Field conditions

The experiment was conducted at the Colegio de Postgraduados, Campus Montecillo, Texcoco, México (19° 29’ N and 98° 53’ W), altitude 2,250 masl, during the 20142015 and 2015-2016 crop cycles. The climate is sub-humid, with and mean temperature of 15 °C and annual accumulated precipitation of 636 mm (^{García, 2004}). Soil is sandy loam with pH 7.1. Climate data were obtained from the meteorological station of the Universidad Autónoma Chapingo, 2 km from the study site. Accumulated annual precipitation was 1043 mm in the first year and 877 mm in the second. Higher precipitations and temperatures were recorded during the spring and summer than in fall and winter in both years (Figure 1).

Figure 1 Mean monthly high and low temperatures and monthly accumulated rainfall (PP) in the two crop cycles. 

The pasture of L. corniculatus was established by transplanting on March 14, 2014; density was 9 plants m^-2 and distance between plants was 33 cm. The plants were obtained from seedlings grown in a greenhouse in a substrate of soil typical to the study site with no fertilization. Before beginning the experiment, sheep were grazed 13 weeks after establishment and uniformizing cutting was done at the beginning of the experiment (September 5, 2014) to 7 cm from the soil. The plots were irrigated every 15 d to field capacity in the dry season; no fertilizer was applied.

The experimental area was divided into 12 sub-plots with an area of 4 m² with 36 plants per experimental unit. Cuttings were treatments and were determined when the pastures reached percentages of intercepted light of 90, 95 and 100 %, plus a fixed cutting defined seasonally (spring-summer: every 28 d, fall: every 35 d, and winter: every 42 d). LI was monitored at soil level (six measurements) using a canopy analyzer, AccuPAR Linear PAR/ LAI ceptometer Modelo PAR 80 (Decagon devices, USA).

Forage yield (kg DM ha^-1) was assessed when biomass was harvested in two fixed squares (0.25 m²) by repetition, established at the beginning of the experiment. This forage was dried at 60 °C to constant weight in a forced-air oven (Felisa, Mod. FE-243A). Botanical and morphological composition (CBM) was calculated by taking a 10 % sample of the harvested forage, separating it into leaf, stem and dead matter (senescent material) and weeds. The contribution to yield of each component (kg MS ha^-1) was calculated. Plant height (cm) was measured in 12 random readings per replication using a 50 cm graduated ruler. The equivalence in yield per cm of plant height was calculated by dividing yield by the cm of plant height by the number of cuttings. The CBM data for leaf and stem were used to determine the leaf/stem ratio (R:L/S: 1), dividing leaf weight by stem weight.

The treatments were assigned to experimental units according to a design of random blocks in divided plots with three replications, where the large plot was the genotype and the small plot the treatment. The data obtained per cutting were organized by season and analyzed with PROC GLM of SAS (^{SAS Institute, 2009}); averages were compared with the Tukey test (p≤0.05).

Forage yield

Differences (p≤0.05) in annual forage accumulation were found only in the second year (Table 1), and the lowest yield corresponded to the fixed cutting with 29 % less than the average of the three percentages of intercepted light (32,173 kg MS ha^-1), indicating less adaptation of the species to fixed cutting management, relative to the cuttings that depended on the percentage of intercepted light. Regardless of the treatment, the highest mean seasonal yields, in both years, were obtained in spring with 10,461 kg MS ha^-1 and the lowest in fall (5,915 kg MS ha^-1) and winter (5,585 kg MS ha^-1). This behavior can be explained by the direct and close relationship between forage yield and optimal temperatures (22 °C) for growth of the species in field conditions (^{García et al., 2014}). Optimal temperatures occurred in spring (Figure 1) benefitting L. corniculatus growth and productivity. In our study, seasonal behavior is similar to forage yield reported by ^{Scheffer et al. (2011)} of five populations of L. corniculatus, but unlike the behavior with fixed cutting established in our study, these authors observed that a high cutting frequency increases yield. ^{Sbrissia et al. (2010)} attribute these changes in yield to environmental conditions in each season of the year. ^{Díaz et al. (1996)} found a seasonal distribution of L. corniculatus forage production of 49, 25, 14 and 12 % for spring, summer, winter and fall, respectively, with 72% in spring-summer. According to ^{Santiñaque and Battista (2003)}, there is an increase in yield of 45% at harvest intervals of 44 d, relative to that at 22-d intervals. For 45-d cutting intervals, ^{García et al. (2014)} observed an average forage yield of 7700 kg MS ha^-1, and the genotype 202700 was the most productive of the 12 L. corniculatus genotypes evaluated. They also pointed out that accumulated production can be influenced by climate, management and the growing habit of the genotype.

