Introduction

Mombaza grass is a cultivar of the species Megathyrsus maximus (Simon & Jacobs), formerly Panicum maximum (Jacq.), its origin is Africa. It was first introduced to America in 1967 (^{Jank, 1995}), it is a perennial grass with good agronomic characteristics, adaptable to soils with low fertility and resistant to drought (^{Papalotla, 2001}), with a dry matter yield of 22.8 t DM ha^-1 year^-1 (^{García et al., 2008}) and 14.89% of PC at 35 days of regrowth (^{Guerdes et al., 2000}). However, there is low availability of seed in the national market.

Mombaza seed production has a high economic potential; however, seed yield is very low, and there is not enough information in Mexico on aspects such as reproduction and agronomic management in seed production, and the low availability and poor quality of the seed are factors that limit its use and the sowing of new areas with this forage grass.

There are several problems in seed production, where the main ones are the low yield and the poor quality of the harvested seed, due to a low number of inflorescences, poor timing of flowering and irregular maturation of the seed (^{Boonman, 1979}). Other problems that this species presents is the height, lodging and fall of ripe spikelets, or inflorescences appear in different stages of development within the bunch, which causes a prolonged flowering period (^{García and Ferguson, 1983}).

The Mombaza grass reaches a height of up to 4.5 m (^{Bogdan, 1977}), which makes it difficult to harvest the seed, and causes lodging. Likewise, the presence of winds at the end of flowering and in times of maturation causes the loss of ripe seeds and, consequently, a low yield. Therefore, to increase seed production, it is essential to decrease plant height, increase the number of stems and achieve homogenization of flowering to achieve an efficient harvest (^{García and Ferguson, 1983}). It has been pointed out that in addition to nitrogen fertilization and plant density, precutting is an agronomic practice that could increase seed yield and quality in tropical forage grasses.

The precut consists of making a uniform cut before flowering, in order to obtain a uniform flowering and reduce the plant height (^{Toledo et al., 1989}). It has been shown that precut intervenes in the synchronization of the flowering, as well as in the decrease of the lodging and, consequently, the seed yield is increased. For example, in the Andropogon gayanus species, with precut made on July 31, a yield of 115 kg ha^-1 of total seed was obtained, 40% more, compared to the control 82 kg ha^-1; ^{Sosa et al. (2001)}.

Seed of M. maximus is not produced in Mexico because the technology to produce it is unknown; in addition to this, there is little information regarding the practice of precutting to improve seed yield in this species. For this reason, it is necessary to generate technologies to produce forage seed, especially in the tropical area, increasing the yield obtained and with a view to implementing a national seed industry to break dependence on imports. Therefore, the objective was to evaluate the yield and quality of guinea grass seed (Megathyrsus maximus) cv. Mombaza, at different pre-cut dates.

Materials and methods

The study was carried out in the experimental field of the University of the Papaloapan Campus Loma Bonita, Oaxaca, whose geographic coordinates are 18° 06’ north latitude and 95° 53’ west longitude, at 30 masl. The climate of the place is warm humid, with abundant rains in summer. The monthly average temperature and annual precipitation are 26 °C and 1 801.4 mm, respectively (^{FAM, 2015}). The soil is of sandy loam texture, with pH 4.9 and 0.8% of organic matter, 14.8, 23.5, 37, 241 and 42.3 mg kg^-1 of N, P, K, Ca and Fe, respectively.

The treatments consisted of eight precuts every 10 days from the uniformity cut, which was carried out on July 1, 2015 (Table 1). These treatments were distributed under a randomized complete block design, with three replications. The size of the experimental plot consisted of four furrows separated 80 cm and 70 cm between bushes. The dimensions of the plot were 3.2 × 5.6 m, for a total area of 17.92 m² and a useful plot of 6.72 m² located in the two central furrows, leaving the last tiller at each end.

The grassland where the experiment was carried out was planted in November 2004. Botanical seed was used and 8 kg ha^-1 of commercial seed was required. At the beginning of the rainy season, a uniform cut was made at a height of 15 cm and then pre-cuts were made on the dates indicated in Table 1. Immediately after each pre-cut and according to ^{Joaquin et al. (2001)} was fertilized with 100, 50 and 50 kg ha^-1 of nitrogen, phosphorus and potassium, respectively. Urea (46% N), triple calcium superphosphate (46% P₂O₅) and potassium chloride (60% K₂O) were used as the fertilizer source. Weeds were controlled by manual machete plating, from the beginning of the experiment to the beginning of flowering.

