common variable in the choice of grains to process. The first two PC were significant and represented 81.83% of the variability of the data (Figure 1C). The standard (^{NMX-034/1 2002}) did not consider the size of the grain; however, it is a variable of interest that impacts cooking and water absorption during nixtamalization (^{Antuna et al., 2008}).

Considering the classification of ^{Salinas and Vázquez (2006}), the large grains present PCG > 38 g, the intermediate grains a PCG of 33 to 38 g and the small grains PCG <33 g. The hybrids Atziri Puma in TEM and H-50 in ATL presented the largest grains (Figure 1C). However, according to the selection criteria of Salinas and Aguilar (2010), the tortilla-dough and nixtamalized flour industries prefer small and medium grains. When extending the environmental vectors towards their negative side, a high correlation was found between H-51AE in TEM, as well as between H-47AE in IXT and ATL, which shows a specific adaptation to these environments, with small grains (Figure 1C).

The IHN prefers intermediate-size corn, while IMT prefers intermediate and small grains, due to the better hydration that favors the yield of nixtamal, dough, and tortilla as they present high humidity (^{Salinas and Aguilar, 2010}). Under this premise, seven of the ten hybrids evaluated presented intermediate size, resulting in an alternative with adequate characteristics for processing for IMT industries.

Chemical characteristics

There are significant differences (p< 0.05), for HIB, AMB and interaction between the two factors was detected in the variables evaluated, a correlation was observed between PROT vs ACEI (r= 0.60, p≤ 0.01), ACEI vs ALM (r= -0.86, p≤ 0.01) and PROT vs ALM (r= -0.82, p≤ 0.01), which confirms what was reported by ^{Vázquez et al. (2015}) and it is visible in the biplot, since they are similar or ‘quasi-symmetric’ with each other.

The most important compound of the corn grain is starch, since it represents – parts of the total (^{Paredes et al., 2009}). In this study, ALM percentages (69-70%) were found with a minimal difference between the hybrids. These values are consistent with that reported for serrated type maize (^{Watson, 2003}). The hybrid H-49 AE was better adapted, in relation to the other hybrids and with a higher percentage of starch, followed by H-47 AE, H-53 AE and Albatros, while the Atziri Puma and Tsiri Puma presented values lower than 69.1, respectively (Table 1).

The hybrids stand out in their ACE content, the lowest percentages were 4.88 and 4.90% (Table 1), higher than that observed in hybrids of normal endosperm by ^{Vázquez et al. (2015}). The highest percentages of ACE were found in larger grains (r= 0.43, p≤ 0.01). According to ^{Vazquez et al. (2015)}, are considered hybrids with high oil content (ACA) when they have more than 5% in the grain. Within this classification, the HIB tested in this experiment can be located, which is relevant in the preparation of tortillas by IMT, due to the relationship between the ACE in grain and the softness of the resulting tortillas (^{Vázquez et al., 2016}), a greater amount of it reduces the retrogradation of starches (^{Putseys et al., 2010}).

Therefore, the presence of oil in the corn kernel can significantly improve the firmness and chewing of the tortilla (^{Vidal-Quintanar et al., 2001}). In JOC and ATL, grains with a higher percentage of oil were produced. The hybrids with high percentage of oil in grain were H-66, Atziri Puma, and Tsiri Puma, these last two are stable in TEM. The hybrids with the lowest percentage were H-49 AE and H-47 AE with 4.9% respectively (Figure 1E).

In protein, between 9.75 and 10.31% were obtained (Table 1), the highest percentages in H-53 AE, Tsiri Puma and Atziri Puma (≥10.22%), while the lowest was obtained in the Albatros hybrid. The results are similar to those reported by ^{Vázquez et al. (2012}) who observed from 9.76 to 12.54% protein in local varieties and in improved hybrids from 8.24 to 11.34%. These variations in proteins are associated with the genetics of the materials tested.

Quality nixtamal-tortillas

There were significant differences (p< 0.05) between the hybrids, environments and in IGA for the quality variables of nixtamal, dough and tortillas. The assignment of cooking times according to the IF allowed to reach humidity in the nixtamal between 45.6% (intermediate) and 47.3% (hard), values that in dough increased from 57.4 and 58.1% (Table 2), an increase that is due to, in hard corns, the longer cooking time, and in the dough to the addition of water during grinding and to condition it, prior to stamping.

