Introduction

In Mexico 8.5 million ha maize (Zea mays L.) are sown each year; 25 % of this surface with improved seed, and 75 % with local varieties (genuine landraces or 25 % with locally adapted, improved varieties, and landraces that have outcrossed with improved varieties or advanced generations of hybrids). The genetic influence of introduced improved varieties has provided advantages for the landrace maize varieties.

The use of improved seed is concentrated in the areas of high productive potential (irrigation or abundant rainfall). Each year the growers are reluctant to purchase new seed, because the cost is the highest in the world (^{Espinosa et al., 2009}; ^{Espinosa-Calderón et al., 2012a}; ^{Espinosa-Calderón et al., 2014}); improved seed is used in only 6 % of the area of the High Valleys of Mexico, 2200 to 2600 masl, improved seed is used (^{Espinosa et al., 2012b}).

As the seed market is controlled by a few private companies, 1000 seeds can cost as much as 2.71 dollars, whereas in the U.S. the price is 1.34 dollars (^{Espinosa et al., 2008}; ^{Espinosa et al., 2009}). This is aggravated by the distortion of the seed system in Mexico due to the closing of the National Seed Producer (PRONASE). This resulted in the domination of large international companies and private companies (^{Espinosa et al., 2003}; ^{Coutiño et al., 2004}; ^{Luna et al., 2012}). Producers have preferred not to depend on commercial seeds but to control their own seed production, as they have done for more than 330 generations (^{Boege, 2009}; ^{Turrent and Espinosa, 2006}; ^{Turrent, 2009}). For this reason, they frequently use second generation hybrid seed (F₂) or more advanced generations obtained from their own plot or from neighboring producers who have bought improved seed (^{Ortiz and Espinosa, 1991}; ^{Coutiño et al., 2004}; ^{Martínez- Gómez et al., 2006}). The loss of heterosis causes a lower yield in the second generation with respect to the first (^{Ortiz and Espinosa, 1991}; ^{Espinosa et al., 1998}; ^{De León et al., 1998}; ^{Espinosa-Calderón et al., 1012a}). According to ^{Martínez-Gómez et al. (2006)}, the sowing of advanced generations has spread in various maize growing regions of Mexico (^{Ramírez et al., 1986}; ^{Coutiño et al., 2004}; ^{Espinosa-Calderón et al., 2012a}), to a greater or lesser degree depending on the availability and supply of seed.

The decrease in yield of generation F₂ may depend on the structure and conformation of each hybrid (trilinear, simple or double), of the nature of the parents and of the degree of endogamy of its lines (^{De León et al., 1998}; ^{Espinosa et al., 1998}). The different ecological areas of Mexico also influence the genotypes that are used (^{Ramírez et al., 1986}; ^{Ortiz and Espinosa, 1991}; ^{Espinosa et al., 1998}; ^{De León et al., 1998}). In the largest area of the Highlands, the hybrids H-48 and H-50 have been sown (^{González et al., 2007}; ^{González et al., 2008}). In the Valley of Mexico Experimental Field (CEVAMEX) other hybrids have been developed in addition to the above, such as H-40, H-66, H-70 and H-51 AE, which have adequate male sterility for producing hybrid seed (^{Espinosa et al., 2012b}).

Although there is a reduction in yield of the F₂ generation with respect to the F₁ generation in maize, it is important to understand the specific behavior of maize hybrids of the Highlands in the F₂ generation with respect to the F₁; that is, the level of yield loss should be established. In the present study an evaluation was made of some materials, with the hypothesis that the yield of F₁ is sufficiently higher than that of F₂ to justify the acquisition of new seed each cycle.

Materials and Methods

In the 2013 spring-summer cycle, the hybrids H-40, H-51 AE, H-57 AE and Puma 1167 were used to evaluate the yield of their F₁ and F₂ generations. The study was carried out in the fields of the Facultad de Estudios Superiores Cuautitlán (FESC), UNAM, Mexico, under dry conditions and with 55 000 and 70 000 plants ha^-1.

The experimental design was complete randomized blocks with four replications, and the statistical analysis was factorial. The factors were as follows: hybrids (4), population densities (2), advanced F₁ and F₂ generations (2), and hybrid interactions x F₁F₂ x PD, hybrids x F₁F₂ x PD. Average values were compared made with Tukey test (p≤0.05).

The plot was prepared with machinery and consisted of clearing, harrowing, disking seeding with 80 cm spacing between rows. The NPK fertilization formula 80-40-00 was used. Although it is not recommended in the Highlands, the fertilizer is applied under May crop conditions, in a single basal application. Nitrogen as urea (46 % N) was used to complete the desired N units along with the diammonium phosphate (18-46- 00 of P) which was the source of P. Sowing was carried out in the second quarter of May 2013, planting three seeds per site with a 50 cm spacing within the rows, which generated 33 plants per 5 m row, equal to 82 500 plants ha^-1, and a final plant density of 70 000 ha^-1. Thinning left 28 plants per 5 m row, and in the 55 000 per ha, thinning maintained 22 plants per row, 35 d after sowing.

Weed control was made with applications of Herbamina and Gesaprim in pre-emergence, one day after sowing, in a proportion of 3 L: 3 kg per ha; and a second application 20 d after sowing with a mixture of 1 L of Sansón, 2 L of Hierbamina and 3 kg of Gesaprim per ha.

Plot size consisted of a row of 5 m length. The rows were sown with a separation of 0.80 cm, depositing three seeds per site every 0.50 cm.

