BROMATOLOGICAL CHARACTERISTICS OF Macrotyloma axillare AS A FUNCTION OF COMBINED SUPPLY OF PHOSPHORUS AND CALCIUM CARACTERÍSTICAS

: The supply of nutrients to forage legumes is significant for their nutritional quality. The study aimed to evaluate the bromatological characteristics of the legume Macrotyloma axillare in response to the combined application of phosphorus and calcium. The study was conducted in a greenhouse on a Red Yellow Ultisol. A fractionated 5 2 factorial was used, with 13 different combinations of phosphorus and calcium application rates (P-Ca): 0-0; 0-20; 0-40; 100-10; 100-30; 200-0; 200-20; 200-40; 300-10; 300-30; 400-0; 400-20; and 400-40 mg dm -3 . Experimental units were set in a randomized block design, with four replications. Neutral detergent fiber (NDF), acid detergent fiber (ADF), hemicellulose, cellulose, lignin, and crude protein (CP) in the leaves and branches were determined in two harvests. In both harvests of the legume we observed that the ADF, hemicellulose, cellulose, and lignin contents in the leaves did not show isolated or combined responses to phosphorus and calcium rates. It was concluded that the combined supply of phosphorus and calcium did not interfere in the nutritional quality of the legume. Calcium supplied alone interfered only the NDF content and phosphorus supplied of the CP of the leaves. Phosphorus rates of 31.5 mg dm -3 and 28.2 mg dm -3 were responsible for the lower NDF content in the leaves and the higher hemicellulose content in the branches, respectively.


INTRODUCTION
In tropical areas, a major challenge faced by cattle farmers is providing adequate nutrition for their animals during the dry season, when pasture, grain residues, and corn husk have poor nutritional quality. An alternative means of improving animal feeding during this period is the use of forage legumes (BARROS-RODRÍGUEZ et al., 2014). The use of forage legumes in pasture has been recommended because of their ability to fix atmospheric nitrogen, allowing an improved availability of nitrogen for plants in consortium or in succession (ROSCHER et al., 2011). In this context, the inclusion of legumes in pasture can improve the biomass production quality and provide forage support for a long period (LÜSCHER et al., 2014).
Among the forage legumes with potential use in animal production systems in tropical countries there is Macrotyloma axillare, a species from South African that is very tolerant to drought and it can survive in situations for up to 8 months in situations with of very dry winter-spring weather and during hot and dry summer periods (BLUMENTHAL; STAPLES, 1993). However, in order to make this legume suitable for animal production systems, its qualitative evaluation is needed to determine whether it is capable of providing a diet which is balanced and rich in protein and energy, as well as to identify its potential toxicity potential for animals (CHAUDHRY, 2008).
In the qualitative evaluation of forage legumes, studying the crude protein (CP), neutral detergent fiber (NDF), and acid detergent fiber (ADF) contents is key, as these parameters can influence legume consumption by animal species (OLIVO et al., 2009). Hemicellulose is part of the NDF and cellulose represents the bulk of ADF (CAMPOS et al., 2016). Lignin is an amorphous structural component, which seems to have a 'cementing' function. Lignin combined with cellulose and hemicellulose forms a ligno-cellulosic complex, making those carbohydrates unavailable for degradation by microorganisms (SARKAR et al., 2009;CHEN, 2014).
The use of forage legumes without prior study can result in low animal production, due to lack of appropriate indications regarding plant nutrition for maximizing animal performance (SITA et al., 2017). From this perspective, an adequate supply of nutrients to forage legumes is of great importance for their quality. Applied individually, phosphorus and calcium interfere with nodulation of forage legumes, root growth, senescence, and foliar abscission, and consequently with their growth (SULIEMAN et al., 2013;WHITE;BROADLEY, 2003;YANG et al., 2012).
In legumes, the total composition of free amino acids in the phloem is affected by phosphorus deficiency. Additionally, phosphorus deficiency has a strong influence on the nodule carbon metabolites content (SULIEMAN et al., 2013). Calcium is required for various structural roles in the cell wall and membranes is a counter cation for inorganic and organic anions in the vacuole, and the cytosolic Ca 2+ concentration ([Ca 2+ ]cyt) is an obligate intracellular messenger coordinating responses to numerous developmental cues and environmental challenges (WHITE;BROADLEY, 2003;YANG et al., 2012). However, there is a lack of studies on the interference between these nutrients when used in combination in terms of providing forage legumes of adequate nutritional quality.
Given the above, this study evaluated the bromatological characteristics of the Macrotyloma axillare legume in response to the combined application of phosphorus and calcium.

