Effect of cutting time on agronomic and nutritional characteristics of nine commercial cultivars of Brachiaria grass compared with Napier grass during establishment under semi-arid conditions in Rwanda

1 Department of Animal Production, Kigal Rwanda Agriculture Board (RAB), P. O. Box 5016 Rwanda. 2 Department of Animal Science and Poultry Science, Earth and Environmental Sciences, University of KwaZulu-Natal, School of Agricultural, Pietermaritzburg Campus, Private Bag X01, Scottsville, 3209, South Africa. 3 International Center for Tropical Agriculture (CIAT), A. A. 6713, Cali, Colombia. 4 United States Department of Agriculture, Agricultural Research Service, Plant Polymer Research Unit, National Center for Agricultural Utilization Research, 1815N, University of St., Peoria, IL 61604, USA.


INTRODUCTION
Intensification of ruminant livestock production is gaining momentum in a number of sub-Saharan African countries due to increasing population pressure and decline in grazing areas (Thornton and Herrero, 2014).Even farmers with access to open grazing land experience frequent feed shortage due to climate change and variability (Chirat et al., 2014).
Feeds and feeding, especially for dairy cows are a critical issue for smallholder farmers in Rwanda, mostly during the dry season (Klapwijk et al., 2014).In the tropics, grasses are the most ecologically reliable and economically justifiable feed resources (Pedreira et al., 2011) because of their morphological characteristics which enable efficient water use, and rapid recovery after periods of drought (Batistoti et al., 2012).Therefore, improving tropical forages is essential for environmental protection and sustainable livestock feed and food futures for livestock and people (Baudron et al., 2015).Brachiaria grasses are among the most important tropical grasses that originated from Africa and improved in Americas through agronomic selection and breeding (Miles et al., 2004).They have demonstrated to be highly productive, nutritive and socially acceptable in Asia and Africa for different livestock production systems (Vendramini et al., 2014).
In tropical areas, the most sensitive attributes to management are metabolisable energy (ME) and crude protein (CP) contents as well as macro and microminerals.In case these attributes are low in forages grown in warm-environment, including Napier grass, they can compromise milk yields of cows fed on these forages (Mutimura et al., 2015).
Results on dry matter production and nutritional quality of some of these grasses have been reported under different cutting regimes of 3, 6, 9 and 12 weeks (Ortega-Gomez et al., 2011).However, the effect of cutting age on nutritional attributes of the mentioned Brachiaria cultivars including commercial hybrids (cv.Mulato and cv.Mulato II) during establishment in semi-arid environments in Rwanda has not been determined.The objective of the current study was to identify the best-bet Brachiaria cultivars and the based on cutting age to optimize nutrient characteristics and productivity in Rwanda.

Site description
A field experiment on evaluation of tropical forage grass cultivars was established on-farm (Field trial was established in October 2013 and data recorded until February 2014) in Bugesera district, in the eastern Province (semi-arid area) of Rwanda.Bugesera district lies between 30°25" E and 2°30" S with an average altitude of 1,400 m.a.s.l.The climate is semi-arid with a long (4 to 5 months) dry season (Figure 1).Annual rainfall ranges between 650 and 900 mm, with the average temperature of the coldest month is lower than 18°C.

Land preparation and experimental design
The field trial included a set of nine The trial was established at on-farm in a randomised complete block design (RCBD) with four replicates.The land used, was before planted to Lablab purpureus, a forage legume.The plot was prepared using a hoe, then the plot was divided into sub-plots of 3×3 m.Grasses were established without fertiliser application using seeds and cuttings (for Napier grass with the spacing of 1 m between plants and rows).
For seeds, the sowing was on continuous rows to the rate of 8 kg ha -1 with spacing of 50 cm between rows.The experimental design was a split-plot where Brachiaria cultivars and Napier grass were the main plots and cutting age of 60, 90 and 120 days after planting (DAP) were subplots.Soil samples were taken and analysed for total nitrogen (N) and soil organic carbon (SOC) contents before planting.Analysis of soil (AOAC, 2006;method ID 984.13) revealed N and SOC contents were 0.3±0.2% and 1.5±0.7%,respectively.

