Forage mass, botanical composition and stocking density of bermudagrass overseeded with forage legumes

1 Submitted on June 18 , 2015 and accepted on September 15 , 2017. 1 This research is part of the first author master’s thesis. 2 Universidade Federal de Santa Maria, Programa de Pós-Graduação em Zootecnia, Santa Maria, Rio Grande do Sul, Brazil. assiszoot@yahoo.com.br; priscilafloresaguirre@yahoo.com.br * 3 Universidade Federal de Santa Maria, Departamento de Zootecnia, Santa Maria, Rio Grande do Sul, Brazil. clairolivo@yahoo.com.br 4 Universidade Federal de Santa Catarina, Departamento de Zootecnia e Desenvolvimento Rural, Florianópolis, Santa Catarina, Brazil. denise.leme@ufsc.br 5 Universidade Federal de Santa Maria, Programa de Pós-Graduação em Zootecnia, Santa Maria, Rio Grande do Sul, Brazil. franciane-ap@hotmail.com; debora.dedi02@hotmail.com *Corresponding author: clairolivo@yahoo.com.br Forage mass, botanical composition and stocking density of bermudagrass overseeded with forage legumes 1


INTRODUCTION
The cultivar coastcross-1 (Cynodon dactylon L. Pers.) has been highlighted for its elevated forage potential, response to fertilization, great adaptability to diverse environments, and flexibility of use as pasture, hay, and silage (Branco et al., 2012).
In most properties, bermudagrass is cultivated alone and demands high levels of chemical fertilizers. The use of fertilizers, especially nitrogen, increases the forage production and the nutritive value. In that practice, production costs are elevated and the distribution of forage is inadequate during the productive cycle. In this context, mixed grasses with forage legumes could improve the productivity of the system, in addition to balancing the allowance of forage, improving the diet of animals and reducing or eliminating the use of nitrogen fertilizer (Azevedo Junior et al., 2012), besides extending the grazing seasons.
Even with this performance, normally expected for grass-legume association, there are few properties using mixed system, due to the slow establishment and low persistence of the legumes in the pastures (Olivo et al., 2014). Also, there is some research to overcome these difficulties. Among forage legumes used for mixed system, the common vetch (Vicia sativa L.) and arrowleaf clover (Trifolium vesiculosum Savi) stand out. These forages feature considerable nitrogen fixation, forage production, and nutritional quality to compose the diet of the animals.
The objective of this study was to compare the performance of forage mass, botanical composition, and stocking density of three pasture-based systems: coastcross-1 and common vetch or arrowleaf clover mixture plus 100 Kg N ha -1 ; and coastcross-1 plus 200 Kg N ha -1 .
The research was conducted from May 2013 to April 2014 (330 days). The monthly temperature and precipitation average were 14.7 to 25.7 ºC and 130.7 mm -1 , respectively. The climate values for the respective period were 18.8 ºC and 134.7 mm -1 , respectively. During the fall, winter, and spring, seven, ten, and two frosts occurred, respectively. Meteorological data were obtained in the Meteorological Station of UFSM, located approximately 500 m from the experimental area. The grazing period was 247 days from August 15, 2013to April 18, 2014 For experimental evaluation, an area of 4691 m 2 subdivided into nine paddocks was used. The treatments were constituted of the following forage systems: bermudagrass (Cynodon dactylon L. Pers.), cv. coastcross-1 + 100 kg N ha -1 year -1 + common vetch (Vicia sativa L. cv. Comum); coastcross-1 + 100 kg N ha -1 year -1 + arrowleaf clover (Trifolium vesiculosum Savi), cv. Yuchi; and coastcross-1 + 200 kg N ha -1 year -1 . The coastcross-1 of the area had already been manually deployed, using seedlings from the subdivision of the clumps, derived from an experimental area of the Dairy Cattle Laboratory. In three paddocks, the common vetch was overseeded on coastcross-1(mid-May.) by direct mechanized plantation, with seeding density of 60 kg ha -1 , pre-inoculated, with spacing planting of 17 cm between rows. In other three paddocks, in the same period, haul of the arrowleaf clover was distributed, being scarified and pre-inoculated using density of 10 kg seeds ha -1 . In all experimental areas, the development of natural annual ryegrass Lolium multiflorum L. Lam reseeding was allowed.
According to the soil analysis of perennial pastures in summer cycle, 90 kg ha -1 year -1 were applied as base fertilizer by using P 2 O 5 and K 2 O. Nitrogen fertilization was subdivided into six applications, being the first held before grazing started and after one, four, six, eight, and ten grazing cycles. On November 27, after the 6 th grazing, mowing of the experimental areas was performed to uniformize the stand areas.
The grazing method used was the rotational stocking, with one day of occupation and the criterion for the start of animal grazing was the sward heights of the various forage systems in the pasture. In the winter period, it was based on the height of the legumes (about 30 cm) and in the summer, it was the height of coastcross-1, close to 25 cm for all areas.
The forage allowance was 6 kg DM (dry matter) for each 100 kg of body weight throughout the year. For analysis, lactating Holstein cows were used, with an average weight of 567 kg and an average production of 19.3 kg milk/day, subjected to two daily milking sections at 07:30 and 17:00 h. After milking, the cows received supplementation concentrate, at 0.9% of body weight, considering the period of lactation and mean milk production, with ad libitum mineralized salt and water. The animals were kept on season grass pastures when they were not in the experimental areas.
At pre-and post-grazing, the forage mass was estimated by double sampling technique (Wilm et al., 1944) Rev. Ceres, Viçosa, v. 64, n.5, p. 500-506, set/out, 2017 with twenty visual estimates and five cuts close to the ground per paddock. The forage from the destructive samples was weighed and then homogenized. A subsample was removed for estimating botanical composition of the pasture and structural composition of the coastcross-1. These structural and botanical components were dried in forced-air oven at 55 °C until reaching the constant weight for estimating dry matter, which was used to calculate the percentage of each fraction. For stocking rate, the value of instantaneous stocking density was divided by the number of the days of each grazing cycle. The subtraction of the pre-and post-grazing forage mass was calculated for the disappeared forage mass.
Data on parameters of grazing cycles in every season were used for statistical analysis, considering 2, 5, 4, and 2 grazing cycles in winter, spring, summer, and autumn, respectively. The experimental design was completely randomized with three treatments (grazing systems) and three replicates (paddocks) in completely split-plot time (grazing cycles). Data were subjected to analysis of variance of using PROC MIXED procedure of SAS (2002), using the Tukey's test for comparison of means. Significance was determined at p < 0.05.
The experimental design was approved by the Ethics Committee and Biosafety of the UFSM, the opinion 113/ 2011 under the Protocol 23081016073/2011.

