INTRODUCTION
Methane is an important greenhouse gas (GHG) produced from enteric fermentation of feed/fodder by ruminant animals. The productivity of livestock in the tropical and sub-tropical areas of developing countries is limited by lower nutritional conditions that are characterized by highly lignified, low digestible feed from poor quality, nitrogen (N)-limited native grass pastures and crop residues, or may also suffer from a general lack of feed during drought (Goel & Makkar Reference Goel and Makkar2012). This sub-standard productivity results in high absolute methane emissions resulting in a very high cost of methane emissions per unit of product (Aluwong et al. Reference Aluwong, Wuyep and Allam2011). This is particularly true when straw-based forages are the main ingredient in ruminants’ diets (Bhatta et al. Reference Bhatta, Enishi, Takusari, Higuchi, Nonaka and Kurihara2008, Reference Bhatta, Uyeno, Tajima, Takenaka, Yabumoto, Nonaka, Enishi and Kurihara2009).
During anaerobic digestion, ruminal microbes usually convert major portions of the carbohydrate (CHO) and protein in feeds to useful end-products such as volatile fatty acids and microbial protein, as well as waste products; mainly methane and carbon dioxide (CO2). The pattern and concentration of these end products depends mainly on the chemical components of the diet (i.e. CHO and protein fractions), their digestibility and intake. Fermentation of plant materials containing low amounts of cell walls results in lower methane production (Johnson & Johnson Reference Johnson and Johnson1995), as well as a decrease in the molar proportion of acetate and an increase in the molar proportion of propionate (Widiawati & Thalib Reference Widiawati and Thalib2007). Fermentation of diets containing high amounts of plant cell walls is likely to produce a higher molar proportion of acetate than propionate (Bhatta et al. Reference Bhatta, Enishi, Takusari, Higuchi, Nonaka and Kurihara2008). Methane from enteric fermentation represents a loss of dietary energy in ruminants up to 12% of gross energy intake (McCrabb & Hunter Reference McCrabb and Hunter1999), and depends primarily on the quantity and quality of the diet as it affects rate of ruminal digestion and passage (Van Soest Reference Van Soest1994; Beauchemin et al. Reference Beauchemin, Kreuzer, O'Mara and McAllister2008). Decreased forage digestibility is generally accompanied by decreased forage intake and increased ruminal acetate: propionate ratio, which favours increased methane production per unit forage consumed (McAllister et al. Reference McAllister, Okine, Mathison and Cheng1996). Tamminga (Reference Tamminga, Phillips and Piggins1992) reported a decrease in methane losses [as a proportion of digestible energy (DE)] with increasing N content in fresh grass and this decrease was hypothesized to be linked to its lower fibre content. Protein degradation in vitro has been shown to be associated with lower methane production than fermentation of CHOs (Cone & Van Gelder Reference Cone and Van Gelder1999; Jentsch et al. Reference Jentsch, Schweigel, Weissbach, Scholze, Pitroff and Derno2007), although increasing dietary N concentrations might also stimulate ruminal methanogenesis (Kurihara et al. Reference Kurihara, Magner, Hunter and McCrabb1999). Enteric methane production could be influenced by the nature of CHOs fermented, such as cellulose, hemicelluloses and soluble residues of the diet (Takahashi Reference Takahashi2001; Santoso et al. Reference Santoso, Kume, Nonaka, Kimura, Mizukoshi, Gamo and Takahashi2003). Moss (Reference Moss1994) reported that digestible acid detergent fibre (ADF), cellulose and hemicellulose are important fibre fractions influencing methane production in the rumen. The information on rumen methane output of feeds from tropical region is largely unknown. In vitro experiments could be used to obtain methane production data from diverse feeds/fodder for further use to estimate methane production from ruminants/livestock fed different feeds/fodder or diets. The objective of the present work was to develop a database on methane production for common feed ingredients and diet combinations fed to ruminants so that rations could be formulated with lowest methane emission.
MATERIALS AND METHODS
Samples
The experiment was conducted at the ICAR-National Institute of Animal Nutrition and Physiology, Bengaluru, India. The samples were collected from different parts of Karnataka state, India.
Collection and processing of samples
Samples comprised dry fodder (14 samples), grass (two), tree leaves (five), cultivated grasses (11), cereal by-products (three), cereal grains (five), oilseed meals (eight), compound feed (five) and total mixed ration (TMR, 21). The TMRs were prepared using locally available feedstuffs, mimicking the feeding practices followed in this region.