Chemical composition and fibre analyses
Above ground biomass was harvested at each DAP from 1 m 2 quadrat and fresh weight was recorded.Harvested samples of each cutting age were divided into two portions; one portion was dried at 105°C for 24 h to calculate dry matter (DM) contents (AOAC, 1990;method ID 9420.5); and organic matter contents by incineration at 550°C for 8 h (AOAC, 1990;method ID 9420.5).The second portion of these samples was dried at 60°C for 48 h, and then milled to pass through 1 mm screen for subsequent analysis.Crude protein (CP) expressed as 6.25 x Kjeldahl Nitrogen (N kg -1 DM) content in the feed (AOAC, 2006;method ID 984.13) using automated systems (Büchi Labortechnik AG, CH-9230 Flawil 1/Switzerland, Type: K-360).Neutral detergent fibre (NDF) was determined according to Van Soest et al. (1991).

In vitro dry matter digestibility and gas production
In vitro gas production (GP) kinetics was measured using automaticcomputerised gas production systems (Pell and Schofield, 1993).Ground forage samples (1 g) incubated at 39°C with buffered rumen fluid (100 ml) in Duran bottles (250 ml).
Two fistulated cows fed ad libitum veld hay and Lucerne supplement (2 kg day -1 ) provided rumen liquor.The rumen content was macerated in a plastic bucket under CO2 flux.The rumen fluid was squeezed through four layers of cheesecloth.The resultant liquor was transferred into a warm vacuum flask for delivery to the laboratory within 20 minutes.The final inoculum was made by adding the buffer solution (67 ml) and rumen liquor (33 ml) to the sample (1 g) in the Duran bottle (250 ml) under continuous CO2 flux.Pressure readings were recorded at 20 min interval for 72 h.
These bottles were removed and their contents transferred into Beckman bottles for centrifugation (BECKMANTM, JLA-16.250,Max 16000 RPM, S/N 13U5193) at 16,000 rpm for 15 minutes at 4°C.Supernatants were discarded and pellets were quantitatively recovered for DM determination by oven drying to constant weight at 60°C for 72 h (Castells et al., 2012).

Statistical analysis
Cumulative gas volumes were computed for each channel of pressure sensor as the difference between the readings at time (ti) and the initial reading (t0), adjusted for control readings (Blank) at corresponding recording times.To determine kinetics of gas volume production combined models (Schofield et al., 1994; Equation 1) were used.Maximum rate of GP at the point of inflection was calculated from the cumulative gas production while the time taken to produce half of gas volume (T½) was estimated based on Sahoo et al. (2010)-Equation 2. Rumen degradability efficiency factor (DEF) was also calculated based on Ouda and Nsahlai (2009)-Equation 3.
Where PF is a partitioning factor The model was run using NEWAY1.SAS (SAS, 2010) which also estimated asymptotic gas production as proxy indicators for organic matter degradability.Organic matter degradability (OMD) and metabolisable energy (ME) values were estimated from in vitro digestibility using the following equations (Menke et al., 1979;Equation 4 and 5): Where V24 = gas volume (ml) at 24 h; CP: crude protein (%).In vitro apparent digestible dry matter (ivADDM) was calculated as follows (Equation 6): Differences in chemical composition and in biological measures among forage cultivars and cutting ages were statistically examined using the general linear model (Equation 7) procedures of SAS (SAS, 2010).
A relationship was established between PF and DEF; and ME of these grasses, using regression procedures.This is to determine if the ME can be estimated by using PF and DEF.