MATERIAL AND METHODS
The experiment was carried out in a greenhouse, located in Nova Odessa, São Paulo, Brazil, during the period from August to December 2015, in a Red Yellow Ultisol (SANTOS et al., 2006) and with the forage species Macrotyloma axillare, access NO279 (Macrotyloma legume). The study was conducted under controlled conditions in order to avoid interference of climatic conditions and to guarantee the adequate water supply and temperature for plant growth. The average temperature in the greenhouse varied from 32.54 at 37.73 o C.
Five application rates for phosphorus (0, 15, 30, 45, and 60 mg dm -3 of P 2 O 5 ) were used and five application rates for calcium (0, 20, 40, 60, and 80 mg dm -3 of Ca 2+ ) combined in a surface response study based on a central experimental design modified through fractionated factorial 5 2 , based on Littell;Mott (1975) (Figure 1). Thus, 13 combinations of different application rates of phosphorus (P 2 O 5 ) and calcium (Ca 2+ ) were used (in mg dm -3 ): 0-0; 0-40; 0-80; 15-20; 15-60; 30-0; 30-40; 30-80; 45-20; 45-60; 60-0; 60-40; and 60-80. The experimental design was arranged in randomized blocks, with four replications. After the soil incubation period, thirty seeds of the Macrotyloma legume were sown per pot. Soon after sowing the micronutrient molybdenum and combined rates of phosphorus and calcium that corresponded to the treatments were applied. In treatments without calcium and phosphorus only distilled water was provided. Phosphorus was supplied as sodium phosphate (NaH 2 PO 4 ) and calcium as calcium chloride (CaCl 2  Fourteen days after sowing the Macrotyloma legume, the plants were thinned until seven plants remained per pot. Two harvests of the plants were performed, the first 53 days after sowing the legume and the second 34 days after the first harvest. The plants were harvested at a height of 5 cm from soil surface. Soon after the first harvest, the treatments were reapplied, as well as the other nutrients previously reported. All the material harvested was oven-dried using forced air circulation at 65°C until constant weight was achieved. Dried leaves and branches were ground separately in a Wiley mill for evaluation of the bromatological characteristics: neutral detergent fiber assayed without a heat-stable amylase and expressed inclusive of residual ash (NDF; VAN SOEST et al.,1991), acid detergent fiber expressed inclusive of residual ash (ADF; VAN SOEST et al.,1991;CAMPOS et al., 2016); lignin (ROBERTSON; VAN SOEST, 1981); hemicellulose (hemicellulose= NDF -ADF; CAMPOS et al., 2016), cellulose (cellulose= ADFlignin; Campos et al., 2016) and crude protein (CP: N (g kg -1 ) × 6.25; JONES, 1931).
After chemical evaluations, all data were subjected to statistical analysis, with response surface analysis using the Statistical Analysis System (SAS, 2000), and P ≤ 0.05 indicating statistical significance. First analysis of variance was performed with the combinations of P and Ca at different application rates, which were significant in the F-test for the P ×Ca interactions; polynomial regression (response surface) was performed using the RSREG procedure. When this interaction was not significant, a first and second degree regression study was performed using the GLM procedure for the effect of phosphorus and calcium rates alone.

Evaluating
the bromatological characteristics in the first and second harvests of the Macrotyloma legume, it was observed that ADF, hemicellulose, cellulose, and lignin contents of the leaves did not show isolated or combined responses to phosphorus and calcium rates. In the first harvest, the average ADF, hemicellulose, cellulose, and lignin content of the leaves of the Macrotyloma legume were 172 g kg -1 of DM, 167 g kg -1 of DM, 112 g kg -1 of DM, and 46 g kg -1 of DM, respectively. In the second harvest, the average ADF, hemicellulose, cellulose, and lignin content of in the leaves were 173 g kg -1 of DM, 125 g kg -1 of DM, 114 g kg -1 of DM, and 55 g kg -1 of DM, respectively.
The ADF, NDF, and lignin content in the case of branches of the Macrotyloma legume did not show isolated or combined responses to phosphorus and calcium rates in the first and second harvests. In the first harvest, the average ADF, NDF, and lignin content of the branches were 477 g kg -1 of DM, 623 g kg -1 of DM, and 88 g kg -1 of DM, respectively. In the second harvest, the average ADF, NDF, and lignin content of the branches were 538 g kg -1 of DM, 689 g kg -1 of DM, and 97 g kg -1 of DM, respectively. The lack of response of the bromatological characteristics to phosphorus and calcium rates in the second harvest may be related to the harvest interval in this study.
The NDF content of the leaves showed an isolated response to phosphorus and calcium rates only in the first harvest of the Macrotyloma legume. With regard the phosphorus rates, the results of NDF in the leaves were adjusted to the quadratic regression model, and the phosphorus rate at 31.47 mg dm -3 was responsible for the lowest NDF content (319 g kg -1 of DM) (Figure 2A). In the case of calcium the NDF content of the leaves was adjusted to the linear regression model, with higher NDF content as the calcium application rate increased ( Figure 2B). The NDF content of the leaves as a function of increased calcium application rates ranged from 317 to 337 g kg -1 of DM.
The hemicellulose content of the branches in both harvests and the cellulose content of the branches only in the second harvest responded to different the phosphorus application rates. In the first harvest, the hemicellulose content of the branches was adjusted to the linear regression model with decrease in this content of 5.96% (151 to 142 g kg -1 of DM) as the phosphorus rates increase ( Figure  3A).
In the second harvest, the hemicellulose content is adjusted to the quadratic regression model and the phosphorus rate at 26,9 mg dm -3 revealed a higher value (156 g kg -1 of DM) ( Figure 3B). The hemicellulose corresponds to the cell wall fraction that is most easily digested by the animals.
With regard to the cellulose content of the branches in the second harvest, it was observed that, as the phosphorus application rates increased, there was an increase in these values, ranging from 425 to 444 g kg -1 of DM depending on the increase in the phosphorus rates ( Figure 3C).  Regarding the CP content of the leaves and branches, responses to phosphorus rates were observed only in the first harvest. In the leaves, the CP content in the first harvest was adjusted to the linear regression model, with a reduction in these values as the phosphorus rates increased ( Figure  4A). The CP content ranged from 172 to 132 g kg -1 of DM as a result of increased phosphorus application rates. Although there were negative changes in CP content depending on the phosphorus Y = 335.8132 -1.1169*P + 0.01784*P 2 (R 2 = 0.83) supply, the values observed could meet the daily CP requirements of animals. In the branches, the CP content in the first harvest was adjusted to the quadratic regression model, and the phosphorus rate at 39.6 mg dm -3 showed a lower CP content (88 g kg -1 of DM) ( Figure 4B). Based on these results, it was found that the CP content of the leaves was higher compared to the branches. The consumption of branches by ruminants is generally lower than that of leaves, due to their lower nutritional value.

