Directory UMM :Data Elmu:jurnal:S:Small Ruminant Research:Vol36.Issue3.2000:

Metabolizable energy of roughage in Taiwan

Mei-Ju Lee

a

, Sen-Yuan Hwang

a

, Peter Wen-Shyg Chiou

b,* a_{Taiwan Livestock Research Institute, Hsin-hua, Tainan, Taiwan}

b_{Department of Animal Science, National Chung-Hsing University, Taichung, Taiwan}

Accepted 27 September 1999

Abstract

The ®xed metabolizable energy (ME) values from the NRC do not represent the true ME values of the various feedstuff used in livestock rations. Therefore, a rapid and effective method for evaluating the ME value of forage crops is required for proper ration formulation to improve production ef®ciency. Dairy goat digestion trials were conducted as the in vivo reference using the method of Menke and Steingass (1988) [Menke, K.H., Steingass, H., 1988. Feed Sci. Technol. 28, 91±97] which derived the amount of gas produced from in vitro fermentation. This method was adapted in this study to evaluate the ME value. In the goat digestion trial, six dairy goats were used for each roughage sample in a total fecal collection trial to determine the digestible nutrients, including energy (DE) and total digestible nutrient (TDN). The in vivo ME value was calculated using the method of Shiemann et al. (1971) [Shiemann, R., Nehring, K., Hoffmann, L., Jentsch, W., Chudy, A., 1971. Energetische Futterbewertung und Energienormen. VEB Deutscher Land-wirtschaftsverlag, Berlin, p. 75. (in German)] (ME1(MJ/kg)ˆ5.2DCP‡34.2DEE‡12.8DCF‡15.9DNFE,

g/g). The in vitro ME value was then estimated from the chemical composition of the feed and amount of gas produced (Gb) from

in vitro fermentation. The value calculated from both with (ME3) and without (ME2) the inclusion of nitrogen free extracts (NFE)

in the prediction equation. (ME2 (MJ/kg)ˆ0.145Gb‡4.12CP‡6.5CP2‡20.6EE‡1.54, g/g; ME3(MJ/kg)ˆ0.118Gb‡

8.75CP‡19.21EE‡3.38NFE‡0.691, g/g). The 12 roughage samples consisted different growth stages of Napier grass Taishi No. 2: (day of harvest; 40, 50, 60 and 65), dwarf Napier grass Taishi No. 1: (Day 40 and 65) and Pangola grass (Day 45) hay (Day 70), corn silage, imported alfalfa hay, timothy hay and Bermuda hay. The correlation between the ME values calculated from in vivo and in vitro without NFE was lower than with NFE inclusion in the equation. A higher correlation between the ME values calculated from in vivo and in vitro without NFE inclusion than with NFE inclusion in the prediction equation was obtained when alfalfa and corn silage were not included. This indicated that the ME value of forage could be estimated rapidly using this in vitro gas method adapted from Menke and Steingass (1988) [Menke, K.H., Steingass, H., 1988. Feed Sci. Technol. 28, 91±97] for practical applications in ration formulation.#2000 Elsevier Science B.V. All rights reserved.

Keywords:Roughage; Metabolizable energy; In vitro method; Gas production

1. Introduction

The nutrient composition including the metaboliz-able energy (ME) value of roughage varies widely

according to different plant genetics, environment and management, i.e. the growing season, fertilizer, irriga-tion, and harvesting stage. At present, forages used in Taiwan is either imported (alfalfa, timothy and Ber-muda) or locally produced (Napier grass, pangola grass, green corn, corn silage, and by-products: brewer's dried grain, peanut vine, soybean vine, soy-bean pod, and soysoy-bean curd residues). The ME content

Small Ruminant Research 36 (2000) 251±259

*_{Corresponding author. Tel.:‡886-4-2870613;}

fax:‡886-4-2860265.

E-mail address: wschiou@dragon.nchu.edu.tw (P.W.-S. Chiou)

of this roughage however, has not been analyzed and the NRC table values are used instead.

Animal feeding faces a bottleneck because this important nutrient-energy can not be properly evalu-ated due to the lack of a rapid and effective method for ME estimation. Estimation of the ME value in vivo using cows or goats is tedious and requires labor and facilities. It is good for research but not for practical applications in feed formulation and feeding improve-ment. A rapid and precise in vitro method to evaluate the ME value for roughage is therefore important for ef®cient dairy production.