Chemical composition
DM was different among grass cultivars across cutting age and at the interaction between cutting age and grass cultivars (P<0.001,Table 1).The DM increased up to 90 DAP and declined at 120 DAP.Brachiaria hybrid cv.Mulato II had the highest and Napier grass had the lowest DM contents (Table 1).
A cluster of three cultivars (MG-4, Basilisk and Piatá) had the second highest DM contents which were not different (P>0.05)within the same cluster.Cultivars Marandú and Llanero had the second lowest DM contents but not different (P>0.05) from DM contents in Mulato, Humidicola and Xaraes (Table 1).
Organic matter content also differed (P<0.001)among grasses and between cutting ages.However, the interaction between cutting age and grass cultivars was not significant (P>0.05,Table 1).Organic matter content increased from 60 DAP to 90 DAP and declined at 120 DAP.This trend was consistent in all except three entries (MG-4, Mulato and Napier grass), where OM contents increased with increase in DAP (Table 1).
In addition, CP content differed (P<0.01)among grasses, between cutting ages (P<0.001) as well as the interaction.The CP contents between Napier grass and among two cultivars of Brachiaria were very strong (P<0.001).Cultivars Humidicola and Piata had the highest but similar CP contents to three cultivars of Brachiaria (Marandú, Llanero and Basilisk).Brachiaria hybrid cv.Mulato II had the least but similar (P>0.05)CP contents to Brachiaria hybrid cv.Mulato and Napier grass.The other entries were cultivars with intermediate CP contents (Table 1).
Crude protein contents declined with DAP and this effect was highly significant (P<0.001).However, these responses were dependent on the cultivars and the interaction effect (P<0.05).This interaction effect showed that cv.Humidicola, cv.Llanero, cv.Mulato and Napier grass lost more CP between 60 and 90 than they did between 90 and 120 DAP (Figure 2).Conversely, cv.Basilisk, cv.Marandú, cv.MG-4 and cv.Mulato II, cv.Piatá and cv.Xaraes lost more CP between 90 and 120 DAP than they did between 60 and 90 DAP.
Neutral detergent fibre (NDF) of tested grasses differed (P<0.001)among grass cultivars, across cutting age and at the interaction between cutting age and grass cultivars.The NDF contents differed highly (P<0.001) between Napier grass and among Brachiaria cultivars.However, cv.Llanero, cv.Mulato and cv.Mulato II had the similar (P>0.05)NDF contents and these values were higher than those of Napier grass and other Brachiaria cultivars, followed by cultivar Xaraes (Table 1).Neutral detergent fibre contents also increased with DAP (P<0.001).However, the magnitude of change depended on the cultivar.In one cultivar (Basilisk), NDF content decreased by 7 to 8% with DAP from 60 to 120.There was a slight (2%) decrease in cultivar Llanero between 60 and 90 days, thereafter it increased slightly (9%).
In cultivar MG-4, NDF content decreased by approximately 25% from 60 DAP to 90 DAP and rebounded by 44% from 90 DAP to 120 DAP, which making a net gain of approximately 7% from the NDF content at 60 DAP to 120 DAP.Napier grass, Mulato II, Piatá and Xaraes gained (19 to 30%) large fibre content between 60 and 90 DAP compared to subsequent increases at 120 DAP.Successive increases in fibre contents in other cultivars were small (Table 1).

The ivADDM, OMD and ME of tested grasses
There was no difference (P>0.05) in ivADDM among grass cultivars although the tendency (P=0.051) was very strong at 90 DAP with cultivar Piatá being the highest in ivADDM.However, the DAP (P<0.001) and the effect of the interaction between DAP and genotypes affected ivADDM.Generally, although there was no effect (P>0.05) of grass genotypes the trend showed that ivADDM increased from 60 to 90 DAP, then decreased to 120 DAP.
Organic matter digestibility content differed (P<0.001)among grass genotypes (Table 2).The effect of DAP was also high (P<0.001)whereas the effect of the interaction between DAP and grass genotypes was evident (P<0.05).In all grasses except cv.Humidicola, OMD increased from 60 to 90 DAP and substantially decreased at 120 DAP.At 90 DAP cv.Piatá had the highest OMD but similar to other grasses except cv.Xaraes and cv.Marandú.
In addition, ME of tested grasses differed among grasses and among DAP (P<0.001).Also, the effect of interaction between DAP and grass genotypes was evident (P<0.05;Table 2).At 90 DAP, most grasses had similar ME except cv.Piatá and cv.Xaraes which had the highest and lowest levels, respectively.Cumulative GP was different (P<0.01)among grasses and highly (P<0.001)among DAP (Table 3); however, interaction between grass cultivars and DAP was not different (P>0.05).The GP of grasses cut at 90 DAP was higher than at 60 and 120 DAP.Degradability showed similar (P>0.05)quickly degradable fraction (A) among tested grasses but differed (P<0.001)among different DAP.Thus, grass harvested at 60 and 90 DAP had highest A compared to grass cut at 120 DAP.