The dry fodder samples were collected after harvesting their grain. The samples from different regions were pooled by combining equal portions into a representative sample. The local grass and cultivated grass were sampled from three random sites using a 1 m2 quadrat to create three field replicates during the pre-flowering stage. Leaf samples (leaves + fine stem < 6 mm diameter) were collected from three trees to get representative samples. Cereal by-products, cereal grains, oilseed meals and compound feed samples were collected from different stalls in local markets and likewise pooled by combining equal portions. The TMR was formulated in the laboratory by mixing the required ingredients. All the samples were oven-dried at 60 °C for 48 h and then ground to pass through a 1 mm sieve in a Wiley mill. Ground samples were stored for chemical and biochemical analysis.
Chemical analysis
The tree leaf samples were analysed in triplicate for crude protein (CP) (AOAC 1997), neutral detergent fibre (NDF) and ADF (Van Soest et al. Reference Van Soest, Robertson and Lewis1991). The NDF was analysed in samples without sodium sulphite and amylase. Both NDF and ADF were expressed with residual ash. Other samples were analysed according to the standard methods of AOAC (1995) for dry matter (DM; 976·63) and N (984·13). Lignin (sa) was determined by solubilization of cellulose with sulphuric acid in the ADF residue (Van Soest et al. Reference Van Soest, Robertson and Lewis1991).
In vitro incubation
Initial incubations were performed to determine the time to achieve the half-time gas production (t 1/2 time) of the substrate. For this, rumen liquor was collected from two cannulated Holstein Friesian crossbred bulls fed a TMR (160 g/kg CP and 9·0 MJ/kg DM of metabolizable energy) containing finger millet (Elusine coracana) straw and commercial concentrate mixture in 1 : 1 ratio. The rumen liquor, strained through muslin cloth, was pooled and used as the source of inoculum. A total of 200 mg air-equilibrated sample was incubated with 30 ml of buffered rumen inoculum (Menke et al. Reference Menke, Raab, Salewski, Steingass, Fritz and Schneider1979) in 100-ml calibrated syringes and placed in a water bath maintained at 39 °C. The incubations were conducted in triplicate for each sample on two successive days and these incubations were performed three times. Incubations without samples served as the blanks with every set. The difference in composition and activity of the rumen inoculum among incubations, if any, was controlled by parallel incubation of reference concentrate and hay standard from Hohenheim University, Germany as suggested by Menke et al. (Reference Menke, Raab, Salewski, Steingass, Fritz and Schneider1979). The gas volumes were recorded at 2, 4, 6, 8, 10, 12, 24, 36, 48, 72 and 96 h. This data were subjected to a graph pad prism program to determine their potential gas production (PGP, ml/200 mg DM), rate constant (k) and t 1/2 (h) time.
In vitro rumen methane output and in vitro dry matter digestibility
Two sets of samples were incubated simultaneously, each in triplicate. Samples in the first set were incubated with 200 mg substrate and 30 ml buffered rumen fluid, and the second with 500 mg substrate and 40 ml double-strength buffered rumen fluid, under identical conditions as described earlier. Each sample was incubated until its t 1/2 time as determined earlier and total gas volume was recorded and analysed for methane concentration, again as described earlier.
After terminating the incubation of the 500 mg samples by chilling the syringes in an ice bath, the syringe contents were transferred to a spoutless 600 ml beaker. The syringes were washed with neutral detergent (ND) solution (100 ml), boiled for 1 h, filtered, washed and dried to determine their DM digestibility.
Methane estimation
After terminating the incubation, the volume of fermentation gas produced was recorded from visual assessment of the calibrated scale on the syringe. Net gas production was calculated as the difference between the total gas produced and the gas produced in blank syringes (ml gas in sample syringe – ml gas in blank syringe). For methane estimation, 1·0 ml of gas was sampled with an airtight syringe (Hamilton Company, Reno, NV, USA) from the head space of the syringe (having one outlet) using a specialized adopter fitted to the silicon tubing and injected into a Thermo fisher gas chromatograph equipped with thermal conductivity detector and stainless steel column packed with Porapak-Q. The temperatures of injector oven, column oven and detector were 60, 100 and 110 °C, respectively (Kajikawa et al. Reference Kajikawa, Tajima, Mitsumori and Takenaka2007). Before analysis of unknown samples, the gas chromatograph was calibrated with standard known samples of methane and a standard curve was prepared with suitable regression equation. After injection of gas from each unknown sample, the area under the curve of peaks occurring at the same retention time of the methane standard was recorded and methane concentration was calculated from the standard curve by linear regression. Based on the methane percentage estimated in the gas produced, methane production in ml was calculated in each sample [methane volume (ml) = methane % × total gas produced (ml)]. The in vitro rumen methane output (IRMO) was expressed as methane in ml/100 mg digestible DM.