The end products of rumen fermentation are micro-bial cells, fermentation acids and gas. Measurement of gas production is a convenient assay of metabolic activity. Net gas production, although does not give a direct value, it is correlated with the extent of digestion (Van Soest, 1982). Application of gas pro-duction from forage fermentation to estimate the ME value was established by Menke et al. (1979) and Menke and Steingass (1988). The gas measuring techniques for assessment of the nutritional quality of feeds has been reviewed by Abreu and Bruno-Soares (1998); Getachew et al. (1998) also applied this technique to estimate the nutritive value of feedgrain. This study is therefore aimed at the appli-cation of gas measurement techniques in establishing a rapid in vitro method for evaluating the ME value of different sources of roughage and thus providing basic information for ration formulation in dairy feeding.

2. Material and methods

2.1. Material

Total of 12 different forage, which included Napier grass, Pangola grass, corn silage, alfalfa hay, timothy hay and Bermuda hay were used in this study. The Napier grass and Pangola grasses used were from different strains and harvesting stages. These grasses included the Taishi #2 Napier grass from Day 40, 50, 60 and 65, the Taishi #1 dwarf Napier grass from Days 40 and 65, green chopped Pangola grass from Day 45 and Pangola grass hay from Day 70. The corn silage was in the milk stage. The alfalfa hay, timothy hay and Bermuda hay were imported from the USA.

2.2. In vivo method

The digestibility trials were conducted on six dairy goats in their dry period. The nutrient digestion coef®cients for each forage sample were measured using the total fecal collection method according to Harries (1970). After a 14 day preliminary period in the metabolic cage with fecal collecting container behind a screen ¯oor which allowed additional separ-ating the feces and urine through the screen ¯oor with funnel underneath for urine collection. During the collection period, forage was cut to 2±9 cm and fed to the dairy goat at 90% the maximum intake in the adaptation period. Premix and water was available all of the time. Feed consumed and feces eliminated were recorded daily. Total feces were collected for seven consecutive days and stored at ÿ18₈C in a freezer. At the end of the trial, the collected fecal samples were mixed and 5% of the collected feces sampled and dried for chemical composition and energy analysis in order to calculate digestibility. Samples of the forages were also collected from the goat digestion trials every day and mixed, dried and ground at the end of the trial.

The total digestible nutrients (TDN) were also calculated from the digestible nutrients. The para-meters included digestible crude protein (DCP), diges-tible crude fat (DEE), digesdiges-tible crude ®ber (DCF) and digestible nitrogen free extract (DNFE). The metabo-lizable energy (ME1) was calculated from the diges-tion coef®cient, which was derived from the in vivo digestion trial using the equation of Shiemann et al. (1971), (ME1ˆ15.2DCP‡34.2DEE‡12.8DCF‡

15.9DNFE; MJ/kg; g/g).

2.3. In vitro method

2.3.1. Incubation facilities

2.3.2. Preparation of rumen liquor

The rumen liquor was taken from the rumen ®stula of a cow that had been fed green chopped Pangola grass ad libitum with 2 kg of concentrate daily for 14 days, and ®ltered through two layers of cheesecloth. Rumen liquor, 200 ml, was placed into a special glass, and then placed into a warm ¯ask ®lled with CO2and mixed with arti®cial rumen ¯uid. The arti®cial rumen ¯uid consisted of (added in order) 400 ml H2O, 0.1 ml solution A, 200 ml solution B, 200 ml solution C, 1 ml resazurine (0.1%, w/v) solution D, and 40 ml reduc-tion solureduc-tion E. This mixture was then kept under CO2 in a 398C water bath and stirred using a magnetic stirrer.

Solution A consisted of 13.2 g CuCl22H2O, 10.0 g MnCl24H2O, 1.0 g CoCl26H2O, 8.0 g FeCl26H2O and made up to 100 ml with water. Solution B con-sisted of 35 g NaHCO3and 4 g NH4HCO3added up to 1000 ml with water. Solution C consisted of 5.7 g Na2HPO4, 6.2 g KH2PO4, 0.6 g MgSO47H2O added up to 1000 ml with water, The solution D consisted of 0.5 g resazurine up to 100 ml with water. The solution E is the reduction solution consisted of 95 ml H2O, 4 ml 1 N-NaOH and 625 mg Na2S9H2O.

