Forage yield

The FY varied significantly between harvest intervals (CI) and season of the year (Table 2). This was affected by the CI (p< 0.0001) and the season of the year (p< 0.0001). A FC interval caused a lower forage production in the winter seasons (3 441 kg DM ha^-1), summer (5 024 kg DM ha^-1) and in the accumulated annual (20 710 kg DM ha^-1), regarding the harvest intervals depending on the percentage of light intercepted. According to ^{Duru and Ducroq (2000)}, the plants with a higher frequency of defoliation (Table 1), present lower forage yield, caused by a reduction in the multiplication of cells.

Thus, at the end of the year the harvesting frequencies 90 and 95% IL were higher than the FC, with values of 28 222, 27 962 and 20 710 kg DM ha^-1, respectively, which could have been caused by poor accumulation of stems at the crown level (^{Da Silva et al., 2010}; Table 3). In general, seasonally, the FY varied from 3 441 kg DM ha^-1 in winter in the FC, to 8 265 kg DM ha^-1 in summer in 95% IL. This last management strategy presented the highest yield in the winter and summer seasons with 5 799 and 7 547 kg DM ha^-1, respectively (p> 0.05).

Therefore, the harvest conditions with 95% IL, in the winter, summer and in the annual accumulation, the FY was higher than the FC (p< 0.05). In the autumn and spring seasons there were no differences between the different harvesting strategies (Table 2). However, if there were differences between seasons (p< 0.05), spring was greater (8 265 kg DM ha^-1) than that recorded in winter (4 899 kg DM ha^-1). The foregoing was probably the result of a greater and lesser production of leaf, stem and plant height, in the different harvest intervals, within each season (Table 3), with which it is highly correlated (^{Rojas et al., 2016}).

Botanical and morphological composition

The morphological components of the species and botanical composition of the prairie were influenced by the CI (p< 0.0001), component (p< 0.0001), and the interaction of CI x component (p< 0.0001). The frequency of harvest of the FC, resulted in less production of stems and dead material, compared to the rest of the treatments (p< 0.05; Table 3). This was probably the result of a shorter regrowth period in the FC of 33 days annual average, with respect to the cut-off intervals depending on the percentage of light intercepted at 55, 60 and 64 days, annual average for 90, 95 and 100% IL, respectively (Table 1).

However, the leaf and the weed did not differ between treatments (p> 0.05); however, the averages show us that the leaf is the component that made the greatest contribution to the annual yield with 12 908 kg DM ha^-1 (Table 3). The leaf contributed 51% of the annual yield (25 194 kg DM ha^-1, Table 2), followed by the stem, weeds and dead material with 32, 10 and 6%, respectively. This was reflected in a higher and lower total forage yield, depending on the amount of leaf produced, between seasons and treatments, given the high correlation between these variables (^{Álvarez et al., 2018a}).

However, the stem is also an important morphological component that allows a spatial arrangement of the plants and accumulation of assimilates that translocate to the leaves (^{Fagundes et al., 2006}). According to ^{Barbosa et al. (2011)}, the accumulation of stems increases in spring and declines as there is a greater deficit of water and low temperatures; during autumn and winter (Figure 2), altering the leaf:stem ratio (Table 4).

Seasonally, the leaf was affected by the CI (p< 0.0001), season of the year (p< 0.0001) and the interaction CI x station (p= 0.0381). It was the highest component in all seasons, followed by stem, weed, and dead material. Its greatest presence was in spring with 5 551 kg DM ha^-1 (CI average; Figure 2).

This may be related to a greater availability of climatic factors that favor its growth such as water, temperature and light (^{Barbosa et al., 2011}). According to ^{Marcelino et al. (2006)}, under suitable environmental conditions, the plants increase the appearance and elongation rates of the leaf, therefore, they make the accumulation of biomass basically depend on the production of leaves.

For its part, the stem was influenced by the CI (p< 0.0001), season of the year (p< 0.0001) and the Station x CI interaction (p= 0.0185). Similar to the leaf, this was higher in spring with 3 204 kg DM ha^-1 (average of the cut intervals), favored by optimal environmental conditions (^{Barbosa et al., 2011}). In the case of the leaf, only in autumn and spring, the treatment with 90% IL, marked differences with the rest of the harvest intervals (p< 0.05) and, on the other hand, the stem was greater in 90 and 95% IL to FC (p< 0.05), in all seasons of the year (Figure 2).

Figure 2 Botanical and morphological composition of bird's foot clover (Lotus coniculatus L.) genotype 260012 at various harvest intervals; 90, 95 and 100% IL and a fixed cut (FC) defined by season (autumn: 35, winter: 42 and spring-summer: 28 days between cuts). 