The seed was harvested manually 18 days after flowering (^{Joaquín, 2002}; ^{Padilla and Febles, 1975}). The flowering time was considered when 50% of the panicles present were in anthesis, that is, with the anthers visible. To determine the anthesis moment, four tillers were selected at random within each useful plot, which were observed every third day (^{Joaquín et al., 2010a}).

The seed harvest was carried out using the traditional technique for harvesting tropical grass seeds (^{Ferguson, 1979}), which consists of cutting all the inflorescences present and subjecting them to a natural sweating process. All the panicles of the tillers were harvested in the two central rows, leaving the tillers of the shore unharvested at both ends. To simulate the sweating process, the harvested panicles were placed in blanket bags, which were grouped on the ground, and covered with the plant material that remained after the panicles had been cut. The sweat period was four days. Subsequently, threshing, cleaning and drying of the seeds were carried out naturally (in the sun). The obtained seed was weighed, packed in paper bags and stored under laboratory environmental conditions.

The variables evaluated were: plant height (cm), number of total panicles per m², number of ripe panicles per m², panicle length (cm), number of seeds harvested per panicle, weight of seeds per panicle (g), purity (%), thousand seed weight (g), total seed yield (kg ha^-1), pure seed yield (kg ha^-1), germinable pure seed yield (kg ha^-1) and seed germination harvested (%). To calculate the total seed yield, the area of the useful plot (6.72 m²) was taken into account. To calculate the yield of pure seed took into account the percentage of purity. The purity percentage was determined by weighing a sample of two grams of seed and separating it into its components: pure seed, vain seed and impurities, as indicated by the ^{ISTA (2005)} rules. To calculate the yield of pure germinable seed took into account the percentage of germination.

For the height of the tiller, four tillering were measured at random per plot and the measurement was made from the base of the tiller to the upper end of the inflorescence. For the number of total panicles per m² and number of ripe panicles per m², all the inflorescences present in four previously randomly selected tillers within each useful plot were counted, and the estimation was made based on the density of tillers.

To measure the length of the panicle, the number of seeds harvested per panicle and the weight of seeds per panicle, 10 panicles were harvested per plot, taken at random within the tillers of the useful plot. The length of the panicle was measured from the insertion point of the first branch, to the upper end of the panicle. The number of seeds harvested per panicle was quantified as the number of spikelets per panicle at the time of harvest. The weight of 1 000 seeds was estimated as the average of eight repetitions of 100 pure seeds per plot and multiplied by 10 (^{ISTA, 2005}).

For the germination test, 400 pure seeds were used in four replications of 100 seeds, which were placed in Petri dishes with a lid 9.5 cm in diameter and 1.5 cm deep, provided with absorbent paper, moistening the substrate with KNO₃ solution at 0.02% and placed on a table inside the laboratory at an ambient temperature of 30 °C and constant light, for 28 days. Three counts were performed at 10, 19 and 28 days, in each count the number of normal seedlings was quantified, according to the ISTA methodology (2005). The germination percentage was estimated from the normal seedlings.

Data were subjected to analysis of variance, based on a randomized complete block experimental design. The comparison of means of the treatments was carried out using the Tukey test, with a significance level of 0.05. In addition, a correlation analysis was performed to estimate the degree of association between seed yield and yield components: number of total panicles per m², number of ripe panicles per m², panicle length, number of seeds harvested per panicle and seed weight per panicle.

Results and discussion

Conclusions

Based on the results obtained, it is concluded that precut has a positive effect on the yield and quality of guinea grass seed, where the highest yield was found with the precut date of July 20. The precut and its positive effect on the increase of seed, is due to the increase in the number of seeds harvested per panicle and the weight of seeds per panicle since they presented the highest correlation indexes. It is recommended to carry out the precut at the end of July, and to continue with this study in this and other cultivars and grass species in order to determine more precisely the precut date and its effect on the yield and quality of the harvested seed.