2.3.3. Incubation

Thirty milliliter of the rumen ¯uid medium-mixture kept at 39₈C was placed into each special syringe using a pipette with an automatic pump. A sample of forage(0.2 g dry matter) was introduced into the syringe. Simultaneously, 0.2 g of standard forage and a mixture of concentrate and forage, at a ratio of 30±70 was introduced into the syringe separately. The standard forage was ryegrass hay with 12.1% CP and 21.1% CF whereas the concentrate (14.1% CP, 9.6% CF) which consist of 55% corn grain, 15% barley, 10% soybean meal, 10% cottonseed and 10% sun¯ower seeds.

Each sample used six replicates. Each replicate of the sample included one blank tube in the experiment. Thirty milliliter of arti®cial rumen ¯uid and rumen-liquor were added separately. During incubation, any gas bubbles in the syringe were removed, the plastic clip on the silicon tube closed, and the position of the piston recorded and placed in parallel on a disk at 390.15₈C (a disc contained 60 syringes) rotating automatically for a cultivation period of 24 h. The rotor stand had two axles. One rotation per minute was suf®cient for continuous mixing of the contents. Time

and temperature were controlled automatically. All of the positions of the piston were read at 6±8 h intervals to record gas production. Gas production was not expected to exceed 60 ml. If 60 ml was exceeded, the piston was moved back to the 30 ml position. The ®nal reading was taken 24 h after the beginning of incubation. Moreover, the gas production in the blank bottle (Gbo), 0 h (U0), 24 h (U24) and the standard samples were recorded and the net gas production computed [Gbˆ(U24ÿU0ÿGbo)(FH‡FHS)/2; where FH stands for standard hay, FHS for standard con-centrate and hay mixture].

2.3.4. Calculation of metabolizable energy

The ME value was calculated from the amount of gas produced and general chemical composition of the forage according to the method of Menke and Stein-gass (1988). The calculation values included ME2 [ME2ˆ0.145Gb‡4.12CP‡6.5CP2‡20.6EE‡1.54 (MJ/kg; g/g)] and ME3[ME3ˆ0.118Gb‡8.72CP‡

19.21EE‡3.38NFE‡0.691(MJ/kg, g/g)].

2.4. Chemical analysis

Approximate analysis of feed and fecal samples were performed according to the methods of the Association of Of®cial Analytical Chemists (1980) (AOAC, 1980). The gross energy was analyzed using an Oxygen bomb calorimeter (Parr 1241 Adiabatic Calorimeter, Parr Instrument Co., USA).

2.5. Statistical analysis

Analysis of the variance and covariance were cal-culated with the general linear model procedure (GLM) of the Statistical Analysis System (1985). Duncan's new multiple range test was used to compare the treatment means and the correlation between the ME estimates and signi®cance test were also calcu-lated according to Steel and Torrie (1960).

3. Result and discussion

3.1. In vivo measurement with dairy goats

The chemical composition of the forage samples was analyzed and is presented in Table 1. Table 2

Table 1

The chemical composition of forage stuffs in Taiwana

DM (%) CP (%) EE (%) CF (%) Ash (%) NFE (%) ADF (%) GE (kcal/g)

Napier grass

Day 65 17.0 8.31 2.56 29.88 11.93 47.32 40.92 3.842

Day 60 20.3 7.10 2.83 30.27 10.10 49.70 42.92 3.728

Day 50 18.8 8.75 2.36 30.33 9.15 49.50 40.05 3.797

Day 40 17.1 9.30 2.71 24.81 9.90 54.59 37.88 3.713

Dwarf napier

Day 65 20.8 7.32 3.01 29.72 10.10 49.85 43.40 3.703

Day 40 16.7 10.03 2.83 23.59 9.30 54.25 37.2 3.781

Pangola

Hay Day 70 87.1 3.04 2.00 32.00 7.02 55.94 46.64 3.965

Fresh Day 45 39.7 6.69 2.07 30.60 9.24 51.40 42.80 3.767

Corn silage 23.8 8.34 2.71 30.56 5.40 52.99 37.37 3.978

Alfalfa hay 87.5 15.03 2.51 29.61 10.78 42.07 37.80 3.842

Bermuda hay 88.1 7.01 2.51 27.30 8.92 54.26 41.33 3.725

Timothy hay 88.6 9.21 2.80 28.31 9.30 50.38 40.92 3.713

a_{DM: Dry matter; CP: Crude protein; EE: Crude fat; CF: Crude ®ber; NFE: Nitrogen free extract; GE: Gross energy.}

M.-J

.