Accumulated forage yield of L. corniculatus in function of the percentages of LI were not different (p>0.05), unlike the findings of ^{Da Silva and Nascimento, (2007)}, ^{Da Silva and Hernández, (2010)}. ^{Mattera et al. (2013)} indicated that forage species should be harvested at 95 % intercepted light, but only when plant density is competitive. In this respect, ^{Rojas et al. (2016)} in alfalfa (M. sativa L.) recorded 95 % intercepted light only in summer; this was explained by the low plant density due to the time of establishment of the pasture. ^{Mattera et al. (2013)} also found that 95 % intercepted light in alfalfa is achieved in spring when combining distances between plants of 10, 15, 20, 25 and 30 cm, while in summer and winter only densities of 10 and 15 cm are present.

Botanical and morphological composition

The percentage of weeds in the pasture was low, 4 % annual average, in both years (Table 2). In this respect, ^{Teixeira et al. (2007)} mentioned that the presence of weeds in a pasture can increase over time as an effect of cutting frequency. Among the morphological components, leaves contributed most to yield with an annual average of 17 247 kg MS ha^-1 (58 %), followed by stems (31%) and dead matter (7 %). This behavior can result from the accelerated rhythm of leaf and stem growth favored by environmental conditions such as photoperiod, temperature and humidity (^{Difante et al., 2009}).

In both years, the treatment with 95 % intercepted light produced the largest amount of leaf, 19,240 kg MS ha^-1 on average, and the fixed cutting had the lowest amount of leaf and also of stem, 14,634 and 7,090 kg MS ha^-1, respectively (p≤0.05), relative to the other treatments (Table 2). These results are similar to those of ^{Giacomini et al. (2009)}, who evaluated B. brizantha cv. Marandu at intercepted light of 95 % and 100 % and two cutting intensities (10 and 15 cm); they found a larger quantity of leaf at 95 % intercepted light at a height of residual forage of 10 cm. In our study, dead forage was not different (p>0.05) among treatments during the first year. In the second year, the highest values were recorded with 100 % intercepted light and the lowest at 95 % intercepted light and fixed cutting. This indicates that the optimal harvest time is when intercepted radiation is at 95 %, which produced the best quality forage and, based on the lowest quantity of dead matter, the highest yield (Table 1, year 2) and the largest amount of leaf (Table 2), similar to the results reported by ^{Montagner et al. (2012)}. Moreover, the minimal presence of dead or senescent matter in the fixed cutting could be because the plant did not reach its optimal physiological moment for harvesting and more light penetrated into the pasture (^{Matthew et al., 2000}), causing higher production of green matter, leaves and stems but of lower weight (^{Villegas et al., 2004}).

Depending on the season of the year, leaves were the component that most contributed to forage yield (Figure 2). The highest and the lowest seasonal yields were found in spring and fall-winter, respectively, with 5813 and 3378 kg MS ha^-1. The treatment with 95 % intercepted light and fixed cutting had yields of 6480 and 5177 kg MS ha^-1. These results indicate effects of season and of the treatments on production of leaves, influenced by the environmental conditions during the study (Figure 1). Moreover, the highest and the lowest quantities of stems were produced during spring and winter, with seasonal averages of 3814 and 1612 kg MS ha^-1. In this respect, ^{Sbrissia et al. (2010)} mention that the production of stems changes in function of the seasonal variations in quantity and quality of light that reaches the pasture, precipitation and temperature. Regarding dead matter and weeds, there were no seasonal differences between the two years.