Statistical analysis
Analysis of variance for chemical analysis of nutrient content, fermentation pattern, in vitro dry matter digestibility (IVDMD) and IRMO was carried out by one-way analysis (SAS Institute 2002) using the model Y ij = μ + Fi + E ij , where Y ij represents the individual observations of the variable and F i is the fixed effect of the ith feed ingredient/diet combination (i = 1–10). The overall mean is expressed as μ and E ij is the random error associated with Y ij not accounted in the fixed effect. Significant differences of feed ingredient/diet combination were considered at the P < 0·05 level.
RESULTS
Composition
Crude protein content (g/kg DM) was least in dry fodder (70·1) and highest in oilseed meals (320), whereas it was similar in local grass and tree leaves (90·7). Cultivated grasses, cereal grains and their by-products contained 115 (g/kg DM) CP. The NDF and ADF contents were highest in dry fodder (711 and 459, respectively) and lowest in oil meals (458 and 213, respectively). Tree leaves contained higher (142) acid detergent lignin [ADL (sa)] than dry fodder (66·4) and local grasses (64·2). In TMR, CP and fibre fractions varied with R : C ratio (Table 1).
Table 1. Composition (g/kg DM) and in vitro dry mater digestibility (IVDMD) of feed ingredients and total mixed ration (TMR)
RS-finger millet straw (E. coracana).
* Feed 1: crushed maize 45 parts + soybean meal 27 parts + wheat bran 25 parts + mineral mixture 2 parts + salt 1 part.
† Feed 2: crushed maize 45 parts + peanut extract 27 parts + wheat bran 13 parts + de-oiled rice bran 10 parts + mineral mixture 2 parts + salt 1 part.
‡ Feed 3: commercial concentrate feed.
The IVDMD figures ranged from 0·48 to 0·87, with the lowest digestibility recorded in tree leaves (0·48). The digestibility of dry fodder was higher (0·508) than tree leaves (0·475) but lower than local grasses (0·557). The digestibilities of cereal by-products and compound feeds were similar (0·61), whereas those of oilseed meals (0·69) were lower than cereal grains (0·87). The nutrient composition of the TMR varied with the level of concentrate in the diet.
Fermentation kinetics
Potential gas production (ml/200 mg DM) ranged from 9·76 to 61·3. The PGP of grasses and compound feeds was similar (39·7), whereas it was least in tree leaves (29·8) (Table 2). The rate constant (mg/h) was maximum in compound feed (0·19) followed by oilseed meal (0·08). The rate constant was similar among the other groups of feedstuffs (0·05).
Table 2. Potential gas production (PGP, ml/200 mg DM), rate constant (mg/h), t1/2 (h) and IRMO of feed ingredients and diet combinations
IRMO, in vitro rumen methane output; RS, finger millet straw (E. coracana).
* Feed 1: crushed maize 45 parts + soybean meal 27 parts + wheat bran 25 parts + mineral mixture 2 parts + salt 1 part.
† Feed 2: crushed maize 45 parts + peanut extract 27 parts + wheat bran 13 parts + de-oiled rice bran 10 parts + mineral mixture 2 parts + salt 1 part.
‡ Feed 3: Commercial concentrate feed.
The t 1/2 time ranged from 9·8 to 19·4 h for local grass (Table 2). The t 1/2 time for dry fodder was 16·5 h and 14·0 h for cultivated grasses; values were similar among tree leaves, cereal grains, by-products and compound feeds at 10·5 h.
In vitro rumen methane output
Methane composition of the total gas varied from 9·79 (tree leaves) to 20·2% (local grasses). The IRMO was expressed as ml methane/100 mg truly digested substrate. Among the straws, IRMO varied from 3·88 (Zea mays) fodder to 12·0 (Sorghum vulgare) with a mean of 6·01. It was 4·67 among the grasses (Table 2). The IRMO was lower (1·34) in fruit tree leaves than cultivated grasses (2·83). Among protein and energy sources, IRMO was higher in cereal by-products (5·92) as compared with cereal grains (2·44), oil meals (2·47) and compound feed (1·12). The IRMO was similar (3·5) among TMR, irrespective of the composition of the concentrate mixture. However, it varied with the level of concentrate in the TMR.