RESULTS AND DISCUSSION
In the experimental period (330 days), thirteen grazing cycles were performed in both systems, as well in the forage system mixed with clover as singular cultivation. In the common vetch mixed system, twelve grazing cycles were performed. The mean interval between grazing was twenty-seven days and the occupation time was one day. Similar results were found by Azevedo Junior et al. (2012) with summer pasture related with better forage quality and animal performance.
For the pasture height in the beginning of the grazing (Table 1), in the evaluations for the winter, values were close to 30 cm (p < 0.05) for the mixed pasture system, which agrees with the methodology indicated. In the pastures constituted by coastcross-1, the height pasture was approximately between 20 and 30 cm in the other seasons (p < 0.05). The utilization of coastcross-1 with sward height around 20 cm was related with 95% luminous interception.
For forage mass (pre-grazing), there was a difference (p < 0.05) between forage systems in winter and in spring, with greater value for mixed system with common vetch. This result is due to the characteristics of this forage legume (Caballero et al., 1995), which is precocious compared with arrowleaf clover, confirmed by the high participation in the botanical composition of the pasture. For the mixed system with arrowleaf clover, although the greatest participation was in the spring, there was no difference between the systems. During this time, there was a large competition between foragers, with greater participation of Coastcross-1. In the summer, the values were similar between forage systems, although the lower fertilization occurred in the mixed system. This result is due to the contribution of the legumes, releasing nitrogen to the system (Sarrantonio, 1992). Similar results were verified by Aguirre et al. (2016) in pastures with Coastcross-1 mixed with legumes of winter cycle fertilized with 100 kg N ha -1 year -1 .
In the autumn, the value of forage mass was higher (p < 0.05) in the system without legumes due to the highest level of nitrogen fertilization, considering that the Coastcross-1 responds to nitrogen fertilization (Humphreys et al., 2012). The lowest value observed in the autumn mixed system was probably due to the decreased residual effect of forage legumes, possibly due to competition between bermudagrass and legume. Also, the forage mass in locals with legumes may be smaller due to its morphological structure, which increase the vertical plane of the forage more than in exclusive grass cultivation, resulting in less forage mass. Thus, even with different morphological structures, Coastcross-1 emits its tillers vertically, which increases its vertical space more than when they are in exclusive cultivation. This probably occurred as a strategy of Coastcross-1 to avoid competition with other species (Barbero et al., 2009).
In the botanical composition of the pasture, the participation of Coastcross-1 was approximately 20% in the winter. In the spring, the lowest value of the Coastcross-1 in the mixed system (p < 0.05) was due to the participation of the forage legumes that usually interferes with the development of the companion grass (Grieu et al., 2001). In the summer and autumn, the Coastcross-1 values were higher. For the annual ryegrass, values were similar in winter and differed between pasture systems in the spring, with lower value in the mixed system (p < 0.05), due to the presence of legumes that also affected the development of the grass. The average participation of forage legumes was 23.7 and 18.8% for common vetch and arrowleaf clover, respectively. It is noteworthy that the presence of arrowleaf clover is longer. According to Roberts (1974), the proportion of forage legumes in the pasture should oscillate between 20 and 40% for the occurrence of positive reflections in animal production. For the other species fraction, it was verified the presence of the spontaneous-growth species (Cynodon spp., Urochloa plantaginea and Paspalum spp.) and the values were higher in the summer for all forage systems; considering that the species are season-warm, the differences found confirm that the presence of legumes implies control of these species in the system. For the dead material fraction, values were higher (p < 0.05), especially in the winter due to the senescence material of the Coastcross-1, due the action of cold and frost.
For sward height of residual forage (post-grazing), it was observed that the height of the forage was lowered at approximately 50% in all seasons of the year (Table 2).
Comparing the seasons, there were differences (p < 0.05), with greater height in the summer and autumn, attributed to the exclusion areas because of the accumulation of feces.
There were no differences between the grazing systems for forage mass (post-grazing) (Table 2) due to the adjustment used for stocking density. Comparing the mean values of forage mass of pre-and post-grazing, the forage used (grazing efficiency) was approximately 48%. This result demonstrated that there was no intake limited per Table 1: Sward height, pre-grazing forage mass and botanical composition of three pasture-based systems (SF) with Coastcross-1 (CC) + common vetch + 100 kg N ha -1 year -1 (CE); CC + arrowleaf clover + 100 kg N ha -1 year -1 (CT) and CC + 200 kg N ha -1 year -1 (CN animal, which normally occurs when the value exceeds 50% (Delagarde et al., 2001). For the botanical composition of residual forage mass, the values of Coastcross-1 were lower than those obtained in the pre-grazing (approximately 32%). This result indicates that the Coastcross-1 was more consumed than other plants, with great participation in the botanical composition of the pasture. For the annual ryegrass and forage legumes, the values were proportionally lower (p < 0.05) due to the greater use of animals, related to the best nutritive value of these forages (Olivo et al., 2014). For other species, the participation value was similar to the initial forage mass value, showing a trend of lower intake due to the composition of this fraction by species of summer cycle. Some plants of the genus Paspalum spp. can be consumed in a similar way to the genus Cynodon spp. due to the similarity of nutritional value. However, as the management was oriented to the pasture of Table 2: Sward height, post-grazing forage mass and botanical composition of three pasture-based systems (SF) with Coastcross-1 (CC) + common vetch + 100 kg N ha -1 year -1 (CE); CC + arrowleaf clover + 100 kg N ha -1 year -1 (CT) and CC + 200 kg N ha -1 year -1 (CN

