Analysis of variance

Combined Anova of localities generated 19 sources of variation, and mean squares, averages, coefficients of variation are shown on (Table 1), in addition a multiple range test was performed to localities, treatments, genotypes, nitrogen (N) and phosphorus (P) doses, for the four variables tested (Tables 2, 3, 4, 5 and 6).

In plant height (ADP), ear height (ADMz) and cob health (SDMz), highly significant differences (p≤ 0.01) were observed, whereas for grain yield (RGr) no significant changes were detected in the source of variation localities. Regarding treatments, ADP and ADMz did not record statistical significance; while SDMz and RGr in t ha^-1, significant effects were observed. Regarding interaction L x G no statistical significance was detected in the four variables (Table 1). This result indicates that when assessing genotypes (VS-535, V-537C and A-7573) in both environments were not affected, because the four traits were statistically similar.

Significant changes occurred in the three traits measured in maize genotypes (VS-535, V-537-C and A-7573), is due to genetic potential and consistency to the environmental conditions to which were exposed (^{Sierra et al., 2002}; ^{Palemón et al., 2011}).

For the nitrogen factor, statistical significance was detected in RGr; while in ADP, ADMz and SDMz, there were no significant changes. These results suggest that the three levels of N (100, 150 and 200) applied to the genotypes showed statistically similar traits in both environments, except RGr (Table 1).

As for phosphorus dose, significant effects were detected in ADMz and SDMz, while ADP and RGr, no significant changes were observed when applying 60, 90 and 120 units of P. These results indicate that phosphorus levels applied in the three genotypes assessed in the localities of Iguala and Apaxtla were not the adequate doses to achieve optimum grain yield.

Regarding interactions genotype by phosphorus (G x P) and nitrogen by phosphorus (N x P), highly significant effects were detected in SDMz and RGr, however, ADP and ADMz, there was no statistical significance. These results indicate that a dose of P applied in the three genotypes significantly affected two variables; while combining N x P did not modify significantly plant and ear height in the three genotypes. In combinations G x N, G x N x P, L x G, L x N, L x P, L x G x N, L x G x P, L x N x P, L x G x N x P, no significant effects were detected in the four variables measured in genotypes (VS-535, V-537C and A-7573) evaluated in the localities of Iguala and Apaxtla. This result indicates that as an interaction is more complex has no effect on the genotype traits when evaluated in two environments, it is important to assess them in more environmental conditions to see if these are stable (Table 1).

Means comparison of localities

Generally, when comparing average plant height of localities, it was observed that in Iguala, was lower (226 cm) than in Apaxtla (256.2 cm); that is, a difference of 30 cm was found. In ADMz and SDMz, the same trend was observed, because in the locality of Apaxtla genotypes showed statistically higher values than those obtained in Iguala (Table 2). The response of varieties VS-535, V-537C and hybrid A-7573 were favored in the locality of Apaxtla, because the altitude of 1182 m and average temperature 30 °C differ to those recorded in Iguala (740 m 37 °C), this behavior indicates that the varieties were selected to suit altitudes from 0 to 1 200 m (^{Gomez et al., 2003}).

In the town of Apaxtla more than 400 kg of grain yield t ha^-1 was recorded; this means that each locality has specific characteristics (p≤ 0.05) for genotypes to express their maximum production potential from one environment to another.

Means comparison of treatments

By comparing 27 treatments average, significant differences (p≤ 0.05) were detected in variables measured, except in plant height (ADP) and ear height (ADMz), this result indicates that applying the nine combinations of nitrogen + phosphorus in each genotype did not modify the size of the plant and ear in both environments, as this were statistically similar (Table 3). Regarding cob health, the highest value observed was in treatment nine (VS-535 + 200N + 120P), while statistically lower values were recorded in three treatments (VS-535 + 150N + 120P, A-7573 + 150N + 120P, V-537C + 100N + 60P), and the remaining 23 combinations were similar to the highest value recorded in VS-535 + 200N + 120P.

Plant size tended to increase when the combination 200N-90P (251.8 cm) was applied to genotype V-537C; while the combination 100N-60P applied to VS-535 showed lower plant height (229.8 cm); i.e. ADP increased 22cm when comparing these two average values by effect of N + P, although not significantly. The results indicate that treatments (VS-535 + 200-60, 245.5 cm, A-7573 + 100-120, 245.4 cm, V-537C + 200-90, 251.8 cm) affected ADP with some combination of N + P to reach the average maximum of both environments. In general, the values obtained in ADMz were below to 1 m; this result indicates that in maize plants, there is a possibility that won’t lodge, because the plant will be balanced for having cobs in the middle and facilitate harvesting manually. The highest average value (8.6) observed in SDMz, was detected in combination 200N + 120P expressed by genotype VS535, statistically surpassed treatments (VS-535 + 150N + 120P, 7.5; A-7573 + 150N + 120P, 7.3 and V-537C + 100N + 60P, 7.4); while the lowest value (7.3) was for hybrid A-7573 with the formulation 150N-120P. These results suggest that the remaining 24 treatments showed statistically similar responses in SDMz of the three genotypes in Iguala and Apaxtla, which implies that the more healthy cobs are, higher production will be achieved.