DISCUSSION
The main objective of the current study was to assess the IRMO of a range of feeds with contrasting chemical characteristic and nutrient composition. Chemical composition of feeds and forages was influenced by factors such as crop type, variety, fertilizer, stage of harvest and environment. Based on their CP contents, dry fodder and local grasses cannot be fed to ruminants as sole diets without supplementation. Higher contents of lignin (sa) in legume straw than in the cereal forages and grasses were recorded because legumes synthesize more lignin for strength and rigidity of plant walls. Nutrient contents of most of the feedstuffs investigated in the present study were within the range of values reported earlier (Singh et al. Reference Singh, Das, Samanta, Kundu and Sharma2002; Chaurasia et al. Reference Chaurasia, Kundu, Singh and Mishra2006; Bhatta et al. Reference Bhatta, Enishi, Takusari, Higuchi, Nonaka and Kurihara2008). Jung & Allen (Reference Jung and Allen1995) described the plant cell characteristics affecting intake and digestibility of forages in ruminants. Higher digestibility of legume straw than cereal straw and stovers may be attributed to their lower NDF, ADF, cellulose and lignin contents. The higher DM digestibility of legume straw (by 10%) than cereal straw reported earlier by Bhatta et al. (Reference Bhatta, Enishi, Takusari, Higuchi, Nonaka and Kurihara2008) is in agreement with the present findings. Further, DM digestion of forages is highly dependent on structural factors such as the relative proportion of cell types present in the plant tissues and the existence of factors restricting microbial access to walls. The low IVDMD of cereal straw in the present study may be attributed to low microbial activity, due to inadequate protein supply to meet their requirements during incubation. The t 1/2 time of local grass was lower as compared with dry fodder due to higher lignification. Cereal by-products, cereal grains and oil cakes were degraded in similar time frames (similar t 1/2).
Methane concentration and IRMO differed significantly among feedstuffs. Such variation in in vitro methane was recorded mainly from straw and agricultural by-products. Variation in methane production from dry roughage may be attributed to significant differences in NDF and ADF fractions and IVDMD, as recorded in the present study. Klevenhusen et al. (Reference Klevenhusen, Bernasconi, Kreuzer and Soliva2008) recorded greater methane outputs from high starch/sugar rather than high fibre feeds when fermented in vitro in a continuous culture system. This is in agreement with the findings of the present study in which feeds with relatively high proportions of non-structural CHOs gave rise to greater methane output than high-fibre feeds such as straw and stover. Getachew et al. (Reference Getachew, Robinson, DePeters, Taylor, Gisi, Higgginbotham and Riordan2005) reported 16% methane (in forages, concentrate ingredients and by-product feeds), which seems to be comparable with dry fodder, cereal by-products and oil meals, and lower in local grasses, home-made feed and higher than other feedstuffs. Among dry fodder, high IRMO was recorded in S. vulgare and Arachis hypogea. These feedstuffs form the bulk of the roughage component in ruminant feeds in the northern Karnataka state in India. If efforts are to be made to ameliorate enteric methane production, then a proportion of S. vulgare and Arachis hypogea should be replaced with feedstuffs having a higher nutritive value in the diet. The methane concentration and IRMO of cultivated grasses and cereal grains were similar. Boadi et al. (Reference Boadi, Benchaar, Chiquette and Massé2004), Beauchemin et al. (Reference Beauchemin, Kreuzer, O'Mara and McAllister2008) and Navarro-Villa et al. (Reference Navarro-Villa, O'Brien, López, Boland and O'Kiely2011) reported lower methane from legumes than grasses. Navarro-Villa et al. (Reference Navarro-Villa, O'Brien, López, Boland and O'Kiely2011) attributed less methane in legumes v. grasses to less extensive in vitro fermentation of legumes.
The lowest IRMO was recorded in tree leaves, mainly due to the presence of tannin. It is well established that tannin present in tropical leaves significantly reduces methanogenesis. Efforts have been made to screen these leaves for their methane suppression properties, so that they can be incorporated in ruminant diets (Bhatta et al. Reference Bhatta, Saravanan, Baruah and Sampath2012, Reference Bhatta, Saravanan, Baruah, Sampath and Prasad2013a , Reference Bhatta, Saravanan, Baruah, Suresh and Sampath b , Reference Bhatta, Enishi, Yabumoto, Nonaka, Takusari, Higuchi, Tajima, Takenaka and Kurihara c ).