The highest dough yield was found in H-51 AE, followed by H-47 AE, Albatros and Tsiri Puma with 2 kg of dough for each kg of nixtamalized corn. The yield of nixtamal and dough was similar to that reported by ^{Gaytan et al. (2013}); ^{Vázquez et al. (2015}); Vázquez et al. (2016). While the humidity in freshly made tortillas (30 min) was 43.5 to 45.1% (Table 2). The JOC locality showed the highest diversity in the 30 min HT between hybrids, the hybrids H-51 AE and Atziri Puma were the most stable in the 30 min HT.

The stored HT (24 and 48 h) was slightly higher than the newly prepared, in the hybrids (Table 2). In TEM the hybrids Tsiri Puma, Albatros and H-53 AE presented high percentages of humidity, and the lowest in the hybrid H-50 (Table 2). In all the hybrids, the tortilla yield was higher than the 1.5 kg of tortilla per kg of corn grain demanded by the industrial tortilla-dough (^{Salinas and Aguilar 2010}). By environment, the lowest average yield was observed in Jilotepec with 1.51 kg of tortilla per kg of corn grain. While the H-70 hybrid presented lower values and the lowest in JOC, although it is mostly adapted to JIL.

In the loss of dry matter (PMS), all the materials registered values lower than the 5 g 100 g^-1 declared in the -034(1) standard (SE, 2002) (Table 2). The results coincide with that reported by ^{Vázquez et al. (2015}), where the mean PMS was 3.31 g 100 g^-1, for six hybrids) from High Valleys of México.

Tortilla consumers demand that they be soft and flexible (^{Mao et al., 2002}). Most of the texture variables showed a significant difference (p> 0.01) due to the effect of the genotype, the environment and the IGA, the results shown include the averages of the five locations. The freshly made tortillas showed greater variability, with respect to the stored ones (Figure 1).

At 30 min after making the softest tortillas, they were the hybrids H-47E, H-53AE and H-49AE, which required less force to break (x= 1.4 N), had a good elongation (EL) (x= 8.3 mm) and their modulus of elasticity (Y) was the lowest (x= 11.8 N cm^-2), this quality of the tortillas can also be related to their oil content (± 5%), as reported by ^{Vázquez et al. (2015}). The next group included H-51AE, H-66, H-50 and Tsiri-Puma, whose tortillas required 1.56 N to break, the Y increased (x= 12.4 N cm^-2) in proportion to the increase in FZA, since its EL was equal to that of group 1 (8.3 mm) (Figure 2).

Figure 2 Breaking force and elongation (A) and Young’s modulus of elasticity (B) in freshly made and stored tortillas (24 and 48 h). Averages of the five locations. 

The loss of flexibility of the tortillas upon cooling (24 and 48 h in refrigeration) is due to the formation of a rigid structure caused by the retrogradation of starch and the association with proteins, fiber and other chemical components (^{Agama-Acevedo et al., 2011}). The greatest increase in FZA occurred at 24 h, this was of the order of 0.28 N (18%), while at 48 h of storage in some tortillas the FZA was reduced, in others it remained the same and only in three genotypes H-66, Atsiri-Puma and Tziri-Puma increased the FZA. Similar behavior was observed in the Y. The elongation of the stored tortillas did not show statistical difference due to the effect of the storage time, the values were 4.3-5.1 mm (24 h) and 4.2-4.6 mm (48 h).

The best tortillas with 24 h of storage were from the H-49AE hybrid, these required less force to break (x= 1.6 N), their EL was 4.4 mm and their Y was 46.8 N cm^-2 (Figure 2 A and B). The good quality of the stored tortillas could be explained by the oil content of these corn (±5%). In this regard, it has been reported (^{Vázquez et al., 2015}) that lipids interact with amylose molecules during nixtamalization, affecting the physicochemical properties of starch and favoring the formation of the amyl-lipid complex, which favors the texture of tortillas (^{Vázquez et al., 2014}).