DISCUSSION
Dry matter content in most tested forage grasses increased from 60 to 90 DAP and declined at 120 DAP for cv.Humidicola, cv.Mulato, cv.Mulato II and Napier grass.At 90 DAP where grasses had higher DM contents, cv.MG-4, cv.Basilisk and cv.Mulato II had greater values.
Previous researchers have reported that DM content is much influenced by genetic makeup of the plant, weather and postharvest handling.For example, the DM of Napier grass obtained at 60 DAP was much lower than that reported in the same grass at the same growth stage (Zetina-Córdoba et al., 2013).This difference might be due to the fact that the grass in Thailand was grown with fertiliser application (NPK, 15-15-15; kg ha -1 ) whereas in this study, tested grasses were established without any fertiliser application.The DM content observed at 90 DAP in this study was similar to that reported on forage cereal crops in eastern China (Qu et al., 2014).
The DM of tested grasses revealed that a good time for high DM concentration was at 90 DAP.For warm-season grass, temperatures below 15 o C can decrease growth of the grass (Moreno et al., 2014).Temperatures throughout our experiment were above 15 o C. For this reason, high DM obtained at 90 DAP might be influenced by the low moisture as the samples at this age were collected during dry season compared to other DAP.However, Napier grass which is the most popular forage cultivar used by farmers in east Africa including Rwanda to feed cattle (Klapwijk et al., 2014) had the lowest DM at 90 DAP.This suggests that tested Brachiaria cultivars might offer more advantages on nutritional characteristics than Napier grass at 90 DAP.
Increase of CP level in a feed should come from conventional feed resources (Baluch-Gharaei et al., 2015).The CP content of most grasses decreased with   -1 DM and were much higher than the CP content (109 g kg -1 DM) reported from Brazil in Brachiaria brizantha when the grass was intercropped with soybean (Crusciol et al., 2014).These results obtained for CP content on cv.Piatá, cv.Marandú and cv.Xaraes were also higher than those reported in Brazil when these grasses were subjected to cutting heights of 10, 20 and 30 cm above the soil (de Pinho Costa et al., 2014).However, results from this study were similar to those of Maia et al. (2014) where inorganic fertiliser (nitrogen and phosphorus) was applied to same grasses after corn was harvested.In addition, the CP contents in cv.Marandú were much higher (170 g kg -1 DM) than those reported in Brazil where cv.Marandú (66 g kg -1 DM) was harvested at 60 DAP and fed to steers (Morais et al., 2011).
This might be due to the soil structure, management practices and weather conditions which are major factors that influence nutritional quality of grasses.The CP values obtained in Napier grass were much higher than CP values (112 g kg -1 DM) of the grass reported in Taiwan at 60 days of growth (Zetina-Córdoba et al., 2013).
In addition, at 60 days of age, cv.Humidicola had higher CP (211 g kg -1 DM) than the rest of these grasses.This could be due to its low germination rates and slower growth under cooler environment (Meena et al., 2014) which might influence its CP in leaves at 60 DAP.At 90 DAP cv.Piatá and cv.Basilisk had high values of CP of 166 and 167 g kg -1 DM, respectively.This could satisfy the daily CP requirement of a lactating cow producing 20 to 30 L of milk per day (NRC, 2001).Our findings suggest that tested Brachiaria grasses, in the short run, can be a good source of CP to cattle without any fertiliser application in local farm prevailing conditions.
The NDF content in feed is one of the major criteria to predict DM intake (DMI) in animal, especially for grazing animals.This is because high NDF content in a feed leads animal to eat less feed (Lardner et al., 2015) and hence affects animal productivity.The NDF content in tested grasses increased with DAP.Cultivars Llanero, Mulato and Mulato II showed higher NDF contents than the rest of these grasses across the three DAP.The NDF content observed in most grasses within each DAP was much lower than values reported in other grasses like Lilum sp.(Fukushima et al., 2015).
The NDF reported in Taiwan on Napier grass was much higher (710 g kg -1 DM; Zetina-Córdoba et al., 2013) when compared to this study results.Cultivars Piatá, MG-4, Xaraes and Marandú showed lower NDF content at all three DAP than values reported in Brazil when these grasses were harvested during four seasons of the year (de Pinho Costa et al., 2014).
When comparing cutting ages, NDF was much higher (385 g kg -1 DM) at 120 DAP, however, these values are in the range of 300 to 400 g kg -1 DM which is the recommended NDF content in feed for good DMI by ruminant livestock (McDonald et al., 2011).Furthermore, OM content of forage grasses used in this study increased from 60 to 90 DAP.However, except cv.Basilisk, cv.Humidicola, cv.Llanero and cv.Mulato II, OM content of the rest of the tested grasses declined at 120 DAP.