The IRMO of compound feed was higher than oil meals and lower than cereal by-products. This was attributed to the type of samples that were collected at the farm gate level. There are various types of compound feeds available for different categories of animals depending on their milk yield.
Oil meals produced comparatively lower methane for two reasons: firstly, fat and other compounds included in the ether extract fraction are mostly not fermented by rumen microbes, and unsaturated fatty acids in particular are known to inhibit the methanogenic microbial system (Czerkawski et al. Reference Czerkawski, Blaxter and Wainman1966; Demeyer & Van Nevel Reference Demeyer, Van Nevel, McDonald and Warner1975). Hydrogenation of unsaturated fatty acids increases propionate synthesis, inhibits protozoa and cellulolytic bacterial activity, and thereby affects the methane production (Czerkawski et al. Reference Czerkawski, Blaxter and Wainman1966). Also, Roger et al. (Reference Roger, Fonty, Andre and Gouet1992) reported that glycerol released from fat hydrolysis suppresses cellulolytic bacterial activity. Secondly, protein is degraded to ammonium (NH4) in the rumen and it can combine with CO2 resulting in ammonium bicarbonate (Getachew et al. Reference Getachew, Blümmel, Makkar and Becker1998). Therefore, NH4 produced as a result of rumen incubation of high-protein sources such as oilseed meals can be expected to combine with CO2, thereby lowering the availability of this substrate for methane production. Among the oil meals, the lowest IRMO were recorded in Crocus sativus and Sesamum indicum (1·1 ml methane/100 mg truly digested substrate). The lower IRMO of Gossypium spp. was due to the presence of high NDF and ADF components.
Many studies in the past have shown that methane production could be influenced by the nature of CHO digested, such as cellulose, hemicelluloses and soluble residue (Macheboeuf et al. Reference Macheboeuf, Coudert, Bergeault, Lalière and Niderkorn2014). Santoso et al. (Reference Santoso, Mwenya, Sar and Takahashi2007) observed a positive correlation of methane production with increased NDF digestion. In the present study, methane production tended to be lower than that reported elsewhere for different forages. Many studies have reported correlations between chemical constituents and methane production (Santoso & Hariadi Reference Santoso and Hariadi2009; Singh et al. Reference Singh, Kushwaha, Nag, Mishra, Bhattacharya, Gupta and Singh2011). Quality of feed/diet has a major effect on methane production, as VFA concentration and their relative proportions are influenced by the nature and fermentation of CHO (Johnson et al. Reference Johnson, Ward, Ramsey and Kornegay1996). The increment in fibre fractions will have a depressing effect on methane production. The fibre fractions decrease methane production by lowering pH (Bhatta et al. Reference Bhatta, Enishi, Takusari, Higuchi, Nonaka and Kurihara2008). Although an increase in VFA production might be expected as the digestibility of feed increases, this is generally accompanied by a concurrent decrease in in vivo methane output (Johnson & Johnson Reference Johnson and Johnson1995) but an increase in in vitro methanogenesis. This difference in methane output between in vitro and in vivo studies when high VFA concentrations are recorded may reflect the strongly buffered systems used with in vitro assays, preventing the pH from declining to a much greater extent than occurs in the in vivo rumen. Such a decline in pH has been shown to reduce fibre digestibility and reduce the activity of rumen methanogens.
Several attempts have been made to predict methane production by determining the amount of crude nutrients in cattle and sheep (Holter & Young Reference Holter and Young1992; Shibata Reference Shibata, Minami, Mosier and Sass1994) and it is known that crude fibre is an important component in methane production. Miller (Reference Miller1995) reported that feed ingredients rich in crude fibre stimulated some species of microorganism within the cellulolytic-methanogen consortium, which serve to couple the degradation of CHOs with the use of hydrogen gas (H2) for the reduction of CO2 to methane.
CONCLUSIONS
The IRMO of various feeds and diet combinations were investigated. Because a substantial amount of dietary gross energy is lost as methane, knowledge of the methane output from these feedstuffs would help in formulating low methane producing diets for ruminants in tropical regions. The results of the current study established that incorporation of tropical tree leaves in the diet and feeding TMR are potential strategies to reduce enteric methane production in ruminants and thereby help in preventing global warming due to enteric methane.
The financial assistance provided to this work by the Indian Council of Agricultural Research (ICAR), New Delhi, under the ‘Outreach Project on Methane’ is gratefully acknowledged.