Lee

et

al.

/

Small

Ruminant

Resear

ch

36

(2000)

shows the apparent digestibility of various types of roughage in the dairy goat digestion trial. In Napier grass from different growth stages, the digestibility of DM and NFE from Days 40 and 50 were signi®cantly higher than those from Days 60 and 65 (P< 0.001). The CP digestibility from Days 60 and 65 were signi®cantly higher than those from Days 40 and 50 (P< 0.001). The CP digestibility of Napier grass cut on Day 40 was higher than that cut on Day 50 (P< 0.05). The digestibility of EE however, did not show a constant trend. Napier grass cut on Day 60 showed signi®cantly higher EE digestibility than grass cut on other days. Days 65 and 40 were next. The digestibility of EE was signi®cantly lowest for Day 50. The digestibility of CF, GE and hence the TDN value were signi®cantly highest on Day 40, followed by Days 50 and 60 in descending order. Digestibility was signi®cantly lowest for Day 65 (P< 0.05). The digest-ibility of NFE also showed a similar trend toward a digestibility decrease as the growth age increased.

Dwarf Napier grass on the other hand, showed signi®cantly higher digestion in most of the nutrients including EE, CF, NFE, GE and TDN for Day 40 over that of Day 65 (P< 0.05). The CP digestibility for Day 65 however was signi®cantly higher than that for Day 40 in dwarf Napier grass (P< 0.05). The nutrient digestibility for green chopped Pangola grass was also

signi®cantly higher for Day 45 than that for hay on Day 70 (P< 0.001). This result was similar to the ®ndings of Chen et al. (1973) and Lee et al. (1991). They indicated that the digestibility of DM, CF, NFE, GE and hence the TDN value in Napeir and Pangola grasses were higher at the earlier stages of growth and decreased as the plant approached maturity. In this trial, the digestibility of Napier and dwarf Napier grasses however showed a lower CP in the early stage than that in the late growing stages. This lower CP digestibility in the early growth stages did not agree with most of the results. Since fecal nitrogen loss and negative nitrogen balance can be exacerbated by the addition of highly digestible carbohydrates to a low nitrogen diet (Van Soest, 1982). The high NFE with low CP content in the early Napier grass growth stage may result in an increase nitrogen loss, hence a decrease in CP digestibility. Furthermore, the digest-ibility of most of the nutrients in Pangola grass hay from Day 70 was signi®cantly lower than that from Day 45, especially for crude protein. This extremely low CP digestibility from Day 45 may be because of the great amount of rainfall, which caused much of the nitrogen to run off. The grass also became over-mature in that summer. Hsu et al. (1990, 1993) also suggested that the most appropriate period to harvest Pangola and Napier grass was at 6±8 weeks of growth because

Table 2

The nutrients digestibility of forage by dairy goat trial*

DM CP EE CF NFE GE TDN (%) Gb(m1/200 mg)

Napier grass

Day 65 61.0b2.3 70.4a2.6 60.4b1.3 65.4c1.4 62.5b2.9 63.1d2.8 58.4d1.2 43.4d2.1

Day 60 63.2b6.2 70.6a1.8 65.3a3.4 67.3b1.8 65.2b3.7 66.4c3.3 61.9b1.5 47.0c1.8

Day 50 70.4a2.9 62.5c1.4 50.9c4.8 72.9a,b1.4 75.8a2.5 71.5b2.8 67.8b1.1 48.1b2.1

Day 40 71.2a2.4 65.2b1.3 61.2b2.2 75.2a2.7 74.2a3.2 75.1a1.5 69.4a1.7 52.1a1.5

Dwarf napier grass

Day 65 64.9b_{1.5 70.4}a_{2.8 64.9}b_{2.9 66.9}b_{2.8 68.2}b_{2.3 65.7}b_{2.6 63.4}b_{1.8 49.3}b_1.6

Day 40 72.3a_{2.2 66.3}b_{3.2 65.7}a_{3.2 76.1}a_{3.2 75.8}a_{2.2 74.5}a_{2.1 69.9}a_{2.0 54.9}a_2.3