As for grain yield t ha^-1, treatment nine showed a higher value (7.1 t ha^-1), although statistically did not exceed 19 treatments, however, lower values were recorded compared with the overall average (5.4), and were represented by VS-535 + 100N + 60P, VS-535 + 150N + 120P, A-7573 + 100N + 60P, A-7573 + 100N + 120P, A-7573 + 150N + 120P, A-7573 + 200N + 120P , V-537C + 100N + 60P (Table 3). The synthetic variety VS-535 was evaluated in the Mountain region (locality Olinalá) and Northern region (locality Teloloapan), applying 120N + 90P ha^-1 and showed the following characteristics: plant height (110.25 cm), cob health (7.2) and grain yield t ha^-1 (4.377), respectively (^{Palemón et al., 2012}). ^{Mukhtar et a.l. (2011)} mention that by increasing nitrogen and phosphorus levels (300 N - 150 P) the thousand grain weight decreases, grains number per ear and grain yield. ^{Kogbe and Adediran (2003)} applied the formulation 100N-40P detecting increased corn grain yield.

Means comparison of genotypes

When comparing phenotypic characteristics of the three maize genotypes, it was observed that V-537C showed the highest ADP and did not exceed hybrid A-7573, while VS535 was statistically lower (7.9 cm) in plant size. ADMz from genotypes VS-535, A-7573 and VS535 were statistically similar. Regarding SDMz, there were no significant changes since the three genotypes have ears with the same statistically health degree in both environments. Genotypes VS-535 and V-537C recorded average statistically equal; however, the hybrid A-7573 was below 900 kg grain yield t ha^-1. The results indicate that the genotypes evaluated in Iguala and Apaxtla were not affected by the environment, because their ADMz and SDMz were similar. VS-535 was evaluated in the localities of Olinalá and Teloloapan, exhibiting similar characteristics to those found in this study (^{Palemón et al., 2012}).

According to ^{Ortiz et al. (2013)} by evaluating 16 native maize populations and as control the hybrid A-7573 recorded 4.073 t ha-1 RGr, lower value (4.9) to that observed in this study. V-537C variety is considered a stable material as it produced 5.8 t ha^-1 grain yield; similar to those reported by ^{Sierra et al. (2002)} average (5.46 RGr) from five locations; while ^{Andres et al. (2014)} reported 4.06 t ha^-1, average from four locations.

Means comparison of nitrogen doses

Comparing the effects of fertilizer doses (100, 150 and 200 kg N) applied on genotypes VS-535, A-7573 and V-537C, it was observed that in phenotypic characteristics ADP, ADMz and SDMz were not detected significant changes; however, there is a tendency that increasing fertilization dose plant height, ear height and cob health increases are not significantly. The results suggest that not finding significant changes in the phenotypic characteristics (ADP, ADMz and SDMz) of genotypes, it is possible to recommend the application of 100 units of nitrogen; however, if the doses of 150 and 200 N are applied, increases the production cost of corn (Table 5).

When comparing doses of 100 and 200 kg of N applied to maize genotypes tended to increase 900 kg in grain yield t ha^-1. ^{Shrestha, (2014)} reported that applying 200 kg N in corn, accelerated days to silking and increased stigma number and yield components (^{Beres et al., 2012}; ^{Ciampitti et al., 2012}; ^{Ciampitti and Vyn, 2013}); however, ^{Kovacs et al. (2014)}, applied 145 units of N and reported an increase in grain yield. Other researchers evaluated the effect of 200 units of N in hybrids BH40625 and BH1576, detecting significant changes in grain yield, leaf area index, dry matter production and protein content (^{Sampath et al., 2013}). ^{Tafteh and Sepaskhah (2012)} reported that the optimum nitrogen application contributes to corn grain yield, in addition it achieves a better leaf area index (^{Ding et al., 2005}), chlorophyll content (^{Majnooni et al., 2011}), available soil moisture, reflects increased grain production (^{Liu and Zhang, 2007}). ^{Mukhtar et al. (2011)} detected significant changes in plant height, thousand grain weight, grain number per ear, grain weight per ear and grain yield when evaluating hybrid YH1898 and YH1921.

Means comparison of phosphorus doses

When comparing the three fertilizations levels with phosphorus, maize genotypes VS-535, A-7573 and V537C, no significant differences were detected for ADP and RGr, while in ADMz and SDMz, there were significant changes; the results suggest that the most desirable values, are those recorded with the application dose of 90 kg ha^-1 (Table 6). When applying the dose of 60 units of P, maize genotypes expressed statistically the same grain yield t ha^-1. The results suggest that low dose (60P) is suitable for the three maize types, because when the dose increased to 120 units of P per ha, ADMz, SDMz and RGr values tended to decrease. From the economic point of view it is not convenient to increase phosphorus dose, given that with 60 units of P, similar values to those registered with the application of 120 units of P measured in the four characters evaluated (Table 6) were observed.

According to ^{Onansaya et al., (2009)} reported that applying 40 units of P, increased plant height, ear length, grain number per ear, protein content (^{Singh and Dukey, 1991}), grain yield (^{Wasonga et al., 2008} ), whereas in our study only trends were observed, probably it is important to increase the number of environments, genotypes and treatments.