The research was carried out in the Patía valley in the south of the department of Cauca-Colombia, a region with dry tropical climatic conditions, at an altitude of 550 m.s.m. with an average temperature of 28 ° C and rainfall of 1200 mm per year. This zone has deep, fertile soils of alluvial origin that are acidic to neutral pH (5.8). The relief is flat with slopes of between 1 and 3%, the soil is moderately deep to very shallow, from very poorly to well drained, with a fine to moderately thick texture, and a pH that is strongly acid to moderately alkaline. The soils undergo moderate erosion and are moderately to highly fertile, and are composed of Fluventic Haplustolls (40%), Entic Haplustolls (30%) and Udic Haplustolls (30%) [9].

Forage mixtures (grasses and legumes) were established in a silvopastoral system where Guazuma ulmifolia (Guacimo), Crescentia cujete (Totumo), and Mangifera indica (Mango) were growing, with live fences of Swinglea glutinosa. The area where the trial took place had low tree coverage and density; the trees were irregularly placed and scattered across the pastures, with most of the trees having been eliminated during the establishment of the pasture as a monoculture. However, the diameters of the trees observed show an advanced ecological succession (Silvopastoral System - Scattered Trees in El Potrero) [10].

The forage species used in the mixture were selected taking into account different results from research carried out by the Agricultural Nutrition Research group (NUTRIFACA) of Universidad de Cauca and the Tropical Forages program of the International Center for Tropical Agriculture. The sowing was carried out in line with technical recommendations, using 8 kilos of grass seed and 2.5 kilos of legume per hectare, conserving a respective proportion of 70:30 over the pasture.

The experiment used a completely randomized block design, where the block factor was influenced by the slope of the terrain, and the treatment factor was determined by the different combinations of Canavalia brasiliensis with the following grasses:

Each treatment was evaluated for agronomic behavior using the methodology of type B regional tests of the International Network of Evaluation of Tropical Pastures (RIEPT) [11].

The following production evaluations were carried out from the 12^th week after sowing:

Vigor. The state of the plant, color, growth and health is evaluated on a scale of 1 to 5, with 1 being the worst and 5 the best, with the entire trial acting as a standard of comparison.

An analysis of variance (ANOVA) was used to determine statistical differences in the blocks, evaluations and established treatments. The Duncan multiple range test was applied for the variables that presented divergences. In the case of categorical ordinals (vigor, incidence of pests and diseases), these were analyzed using correlations, frequency tables and chi-square, using the statistical program SPSS V 25.0.

The Duncan averages test shows three groups. The first group consists of treatments T6, T3 and T5; the second T3, T5 and T2; and the third T5, T2, T4 and T1. The test showed that the mixed treatments are taller than the grasses by themselves (Table 1).

Table 1

Source: The Authors.

These results can be attributed to the benefits that the legume can provide to the grass in terms of nitrogen available in the soil [14,15].

Li et al. [16] assert that the incorporation of legumes in pastures contributes to the improvement of soil fertility since they supply the soil with nitrogen through the symbiotic fixation of this element. This is reflected in the improved growth of the plant species, compared to those not planted in combination with legumes. Conrad et al. also state that the use of these plants enriches the soil with nutrients and improves its physical and chemical characteristics, contributing to changes in the architectural aspects of the pasture, that is, the height of the plants [17]. This observation was reflected in the results obtained in the present study, where the presence of Canavalia brasiliensis improved the growth of the grasses.

Specifically, it was observed that, in terms of height, of the treatments studied, Panicum maximum cv.

Mombasa showed the most growth, with an average height of 167.3 cm, while Bracharia hybrid cv. Cayman showed the least (93.4 cm), probably due to the size characteristics of each species [18], since this type of forage species has a maximum height of 60 to 90 cm [19].

In relation to the production of green forage (GF), the best contender was T3 Panicum maximum cv. Mombasa and the worst was T6 Panicum maximum cv. Mombasa + C. brasiliesis. The hybrid T2 Bracharia brizabtha cv. Toledo produced the most dry matter (DM) and T4 Bracharia hybrid cv. Cayman + C. brasiliensis produced the least (Table 2).

Table 2

Source: The Authors.

The results obtained show that the grass performs better on its own than when it is mixed with the legume. This can be explained by the way that the legume grows, as it is a climbing plant with strong, thin vines that seek support on other substrates or species [20] and can start to crowd the light of neighboring plants.

Two weeks after the forage species were established on the prairie, Canavalia brasiliensis showed greater growth and therefore started to compete with the grasses [19] for space and solar rays which made their photosynthetic process inefficient. This, in the end, was reflected in the production of biomass [21] as the progressive change in the quantity and quality of the light manifests in alterations in the growth and development of forage plants, reducing the yield of dry matter [22].

However, the decrease in productivity of the mixed grasses is compensated by the productive contribution of the legumes planted alongside them [23], since these contribute to the improvement of the animals’ diet, upping the protein content and digestibility of the dry matter as feed [24]. Studies carried out at the International Center for Tropical Agriculture (CIAT - Colombia), show that including species such as Canavalia brasiliensis improves protein content by 3% and increases gross energy, as well as decreasing the indigestible neutral detergent fiber and lignin [25]. This is important, as the quality of the forage affects passage rates, the level of consumption and the production of methane, which in turn affect animal productivity [26].

Added to the above, drought tolerance in livestock systems in the dry tropics allows for system stability throughout the year, supplementing the cattle’s diet during periods of drought [27]. Finally, the legume in the pasture contributes organic matter to the soil which improves fertility and nutrient cycling [28].