Low stem production in FC can be explained by a process of intra-specific competition for growth factors such as light, which gives rise to the process known as ‘size/density compensation of stem population’ (^{Matthews et al., 1995}; cited by ^{Barbosa et al., 2011}). However, the rate of appearance of leaves determines the production of stems, since this reduces and starts the death of stems when 9% of interception of incident light on the grassland is exceeded and the quality and quantity of light is limiting, therefore, the size of the stems is increased to compensate for the number of these (^{Mazzanti et al., 1994}).

Furthermore, when 100% of the light intercepting the plant canopy is intercepted in the grassland, the composition in the light spectrum changes inside the canopy, reducing infrared light (^{Gautier et al., 1999}) and this alteration reduces tillering and stem biomass (^{Barbosa et al., 2011}). As the number of leaves increases, the plant invests resources in stem elongation in an attempt to place leaf mass on top of the grassland canopy (^{Da Silva and Hernández, 2010}).

Therefore, the quantity and quality of light is progressively reduced at the level of the basal area, causing the presence of dead material as a consequence of leaf senescence, as a way of keeping the number of live leaves per stem relatively stable (^{Sbrissia and Da Silva, 2008}), which was evident in the treatment of 95 and 100% IL in all stations, with the presence of dead material (Figure 2).

However, weeds and dead material were only affected by the season of the year (p< 0.0001), registering their highest presence in spring with an average of 1 356 and 859 kg DM ha^-1, respectively, which can be reflection of environmental conditions rather than the harvest interval (^{Rojas et al., 2016}).

Leaf: stem ratio (R:L/S)

The R:L/S was affected by the season of the year (p< 0.0001) and the cutoff interval (p< 0.0001) and by the CI x Station interaction (p< 0.0002). Both the seasons and the annual average show that the FC presented the highest R:L/S values and the lowest in the harvest intervals depending on the percentage of light intercepted (90, 95 and 100%).

This may be related to a higher harvest frequency, since the FC had an IC of 33 days (annual average) and the average of the plots managed with 90, 95 and 100% IL was 60 days (Table 1), which that caused an increase in leaf production with respect to the stem (^{Barbosa et al., 2011}) and did not allow the species to be harvested at its optimum moment, but in the phase of accelerated growth, where the highest percentage of young leaves is found and little number of stems (^{Álvarez et al., 2018b}).

According to ^{Lemaire and Chapman (1996)}, a greater growth of the leaf is a plastic response of the plant to a high defoliation frequency, considered as a strategy of the plant as an escape to defoliation. Seasonally, the highest average values were recorded during autumn and winter with 1.9 and 2, respectively (Table 4), which may be related to a greater amount of light transmitted at the level of the basal part of the plant, in the seasons where it has lower heights (Table 5).

This behavior reflects a greater weight of the leaf with respect to the stem, more distant in autumn and winter compared to spring and summer. This may be the result of seasonal fluctuations in the amount of light received by plants, causing alterations in the amount of leaf produced (^{Matthew et al., 1999}).

These results are similar to those reported by ^{Álvarez et al. (2018a)} in the genotype 202700, who found mean values of 2.1 for autumn and winter compared to spring and summer with 1.7, due to a higher density of stems, but of less weight. ^{Giacomini et al. (2009)}, mention that there is a translocation of assimilates from the leaves to the stems higher in spring and summer, which causes a higher weight of stems, but less number of these.

Plant height (PH)

The PH was affected by the season of the year (p< 0.0001), the harvest interval (p< 0.0001) and the Station x CI interaction (p= 0.0002). According to the annual and seasonal average, with the exception of autumn where there were no statistical differences (p> 0.05), the frequency of harvest of the FC showed lower plant height with an annual average of 21 cm, compared to the rest of the treatments with 26.6 average cm (Table 5). Seasonal behavior showed that the averages in spring and summer were higher in plant heights with 29 cm, compared to winter with 19 cm (p< 0.05).

These results were consistent with the productive behavior of the species in forage yield (Table 2). This could be promoted by a shorter harvest interval of 33 d in the FC compared to 60 d of the grasslands managed with the percentage of light intercepted strategy (90, 95 and 100%) since, at a higher height it corresponded to a higher yield of seasonal forage and between treatments, due to the high correlation that exists between these two productive variables (^{Halk et al., 2012}; ^{Álvarez et al., 2018b}).

Furthermore, the highest values, found in spring and summer, may be the result of more favorable conditions for higher stem growth (^{Rojas et al., 2016}). However, some works in the species Lotus corniculatus L., relate the height of the plant and forage yield with the growth habit of the genotype, since these can range from erect to prostrate (^{Schefer et al., 2011}).