SF Average CV (%)
Coastcross-1, certainly harming intake of Paspalum spp., Cynodon spp. and Sida santaremnensis were less consumed by animals compared with the pasture with Coastcross-1. For dead material, the values were higher in relation to the pre-grazing, as expected, due to the selection of animals by the green forage mass, especially of leaf blades (Table 3) and the damage caused by trampling. The relation between leaf blade/stem + sheath of Coastcross-1 in the pre-grazing (Table 3) presented no difference between pasture systems, demonstrating that the inclusion of forage legume did not interfere in the structure of the Coastcross-1 plants. Despite the low forage availability in the winter, there was a high value of leaf blades (Table 1). In the other seasons, the values were uniform, indicating that forage can have a strategic importance in the southern region of Brazil, especially in the autumn, period with shortage of forage.
Similarly, for the values of post-grazing, the relation between leaf blade/stem + sheath, there was no interference due to the presence of forage legume. The data of post-grazing were more uniform compared with pre-grazing due to the adjustment of stocking density.
The obtained average value of 0.54 was suitable for the recovery of the pasture, providing high number of grazing cycles.
In the average stocking density (Table 4), the results were related to the initial forage mass, with difference (p < 0.05) between the systems. The higher values were found in the winter and spring for the mixed system with vetch, due to the contribution of this forage legume in forage mass (Table 1). In the summer, there was superiority (p < 0.05) of this mixture system, attributed to the residual effect of this forage legumes to the system, with the degradation of nodules and plant parts (Bordeleau & Prévost, 1994;Butler et al., 2012), implying more forage mass and, consequently, more stocking density. In the fall, the highest stocking density was observed in the pasture without forage legumes, demonstrating that the nitrogen fertilization effect was more effective in this period.
In the mixed system with arrowleaf clover, it was observed that between seasons the presence of this forage legume implied stocking density values similar to the system without forage legume. Considering the period of

SF Average CV (%)
the introduction of forage legumes at the end of pasture usage, the mean of the stocking density of the forage systems was 7 cows ha -1 , ranging from 5.6 in the winter and 9.1 in the summer. In a study of Coastcross-1 subjected to irrigation and fertilization with 300 kg N ha -1 year -1 , using Holstein cows, supplemented with 6 kg of concentrate day -1 , Alvim et al. (1997) observed stocking density of 3.7 cows ha -1 in the dry season and 6.4 cows ha -1 during the rainy season. Scaravelli et al. (2007) found stocking density of 5.05 AU ha -1 , assessing pastures of Coastcross-1, from January to May, with nitrogen fertilization of 80 kg N ha -1 .

CONCLUSION
The inclusion of the common vetch plus 100 Kg N ha -1 resulted in the highest forage mass and stocking density in winter and spring. For Coastcross-1 without forage legume and receiving 200 Kg N ha -1 , the best performance was found in the autumn. The inclusion of the forage legumes did not affect the structural composition of Coastcross-1.