DISCUSSION
The NDF content of the leaves of the Macrotyloma legume depended on phosphorus application rates and indicated the requirement for sufficient amounts of phosphorus in the soil (TÜRK et al., 2009), since feed consumption is negatively correlated with NDF, i.e. the slower the degradation, the longer the forage remains in the rumen and consequently, limits the intake of forage (DU et al., 2016). Calcium negatively interfered with the NDF content of legume leaves, since high NDF content can affect feed intake by ruminants because it limits feed intake as a result its slow fractional degradation rate and its low passage rate through the ruminal environment, something which can compromise the productive performance of animals (MONÇÃO et al., 2016). The NDF content is related to consumption modulation, energety density of food, mastication, passage rate, and digestibility of food ( VAN SOEST, 1991;EDOUARD et al., 2008). Besides, FDN and ADF fractions are the least digestible fraction of plants and, in the case of plant leaves, the optimal situation would be having low levels of these fractions (TEIXEIRA et al., 2010).
The NDF content of the leaves of the Macrotyloma legume that were reported in this study, both in relation to the isolated phosphorus Y = 7.9130 -0.0570P +0.00072P 2 (R 2 = 0.80) supply and in relation to the isolated calcium supply, are below the levels observed by Teixeira et al. (2010)  Hemicellulose and cellulose levels in the Macrotyloma legume, depending on the phosphorus application rates, behaved in an opposite way to what is preferred for feeding ruminants, since higher hemicellulose content and lower cellulose content are interesting, considering that bacterial flora into rumen break these components into short chain fatty acids (SCFAs), mainly the acetic, propionic, and butyric acid, which represent the most important source of energy in animals that are fed a forage based diet (PARIZ et al., 2011).
The hemicellulose content observed in the branches of the Macrotyloma legume in this study in relation to the isolated supply of phosphorus is below the levels observed by Teixeira et al. (2010) for the legumes Arachis pintoi, Calopogonium muconoides, Clitoria ternatea, Desmodium heterocarpon, Pueraria phaseoloides and Stylosanthes guianensis cv. Bandeirante, cv. Cook and cv. Mineirão. In the second harvest, the average values of cellulose in the branches were three times higher than the average hemicellulose content, being even higher than those observed by Teixeira et al. (2010) for the legumes Arachis pintoi, Calopogonium muconoides, Desmodium hetero-carpon, Pueraria phaseoloides, Stylosanthes guianensis cv.
Bandeirante, cv. Cook and cv. Mineirão and Stylosanthes macrocephala cv. Pioneiro. The CP content of the leaves and branches of the Macrotyloma legume depending on the phosphorus rates of application, would not limit animal production. PB content below 70 g kg -1 of DM limits animal production, as it promotes decreased voluntary consumption, lower digestibility coefficients, and negative nitrogen balance (KOSCHECK et al., 2013). The high CP content in foraging species favours voluntary consumption by the animals (EVITAYANI et al., 2004).

CONCLUSIONS
The combined supply of phosphorus and calcium did not influence the nutritional quality of the Macrotyloma legume.
Calcium supply alone influenced only the NDF content of the leaves and in the case of phosphorus the CP content of the leaves.
Phosphorus application rates of 31.5 mg dm -3 and 28.2 mg dm -3 were responsible for lower NDF content of the leaves and higher hemicellulose content of the branches, respectively.

ACKNOWLEDGEMENTS
This work was financially supported by the São Paulo Research Foundation (FAPESP -grant number: 2014/22866-2 "Responses of the Macrotyloma legume to rate combinations of phosphorus and calcium").