Pangola grass

Hay, Day 70 54.1b_{1.7 34.0}b_{2.0 55.2}b_{2.5 54.0}b_{2.4 53.2}b_{1.0 47.1}b_{0.8 50.6}b_{1.0 40.3}b_1.9

Fresh, Day 45 79.8a_{4.1 61.9}a_{1.3 62.3}a_{5.2 81.1}a_{1.7 81.8}a_{1.3 78.2}a_{4.1 74.3}a_{1.1 58.3}a_2.5

Corn silage 65.23.5 52.30.9 64.43.3 68.81.0 63.23.3 61.21.7 63.61.8 53.41.9

Imported forage

Alfalfa hay 60.26. 1 66.32.7 55.33.6 43.61.0 78.51.5 54.02.0 59.01.1 44.31.6

Bermuda hay 61.55.3 65.21.9 64.13.0 57.11.7 60.11.6 60.02.5 56.40.5 47.12.7

Timothy hay 62.41.3 67.82.3 63.11.7 70.23.4 65.22.5 63.40.9 62.92.0 48.92.5

* a,b,c,d_{: Means of the same column within the same forage species with the different superscript were signi®cantly different (P< 0.05).}

the digestibility of DM and other nutrients was richest at this stage of growth. The DM digestibility and TDN of grasses over 8 weeks of growth determined by the in vitro method of Tilley and Terry (1963) declined as growth advanced. Pangola grass harvested on Day 70 was over mature, with ligni®ed cell walls lowering the digestibility. Van Soest (1982) suggested that a decline in digestibility occurred as grass is harvested later in its growing stage because of the ligni®cation in the cell walls.

The DM, CP, EE, CF, GE digestibility of corn silage was lower than the values for regular corn silage, especially for CP, hence the lower TDN value (63.6%) shown in this trial. This low digestibility might be attributed to both the harvesting stage and chopping. Corn used in this trial was harvested in the milk stage and chopped to 5±6 cm in length. This processing might cause dif®culties for goats to com-pletely masticate the corn ear. High water content in the forage corn may also result in a poor ensiled quality (75%) with lower digestibility.

The imported alfalfa hay contained only 15% CP which re¯ected over mature harvesting. The digestible nutrients and TDN were similar in both dairy goats and dairy cows when fed the same grades of alfalfa (National Research Council, 1988). The digestibility of the imported Bermuda hay and timothy hay were intermediate to that of Pangola and Napier grass harvested at Day-60 with a TDN of 56.4 and 62.9%, respectively. These values were also close to the (National Research Council, 1988) estimations for dairy cattle.

3.2. Correlation of ME estimated from the in vivo and in vitro trials

Table 3 presents different prediction equations. Table 4 presents the DE, TDN and ME of the forage plants used in Taiwan. The correlation coef®cients of the different ME estimate are presented Table 5.

In a comparison of the in vivo estimates of energy values, the relationship between ME values from in vivo trial, ME1, TDN and DE was highly correlated. The correlation coef®cient was 0.9981 for TDN and ME1, and 0.9997 for DE and ME1with the 12 different forages (nˆ12). When ME estimates were compared with in vivo estimates (ME1) and in vitro gas produc-tion (ME2and ME3), the correlation between the ME1 estimated from in vitro gas production and the ME2 from the gas production, CP and EE was high. But the correlation was still lower (rˆ0.8790) than the ME1 and ME3(rˆ0.8872) that included variable NFE into the predict equation in addition to the variables of ME2.

In a comparison of the ME estimates from the in vivo (ME1) with the NRC (ME4) and the in vitro gas production (ME2and ME3), the correlation of ME1 and ME4 (rˆ0.9444, P< 0.0001) was higher than both the correlation of ME2 and ME4 (rˆ_0.8790, P< 0.0001). It was also higher than the correlation of ME3and ME4 (rˆ0.8872, P< 0.0001). Since both ME1and ME4were calculated from same set of data derived from in vivo dairy goat digestion trial. From higher correlation coef®cient of ME3and ME1to the correlation of ME2and ME1, and ME3 and ME4to