Harvesting was done manually in the second half of November. All of the ears, including the damaged ones, were collected. In a representative sample of five ears, grain moisture was quantified with a Stenlite electric determiner; percentage of grain/cob was calculated after shelling the recently harvested ears and obtaining the quotient of grain weight over the weight of grain with cobs; data was also obtained of the variables ear length, rows per ear and grains per row. Previously, days to male flowering were determined in five plants per plot, when 50 % of the plants in the plot release pollen, as well as days to female flowering, when 50 % of the plants had at least 3 cm of stigma exposed. Plant height was measured from the base of the stem to the spike insertion node; ear height was measured from the base of the stem to the insertion node of the top ear. Grain yield was calculated using the following formula:

Yield = (FW X % DM x % G} X CF)/8600

where PC=field weight of the total of harvested ears per plot expressed in kg; % MS=percentage of dry matter of the grain sample of five recently harvested ears; % G=percentage of grain; FC=conversion factor for obtaining yield per ha which is obtained by dividing 10000 m2 / useful plot size in m²; 8600=constant value which permits the estimation of yield with 14 % moisture, which is used for commercial grain.

Results and Discussion

The factorial statistical analysis detected highly significant differences for yield between hybrids and F₁F₂ and significant differences between population densities. The interactions were not significant. The yield mean was 8422 kg ha^-1 and CV was 12 % (Table 1).

The highest yield of the hybrid Puma 1167 (9989 kg ha^-1) was similar to that of H-57 AE (8334 kg ha^-1), considering the mean of the F₁ and F₂ generations, and the two population densities under evaluation; both were statistically higher than H-51 AE and H-40, and were similar to each other (Table 2). H-57 AE was outstanding, because this hybrid, after evaluations in experiments for several years, is in the final stage of its validation. It was registered by INIFAP in the Colegio Nacional de Variedades Vegetales (CNVV); it is registered as H-53 AE with the number 3153-MAZ-1658-300615/C (^{Espinosa et al., 2012b}). The results confirm the favorable characteristics of this hybrid with androsterility (^{Espinosa et al., 2009}), given that it was superior to H-51 AE, which was released by INIFAP in 2012 (^{Espinosa et al., 2012b}).

The male flowering of H-51 AE occurred in fewer days (78) than the other hybrids (80). For female flowering, the weight of 200 seeds, ear length, number of grains per ear, there were no significant differences. However, in plant height, volumetric weight and grains per row, two groups were defined (Table 2).

The comparison of means of the F₁F₂ generations, considering the mean of the four hybrids and the mean of the two population densities (Table 3), showed that the F₁ generation yielded 9,040 kg ha^-1, and was statistically different from F₂, which yielded 7,804 kg ha^-1; that is, F₁ represented 115.8 % with respect to F₂. This result is similar to that obtained in other studies (^{Ortiz and Espinosa, 1991}; ^{Valdivia and Vidal, 1995}; ^{Espinosa et al., 1998}; ^{Coutiño et al., 2004}; ^{Espinosa-Calderón et al., 2012a}).

The mean values of the variables male flowering, female flowering, plant height, ear height, volumetric weight, grains per row and grains per ear were not different (p>0.05) in the F₁ and F₂ generations. Therefore, probably the difference in yield of F₂ with respect to F₁ was due to other components which were not analyzed in the present study. These results are different from those obtained by ^{Espinosa-Calderón et al. (2012a)}, because these authors detected differences in ear length and grains per ear. In both cases the mean of the F₁ generation was higher with respect to F₂ (Table 3).

Average yield of the four hybrids and the F₁ and F₂ generations was higher with the plant density of 70 000 plants ha^-1, compared to 50 000 plants ha^-1 (Table 4). These results coincided with those obtained with 70 000 plants per ha and other hybrids (^{Virgen et al., 2010}; ^{Virgen-Vargas et al., 2014}).

In all cases the yields of the F₁ generation were higher than those of the F₂ generation, with 13.9 % in H-51 AE, and 119.4 % in H-40 (Table 5). These differences between F₁ and F₂ are documented in other studies (^{Ramírez et al., 1986}; ^{Ortiz and Espinosa, 1991}; ^{Valdivia and Vidal, 1995}; ^{De León et al., 1998}; ^{Espinosa et al., 1998}; ^{Coutiño et al., 2004}; ^{Espinosa et al., 2005}; ^{Martínez-Gómez et al., 2006}; ^{Espinosa-Calderón et al., 2012a}).

Yield varied from 1,029 kg in F₁ with respect to F₂, of the hybrid H-51 AE, to 1399 kg in the hybrid Puma 1167 (average in the four hybrids of 1236 kg), equivalent to ($ 2500 per Mg of grain) a total of 3090 pesos for the commercialization of this grain. If new seed is acquired, the difference in favor of the producer would be 1690 pesos, even though the cost of the bag of seed would be 1400 pesos; thus, the purchase of seed would be profitable (^{Valdivia and Vidal, 1995}; ^{Espinosa et al., 1998}; ^{Martínez-Gómez et al., 2006}; ^{Espinosa-Calderón et al., 2012a}).

Conclusions

The hypothesis that on the average the F₁ generation exhibited higher yield and statistical difference with respect to F₂ was confirmed. The trilinear hybrid Puma 1167 showed a higher average yield than H-40. The results confirm that it is not convenient to use seed of generation F₂ due to its lower yield, which justifies the acquisition of new seed each cycle. If new seed is acquired each cycle, even with its present high cost, there would be a positive difference for the producer’s.