Height of plant (cm)

For this, four plants were taken at random from each plot and measured from the soil to the longest leaf apex and panicle apex. This activity was carried out every 10 days from the first weeding until presenting anthesis.

Index of leaf area (cm²)

It was determined in the period of anthesis and in milk grain state in each experimental unit, for this, a complete plant was extracted and the leaf area of each leaf, present in each stem, was measured, using the following formula: leaf area (long x width of each foliar leaf of the plant) x 0.75 (^{Carrillo and Ruiz, 2004}).

Morphological composition (g MS pl^-1)

From the plants to measure the leaf area, they were separated into leaf blades, stems, panicles and dead material, then the material was placed in paper bags for drying in a stove (CIDERTA^®) of forced air at 55 °C for 72 h and weighed on a scale (Lecsa^®; g or kg).

Seed production ha^-1 (kg)

It was obtained by weighing the totality of caryopses of each experimental unit and by a rule of three, it was related to 1 ha.

Weight of a thousand seeds (mg)

Eight repetitions of 100 caryopses taken at random from each experimental unit were counted and weighed using an analytical scale (Ohaus^®) and the weight of 1000 caryopses was obtained by multiplying by ten the arithmetic mean of the eight repetitions (^{ISTA, 2012}).

Test weight (kg 100 L^-1)

A volume of one liter of seeds was weighed four times for each of the 36 experimental units and was related to 100 L.

Purity (%)

From the seed obtained from each experimental unit, four repetitions of 25 g of each were weighed and separated into three components: seed of the species of interest, other seeds and inert material (stones, soil particles and fragments of plants). According to the ^{ISTA (2012)} it considers as seeds those that are complete or intact and also those that are smaller than normal, wrinkled, diseased or germinated, as long as they can be identified as belonging to the species analyzed. To obtain the purity, the following formula was applied: purity = weight of pure seed (g) x 100/total weight of the sample.

Moisture content (%)

The direct stove method was used, which consisted of weighing a sample of 24 g, obtained from the batch of seeds harvested in each experimental unit, which was separated into four subsamples of 6 g and each was placed in a metal container with lid, which were weighed before and after entering the stove. The stove was calibrated at 103 ±1 °C and the samples were maintained for 6 h. The moisture content was determined by the difference in seed weight before drying and after drying (^{Moreno, 1996}).

Number of seeds per kg

In each experimental unit four repetitions of five grams were weighed and the number of caryopses was counted and subsequently, by rule of three, the respective content per kg was calculated.

Water efficiency It was obtained by dividing the amount of seed of each treatment between the water consumed in the irrigations (m³) plus the accumulation of rainfall.

Statistical analysis

A completely randomized experimental design was used and the data obtained were analyzed by the GLM procedure of ^{SAS (2009)}. The means of the evaluated variables were compared by the Tukey test (α= 0.05).

Discussions

The characterization of pearl millet MF13 seed before planting (Table 1), shows good germination with an adequate moisture content (^{Moreno, 1996}), as well as good germination.

Plant height

Only differences in height were observed at the second sampling date (p≤ 0.01, Table 2), in the control treatment the lowest height (70 cm) was recorded and in the remaining dates, as well as no differences were observed on average (p> 0.05). Therefore, the use of mycorrhizae or chemical fertilizers, as well as their combination, under irrigation does not affect the plant height between fertilizers and control, which differs with ^{Plana et al. (2008)}; ^{Díaz-Franco et al. (2008b)} who report that mycorrhizal inocula promote the development of cereals.

Leaf area

No differences were observed in the length and width of the leaves with the assigned treatments (p> 0.05, Table 3), so the use of mycorrhizae and chemical fertilizers, as well as their combination, do not have an impact on the greater leaf area of the plants. pearl millet, which differs with Dell’Ámico et al. (2002) who reported 20% more leaf area when using mycorrhizae compared to the control (without inoculation). In the morphological composition in physiological state of anthesis of the plant (Table 3), significant differences were observed only in biomass of MS of stems (p≤ 0.01), the highest value was result of fertilization 120-60-00 (368 g MS pl^-1) and the minor the control (167 g MS pl^-1); however, no differences were found in leaves, panicles, dead material or in total (p> 0.05). When fertilizing with both sources, higher biomass of dry matter was obtained per plant, therefore, the use of mycorrhizae and fertilizers, as well as their combination was important to obtain heavier stems and in case that it is decided to provide the green milled cattle, it covers protein requirements in cattle of 350 kg (12%, ^{NRC, 2001}; ^{Urrutia et al., 2014}).

Morphological composition (g MS pl^-1)

The milk grain state of the pearl millet plants only showed difference in the weight of stems (Table 4, p≤ 0.01) and the control resulted in lower weight of stems and was different to biofertilization and chemical fertilization and their combinations (p ≤ 0.01), which agrees with ^{Díaz-Franco et al. (2008a)}; ^{Díaz-Franco et al. (2008b)}; ^{Plana et al. (2008)}. It is worth mentioning that, on average, stems contribute 65%, panicles 25%, leaves 7% and dead material 2%, so in this crop, the application of mycorrhizae or combinations with mineral nutrition or with mycorrhizae will have an impact on the production of green matter; for the foregoing, ^{Urrutia et al. (2014)} in pearl millet mention that the protein content for the stages of the seventh leaf, flag leaf, booting and flowering were: 15.4, 13, 12.7, 12.2, respectively, while for the in vitro digestibility of the dry matter (DIVMS) were: 81.7, 72.5, 74.5 and 69%, respectively, and metabolizable energy (ME) was not affected (1.9 Mcal kg^-1), therefore, Pearl millet is a good forage alternative for the semiarid zone of Mexico.

Seed production ha^-1 and water use efficiency

Significant differences were observed in seed production (p≤ 0.05, Table 5) and it is in the range of production reported by ^{Hernández et al. (2013)} and the highest yield was obtained when fertilizing with 120-60-00, which was similar to mycorrhizae INIFAP (p> 0.05), which coincides with ^{Díaz-Franco et al. (2008a)}; ^{Díaz-Franco et al. (2008b)}; ^{Plana et al. (2008)}; ^{Mena et al. (2013)} and on the other hand the production of Azo Fer^® mycorrhizal seeds, combinations and control were similar (p> 0.055). Therefore, with adequate plant nutrition, increased seed production (^{Villanueva et al., 2001}; ^{López-Castañeda, 2006}) and therefore greater use of water.

Physical analysis of the seed produced

No differences were observed in the physical analysis variables of seeds from chemical fertilizers, mycorrhizae, combinations and control (p> 0.05, Table 6) and the biological or chemical fertilization influences yield, but not in hectoliter weight and weight of a thousand seeds; however, ^{Noda (2009)}, in wheat, reports increases of 20 to 50% in the number of grains per spike. It should be mentioned that the hectolitre weight of the treatments was 3% lower than that found in the characterization of seeds before sowing, while the weight of one thousand seeds of the control was 20% higher compared to that found in initial characterization and, on the other hand, comparing the amount of seeds per kg, this was greater during the initial characterization of the seed since the caryopses were smaller compared to that obtained with the evaluated treatments.