Generally, the mean OM content of tested grasses at 90 DAP was higher (898 g kg -1 DM) than that of 60 DAP (866 g kg -1 DM) and 120 DAP (893 g kg -1 DM).These values of OM content in tested grasses were similar to those reported in meadow grasses (91.3%) in Armenia (c) (Khachatur, 2006).However, OM values of Napier grass were similar (891 g kg -1 DM) to those reported in Taiwan on Napier grass at the age of 60 days of growth (Zetina-Córdoba et al., 2013).
Dry matter digestibility is one of many factors influencing animal productivity (Mathison et al., 1995).This is also influenced by the availability of the degradable materials of the feed (Campos et al., 2013).The ivADDM of tested grasses differed at 90 and 120 DAP.High ivADDM of these grasses was obtained at 90 DAP.This might be due to the high DM content at this age of harvest.At 120 DAP, ivADDM decreased except for cv.MG-4, Napier grass and cv.Mulato II.Most grasses showed low ivADDM with values below 50% except for cv.Piatá, cv.Humidicola, cv.Basilisk and cv.Mulato which had 564; 550; 524 and 508 g kg -1 DM, respectively at 90 DAP.
Similar results were reported on Saccharum officinarum and Panicum maximum by Singh et al. (2012), but these were lower than results reported in Brazil using the same Brachiaria cultivars (Maia et al., 2014).Our results on ivADDM from Napier grass were lower than reported by Singh et al. (2014).
In addition, OMD increased with cutting age of these grasses until 90 DAP and declined at 120 DAP.These same grasses had high OMD at 90 DAP.The high OMD in cv.Piatá, cv.Basilisk and cv.MG-4 could be explained by their high CP contents (Sampaio et al., 2010) at 90 DAP with reasonable values of NDF content in grasses.Other researchers have reported high OMD (>64%) in cv.Mulato II and cv.Cayman (Vendramini et al., 2014).
The OMD in cv.Humidicola was higher than that reported by Nogueira Filho et al. (2000) in the same grass but lower than in Napier grass.Furthermore, ME in tested grasses at 60 and 90 DAP did not differ.However, high ME (9 MJ kg -1 DM) was obtained in cv.Piatá cut at 90 DAP followed by cv.Basilisk, cv.MG-4 and cv.Mulato II.The ME values observed in these grasses were higher than those reported on available forage grasses in smallholder farms in Rwanda (Mutimura et al., 2015).
Instead, these grasses showed similar ME content compared to some temperate grass cultivars (Fulkerson et al., 2007).The Napier grass which is considered as control had similar ME content reported by latter authors.Although there was variation in ME content among tested grasses, some of these grasses might not satisfy the ME requirement for dairy cow with live weight of 450 kg producing 20 to 30 L of milk per day (NRC, 2001).Nevertheless, cv.Piatá can meet the daily ME requirement for dairy cow of 650 kg live weight producing 16 L per day (Geraghty et al., 2010) if it can eat 17 kg of DM per day.Moreover, grasses with ME above 7 MJ kg -1 DM might be better to supply energy to ruminant livestock based on dairy dry matter intake (Datt et al., 2008).
Gas production parameters of a feed are crucial factors in animal nutrition because they can be used to predict dry matter intake by the animal.Gas production (GP), rate of degradability and half time (T 1/2 ) of grasses were higher at 90 than at 60 and 120 DAP.Cultivar Piatá had the highest GP at 90 DAP followed by Mulato.This might be due to high DM content in the grass at 90 DAP.The B of cv.Piatá was also higher than that of other grasses which revealed that this grass cultivar had a high degradable fraction.Interestingly, the rate of degradation in cv.Piatá was also high although similar to cv.Basilisk and cv.Humidicola.The high rate of degradation might be influenced by the high energy content of these grasses.Negrão et al. (2014) reported that the increase of rate of degradation in cv.Basilisk was influenced by increasing levels of rice bran as source of energy.The rate of degradation of cv.Mulato II and cv.Basilisk was much lower than that reported in previous research on Napier grass (Mutimura et al., 2015).The time taken to produce half of gas volume (T 1/2 ) suggests that cv.Piatá, cv.Mulato, cv.Mulato II, cv.MG-4 and cv.Basilisk might serve as a good source of forage which can increase DMI by the animal.Furthermore, although Napier grass had reasonable degradable fractions, it required a longer time (26 h) to be degraded than the other grasses.This means that at this growth age the DM intake of Napier grass by a ruminant livestock might be reduced due to the extended length of time taken in the rumen (Negrão et al., 2014).
The DEF which is influenced by the time to produce half of gas volume, was correlated with ME.The ME increased with increase of PF as well as with the increase of DEF.Similar trend was observed by Ouda and Nsahlai (2009) who reported the increase of DEF of grass hay when supplementation ratio of legumes was increasing.As the DEF is a proportional of PF and T 1/2 , forages with small values of T 1/2 will have high values of DEF.Also, when T 1/2 remains constant, DEF increases with increase of PF.This means that high rumen degradability depicts microbial efficiency and high values of ME in a forage grass indicating that the ME can be estimated using DEF and PF.