Table 3

The predict equations of metabolizable energya

Item Equation for ME estimation Source

ME1 15.2DCP‡34.2DEE‡12.8DCF‡15.9DNFE In vitro digestion trial

ME2 0.145Gb‡0.00412CP‡0.00650CP2‡0.0206fat‡1.54 Menke et al., 1979

ME3 0.118Gb‡8.72CP‡19.21fat‡3.38NFE‡0.691 Menke et al., 1979

ME4 0.82DE National Research Council, 1988

ME5 3.16‡0.0695Gb‡0.00007300 G2b‡0.00732CP‡0.02052fat Steingass, 1980

ME6 1.56‡0.1390Gb‡0.007400CP‡0.01780fat Close and Menke, 1986

ME7 ÿ0.58‡0.1590Gb‡0.0102CP‡0.03140fat Rohr et al., 1986

ME8 1.20‡0.1456Gb‡0.00076575CP‡0.01642fat SchoÈner, 1981

a_{DE: Digestibility of gross energy; TDN: Total Digestible of nutrient, where DCP, DEE, DCF, DNFEˆg digestible crude protein, fat,}

ME2and ME4,this indicated that when all of the 12 forages were evaluated, ME3was highly correlated to in vivo estimates as compared to that for ME2. It appears that the ME estimate on all forages measured

will be more precise with NFE inclusion in their equation.

Comparison of the ME4of the NRC estimate to the other in vivo result without inclusion of corn silage

Table 4

The digestible energy, TDN and ME of forage stuffs in Taiwan

Item DE (MJ/kg) ME1(MJ/kg) ME2(MJ/kg) ME3(MJ/kg) ME4(MJ/kg)

Napier grass

Day 65 10.15 8.62 8.38 8.64 8.32

Day 60 10.35 9.15 8.77 9.07 8.49

Day 50 11.35 10.04 9.08 9.25 9.31

Day 40 11.67 10.31 9.75 10.01 9.57

Dwarf Napier grass

Day 65 10.18 9.40 9.12 9.41 8.35

Day 40 11.79 10.48 10.25 10.42 9.66

Pangola grass

Hay Day 70 7.82 7.36 7.48 7.82 6.41

Fresh Day 45 12.33 11.00 10.35 10.31 10.11

Corn silage 10.18 9.31 9.83 10.04 8.35

Alfalfa hay 8.67 8.89 9.54 9.13 7.11

Bermuda hay 9.35 8.43 8.77 9.18 7.67

Timothy hay 9.85 9.32 9.28 9.50 8.08

Table 5

The correlation coef®cients of metabolizable energy estimatesa

Item DE TDN ME1 ME2 ME3 ME4 ME5 ME6 ME7 ME8

DE 0.9980 0.9980 0.9024 0.8874 0.7825 0.8198 0.8177 0.8174 0.8837

0.0001 0.0001 0.0001 0.0002 0.0044 0.0001 0.0001 0.0001 0.0001

TDN 0.9981 0.9998 0.9416 0.9126 0.7600 0.8231 0.8252 0.8290 0.9287

0.0001 0.0001 0.0001 0.0007 0.0066 0.0001 0.0001 0.0001 0.0001

ME1 0.9997 0.9981 0.9521 0.9272 0.9673 0.8302 0.8352 0.8368 0.9283

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

ME2 0.7747 0.8608 0.8790 0.8297 0.9024 0.9806 0.9845 0.9856 0.8242

0.0051 0.001 0.0001 0.0016 0.0001 0.0001 0.0001 0.0001 0.0001

ME3 0.7602 0.8753 0.8872 0.9652 0.8774 0.9554 0.9627 0.9644 0.7953

0.0001 0.0002 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

ME4 0.8745 0.9410 0.9444 0.8790 0.8872 0.9995 1.000 0.9999 0.7877

0.002 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

ME5 0.6562 0.7697 0.7720 0.8841 0.9481 0.9996 0.9999 0.9993 0.7849

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

ME6 0.6591 0.7805 0.7832 0.8773 0.9452 1.0000 0.9995 0.9999 0.7877

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

ME7 0.6585 0.7817 0.7843 0.8820 0.9485 0.9999 0.9993 0.9999 0.7890

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

ME8 0.8519 0.8411 0.8385 0.5059 0.5820 0.6083 0.7849 0.6083 0.6083

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

a_{The correlation coef®cient values in the lower and left side of diagonal of the table were calculated on the ME value which derived from}

forages without inclusion of corn silage and alfalfa hay (nˆ10). The right and upper side of the table content values derived from ME values of the 12 feeds (nˆ12).