Figure 2 Botanical and morphological composition of L. corniculatus, genotype 202700, in function of percentages of intercepted light and one fixed cutting (CF^¶) defined seasonally (fall: 35d, Winter: 42 d and spring-summer 28 d between cuttings). 

Plant height

There were differences among treatments applied and seasons of the year (p≤0.05), except in summer of the second year (p>0.05). The highest average annual heights were recorded in 100 and 95 % intercepted light with 23 cm in the first year and 24 cm in the second, while the lowest heights were in the fixed cutting with an average of 19 cm in both years. These results are related to plant age because the higher the percentage of intercepted light the more the days of plant regrowth (51 d, average of the two years¹), relative to fixed cutting (33 d, both years⁴), which was always at a younger age with more frequent cuttings, affecting forage yield (^{Villegas et al., 2006}). In this respect, ^{Santos et al. (2004)} pointed out that taller height means higher forage yield, but ^{Scheffer et al. (2011)} related height and yield of L. corniculatus to growth habit (erect and prostrated).

In the first year, the tallest heights were found in spring and summer with 24 cm and the shortest in winter with 18 cm, while in the second year, the tallest heights were found in spring with 25 cm and in fall and winter the shortest, with 19 cm. The tallest plants were found in the seasons with the conditions of humidity and temperature that were optimum for growth and yield of the species (Figure 1). These results are similar to those of ^{García et al. (2014)}, who evaluated 12 genotypes of L. curniculatus. Also, according to ^{Marcelino et al. (2006)}, with adequate conditions of photoperiod, temperature and humidity, forage plants have accelerated growth and change in height in different seasons (^{Difante et al., 2009}), correlating positively with forage yield (^{Da Silva and Nacimiento Jr., 2007}), as occurred in our study.

Plant height can be associated with forage production (Hakl et al., 2010). Moreover, production of L. corniculatus forage is related to height, but also to the growth habit of the genotype, with the lowest yields and heights in prostrated genotypes (^{García et al., 2014}). Height of the pasture gives an idea of the forage produced. Therefore, when calculating the equivalence of the highest average seasonal yields and of the treatments, each cm of height corresponds to 178 kg MS ha^-1 for spring and 184 kg MS h^-1 for 95 % LI, averages of both years.

Leaf: stem ratio

The highest annual average leaf: stem ratio occurred with fixed cutting (Table 4), followed by 90, 95 and 100 % intercepted light in both years. Fixed cutting surpassed the average of the treatments with 95 and 100 % LI by 26 % (2.3 vs 1.7: 1) in the first year, and in the second year by 39 % at 100 % LI (2.8 vs 1.7:1). In the fixed cutting treatment, the highest leaf: stem ratio was the result of more frequent cuttings (33 d yearly average⁵), which did not permit harvesting the species at its optimal moment, but rather in the phase of accelerated growth when the largest percentage of young leaves and less stem is found. Consequently, there was more light penetrating into the interior of the pasture (^{Barbosa et al., 2011}). Also, ^{Montagner et al. (2012)} mention that, after cutting, the quality and quantity of incidental light inside the pasture is altered by the cutting interval, which causes variations in the production of leaves and stems and, therefore, in the leaf: stem ratio.

The leaf: stem ratio in both years was different (p≤0.05; Table 4) between seasons. It was higher in fall and winter, with 2.1:1 and 2.5:1 for the first and second years. The lowest values of the leaf: stem ratio were recorded in spring, in both years, and were different to summer only in the second year (p≤0.05). These results are similar to those reported by ^{Barbosa et al. (2011)}, who found a higher leaf: stem ratio in seasons when plant growth is slower (fall and winter), due to a higher density, but lower weight, of stems. Moreover, a lower leaf: stem ratio in spring and summer results from heavier individual stems (^{Villegas et al., 2004}). In this respect, ^{Giacomini et al. (2009)} mention that translocation of assimilates, from leaves to stems, is greater in spring and summer, resulting in heavier stems. The proportion of leaf relative to stem was higher in all seasons and treatments, with an average of 1.9:1 (first year) and 2.1:1 (second year).