Conclusion
Among 10 grasses tested there were significant differences in terms of nutritional characteristics across the cutting ages.Most grasses had slightly similar nutritive values but cv.Marandú, cv.Basilisk and cv.Piatá were superior in nutritional attributes compared to the rest of these grasses.This is because their DM contents were not higher but their ivADDM, GP, potential degradable fractions, DEF, ME and time taken for rumen degradability values were superior to the rest.Napier grass as the major feed resource of smallholder farmers in east Africa was found to be among the lowest in its nutritive value attributes among the tested grasses.The ME contents decreased from 90 to 120 DAP.The most promising Brachiaria cultivars identified were B. decumbens cv.Basilisk, B. brizantha cv.Marandú and B. brizantha cv.Piatá, because of their nutritional characteristics which were higher than the other cultivars.Although these Brachiaria grasses have shown good nutritional quality, evaluation of their effect on livestock performance is of great importance.

Figure 2 .
Figure 2. Effect of the age of the plant on losses in CP contents in different cultivars of Brachiaria and Napier grass.

Figure 3 .
Figure 3. Relationship between ME and PF (a), between ME and DEF (b) and between DEF and PF (c) of tested grass cultivars.

Table 1 .
Mean dry matter (DM), OM, CP and NDF (g kg -1 DM) of Napier grass and Brachiaria cultivars when harvested at 60, 90 and 120 days after planting in semi-arid zone.

Table 2 .
Mean values of ivADDM, OMD and ME for the tested grasses at different days after planting.

Table 3 .
In vitro digestion parameters of experimental grasses cut at 60, 90 and 120 days after planting.