and alfalfa (nˆ10), the correlation between ME1and ME4 was the highest (rˆ0.9673, P< 0.0001). The correlation coef®cient for ME1and ME2(rˆ0.9521, P< 0.0001) was higher than the correlation for ME1 and ME3(rˆ0.9272,P< 0.0001) from the remaining 10 forages. The correlation of ME2 and ME4 (rˆ0.9024, P< 0.0001) was also higher than the correlation of ME3 and ME4 (rˆ0.8774, P< 0.0001). This indicated that ME2and ME3were appropriate for estimating the ME of forage. ME2 was more practical for estimating the ME of forage than ME3. This represents that inclusion of NFE in addition to EE, CP and gas production in the predicted equation did not improve the precision of the ME estimates. For forage, both CF and NFE procedures are known to be inadequate for accurate determinations. The CF method tends to underesti-mate true ®ber content, especially in immature forage with high hemicellulose content. Data presented in Table 2 clearly demonstrates this concept as CF values which are suppose to represent total ®ber are consis-tently lower than ADF which represents only the cellulose content of the forage. Therefore, CF should be replaced by ADF in the predicting equation to improve the precision. The ME estimated on tropical forage will be more precise without NFE inclusion in the equation.

The NFE procedure also tends to overestimate values for forage. The values presented in Table 2 are considerably high. The actual values for starch would be much less. Take Bermuda grass for example, Sniffer et al. (1992) indicated 79.5% carbohydrate (CHO) with 66.6% ®brous and 12.9% non-®brous CHO with only 0.8% starch and 53.1% slow degrad-able CHO. From our laboratory showed that Napier grass with 56.0% ®brous CHO and 29.4% non-®brous CHO with only 1.7% starch and 46.5% slow degrad-able CHO (unpublished data). Since high correlation achieved in this trial, it appears that underestimation of total ®ber was compensated for by an overestimation of soluble carbohydrates.

The estimates for forage ME, which were derived from in vitro methods using gas production, CP and EE to calculate ME value (ME2) was initiated by Menke et al. (1979) and were modi®ed by Menke and Steingass (1988). This revised equation is ME6ˆ1.56‡0.1390Gb‡0.007400CP‡ 0.01780-EE. The correlation of ME6and ME1estimated from

the goat digestion trial in this study (rˆ0.7832) was lower where the correlation of ME6 and ME2 (rˆ0.8773) was lower than that of the correlation of ME6and ME3(rˆ0.9452). When comparing the ME estimates of gas production, CP and EE from this study to other data, the correlation coef®cients were also very high. The ME1and ME5(rˆ0.7720), which were estimated from the goat digestion trial (Steingass and Menke, 1980) and ME7(Rohr et al., 1986) and ME8 (SchoÈner, 1981) which was calculated using regression were high. These values however were still lower than correlation coef®cient determined from ME1and ME2.

The amount of gas produced from this trial showed a trend toward decreasing as the period of forage growth advanced. The average gas production was around 40±58 ml/200 mg (Table 2) and agreed with the value of Zinash et al. (1996). They also found a decrease in gas production as the forage growing period was prolonged.

It appears that the ME of forage estimated from this trial is easy to measure and provides a reasonable estimate and can be used to replace the digestion trial. This however, can not provide precise estimates for lactating dairy goat because this study utilized dried dairy goats at a maintenance level to generate the in vivo data. Lactation stage has a great impact on nutrient intake and subsequently, nutrient digestibility and utilization.

4. Conclusion

References

Abreu, J.M.F., Bruno-Soares, A.M., 1998. Chemical composition, organic matter digestibility and gas production of nine legume grains. Anim. Feed Sci. Technol. 70, 49±57.

Association of Of®cial Analytical Chemists, 1980. Of®cial Methods of Analysis, 13th ed. AOAC, Washington, DC. Chen, M.C., Chen, C.P., Tan, S.Y., Chung, P., 1973. The nutritive

value of napier grass and pangora grass in Taiwan. J. Chin. Soc. Agric. Sci. 83, 46±53 (in Chinese, with English abstract). Close, W., Menke, K.H., 1986. Selected topics in animal nutrition.

Animal Nutrition, 2nd ed. University Hubenelm, Germany, pp. 26±66.

Getachew, G., BluÈmmel, M., Makkar, H.P.S., Becker, K., 1998. In vitro gas measuring techniques for assessment of nutritional quality of feeds: a review. Anim. Feed Sci. Technol. 72, 261± 281.

Harries, L.E., 1970. A digestion and balance trial for sheep, deer, antelope and relagted species. In: Nutrition Research Techni-ques for Domestic and wild Animals, Vol. 1. Utah State University, Logan, Utah, pp. 5301±5303.

Hsu, F.H., Hong, K.Y., Lee, M.J., Lee, K.C., 1990. Effect of cutting intervals on forage yield, nutrient composition and silage quality of Napiergrass. J. Chin. Soc. Agric. 151, 77±89 (in Chinese, with English abstract).

Hsu, F.H., Hong, K.Y., Lee, M.J., 1993. Effect of cutting intervals on forage yield and quality of pangolagrass and coastal cross. II. Bermudagrass. Taiwan Livest. Res. 26 (1), 91±97 (in Chinese, with English abstract).

Lee, C.F., Buu, R.H., Shy, Y.M., Chen, M.C., 1991. The nutritive value of pangola grass A253 at different stages of growth. Taiwan Livest. Res. 24 (1), 59±65 (in Chinese, with English abstract).

Menke, K.H., Raab, L., Salewski, A., Steingass, H., Fritz, D., Schneider, W., 1979. The estimation of the digestibility and metabolizable energy content of ruminant feedingstuffs from the gas production when they are incubated with rumen liquor in vitro. J. Agric. Sci. Camb. 93, 217±222.

Menke, K.H., Steingass, H., 1988. Estimation of the energetic feed value obtained from chemical analysis and in vitro gas produc-tion using rumen ¯uid. Anim. Feed Sci. Technol. 28, 91±97. National Research Council, 1988. Nutrient Requirements of Dairy

Cattle. National Academy Press, Washington, DC.

Rohr, K., Fliegel, H., Potthast, V., 1986. UÈ berpruÈfung verschied-ener Methoden zur SchaÈtzung des Energiegehaltes von Mischfuttermitteln fuer WiederkaÈuer. Z. Tierphysiol. Tierer-naÈhr. Futtermittelk. 56, 129±130 (in German).

SchoÈner, F.J., 1981. SchaÈtzung des energetischen Futterwertes von Milchleistungsfuttern (ErgaÈnzungsfutter fuer MilchkuÈhe) unter besonderer BeruÈcksichtigung des Hohenheimer Futterwert-testes. Bonn, Univ., Diss. (in German).

Shiemann, R., Nehring, K., Hoffmann, L., Jentsch, W., Chudy, A., 1971. Energetische Futterbewertung und Energienormen. VEB Deutscher Land wirtschaftsverlag, Berlin, p. 75 (in German). Sniffer, C.J., O'Connor, J.D., Van Soest, P.J., Fox, D.G., Russell,

J.B., 1992. A net carbohydrate and protein system for evaluating cattle diets: II. Carbohydrate and protein availability. J. Anim. Sci. 70, 3562±3577.

Statistical Analysis System, 1985. SAS User's Guide, version 5 ed. SAS Institute Inc., Carry, NC.

Steel, R.G.D., Torrie, J.H., 1960. Principles and Procedures of Statistics. McGraw-Hill, New York.

Steingass, H., Menke, K.H., 1980. Die Bestimmung der Verdau-lichkeit und der Gehalte an umsetzbarer Energie und Nettoenergie-Laktation im Hohenheimer Futterwerttest. Kraft-futter 11, 534±536 (in German).

Tilley, J.M.A., Terry, R.A., 1963. A two stages technique for the in vitro digestion of forage crops. J. Br. Grassld. Soc. 18, 104±111. Van Soest, P.J., 1982. Chemistry of forages and feeds. In:

Nutritional Ecology of the Ruminant. Cornell University Press, New York, pp. 75±151.

Zinash, S., Emyr, O., Mewa, S., Michael, K.T., 1996. Prediction of in situ rumen dry matter disappearance of Ethiopian forages from an in vitro gas production technique using a pressure transducer, chemical analyses or in vitro digestibility. Anim. Feed Sci. Technol. 61, 73±87.