Livestock Research for Rural Development 9 (3) 1997

Citation of this paper

A study of washing losses and in vitro gas production characteristics of nine leaves from tropical trees and shrubs for ruminants

J Ly*, Nguyen Van Lai and T R Preston

University of Tropical Agriculture, Thu Duc, Thanh Pho Ho Chi Minh, Vietnam
*Present address: Swine Research Institute, Carretera del Guatao km 1, Punta Brava. La Habana, Cuba



Nine sun-dried leaves from tropical trees and shrubs were examined for washing losses of DM (WL) and in vitro gas production characteristics of the rumen ecosystem. In vitro gas production was compared with WL. The gas volume production could be explained by the amount of total short chain fatty acid (SCFA) produced over 96 hr (R2 = 0.88; P<0.001). The gas:SCFA production ratio was 23.8 ml/mmol . Potential gas production a + b as described by the exponential model p = a + b (1 + e-ct) was significantly (P<0.05) correlated with WL of either unwashed (r = 0.74) or washed leaves (r = 0.67) after 96 hr of incubation. Actual gas production at 3, 6, 12, 24, 48 and 96 hr could be predicted from WL of leaves (0.49<R2<0.77; P<0.05). The highest correlation between both parameters was obtained at 6 hr (R2 = 0.77; P<0.001). This same interdependence could not be established with washed samples. It is concluded that the washing loss test has good potential as predictor of in vitro gas production in the rumen ecosystem.

Key words: In vitro gas production, tropical trees and shrubs, leaves, washing losses, dry matter, sun-dried basis


The in vitro gas production technique developed by Menke et al (1979) has been used by Blummel and Orskov (1993) to determine gas production at several incubation times, and these values could describe the pattern of fermentation of feed, by using the equation of McDonald (1981). Furthermore, this technique has been proposed as a valuable tool to predict ruminant performance traits of economical implication (Blummel and Orskov 1993).

On the other hand, it has been suggested that washing losses of dry matter (WL) could be a very simple parameter to assess the overall process of leaf degradation in the rumen ecosystem (Ly and Preston 1997). In this paper the possibility of the interdependence, if any, among the WL dry matter and in vitro rumen degradation characteristics from leaves of tropical trees and shrubs was explored.

Materials and methods

Samples of nine different leaves from tropical trees and shrubs were studied. These same types of leaves had been used in a preliminary trial reported elsewhere (Ly and Preston 1997). The origin of the samples and the procedure of collection (March 1997, during the dry season) and preparation of the leaves was the same.

Washing losses of dry matter

The zero-time washing loss (WL) of dry matter was considered to be A, the readily degradable fraction of leaves in the rumen, as defined by Orskov and Riley (1990). The WL value was estimated after 90 min of washing sun-dried samples in a semi-automatic washing machine as outlined by Ly and Preston (1997). The WL value was determined in triplicate.

In vitro gas production

The samples of either unwashed or washed leaves were incubated with rumen fluid in calibrated syringes (100 ml) following the procedure of Menke et al (1979) and Menke and Steingass (1988). The syringes were incubated in a water bath at 39oC with a polyurethane lid fitted with holes to hold the syringes upright in the water bath.

Rumen fluid was collected at 9:00 am, about one hour post-feeding, from one of three rumen fistulated Bos indicus heifers fed a roughage diet of rice straw. The rumen fluid was filtered, mixed with a sodium and ammonium carbonate buffer (35 g NaHCO3 plus 4 g NH4HCO3 per liter), in a ratio 1:2 (v/v). About 200 mg of air dry milled samples plus 30 ml of rumen fluid and buffer were incubated in triplicate. The syringes were gently shaken one hour after the start of incubation and thereafter every three hours during the first 48 hr of incubation. A blank of rumen fluid plus buffer was included for comparison. Gas production was recorded at 3, 6, 12, 24. 36, 48, 60, 72, 84 and 96 hr. Net gas volume at each incubation period was calculated by subtracting the mean gas volume of the blanks from the mean volume of gas in syringes with samples. The volumes of gas were not corrected according to a standard.

Chemical analysis

Dry matter was estimated in the leaf samples by microwave radiation until constant weight according to the recommendations of Undersander et al (1993). Total short chain fatty acids (SCFA) and pH values were determined in the supernatant of samples incubated in the gas syringes. Total SCFA were estimated by distillation from a saturated solution of MgSO4 in H2SO4 4 N (J Ly 1996, unpublished data). Briefly, 5 ml of the supernatant from syringes were mixed with 300 ml of the acid, saturated MgSO4 solution and distilled until 200 ml of distillate was collected and then titrated with KOH 0.05 N. Five ml of a standard solution of acetic acid (analytical grade, 3.00 g per liter) was distilled in the same manner by quadruplicate. The acetic acid recuperation was 99.50.2 %. The pH values were determined by a glass electrode in an aliquot of the same supernatant.

Statistical analysis

Analysis of variance was used to determine differences in the WL of the samples of leaves and to compare the effect of washing leaves on total gas and SCFA production over 96 hr. Differences between treatment means were tested using the t-test and the Duncan multiple range test in the appropriate cases (Steel and Torrie 1980).The results from gas volume recording were fitted to the exponential equation of the form p = a + b (1 - e-ct) according to McDonald (1981), where p represents gas production at time t, a + b the potential gas production, and c the rate constant. The estimations were carried out as outlined by the Chen program (Chen 1995). When it was convenient, the data were submitted to regression analysis using the Statistical Analysis System procedure (SAS 1982).

Results and Discussion

There were significant differences (P<0.001) in the WL value among the nine types of leaves used in this study (Table 1), as had been previously observed (Ly and Preston 1997). The highest WL values were for gliricidia (Gliricidia sepium), hibiscus (Hibiscus rosasinensis) and leucaena (Leucaena leucocephala), and the lowest for nipa palm (Nipa fruticans). Intermediary values of WL were found for the other five types of leaves. Interestingly, a trend was found for an inverse relationship between the DM content of leaves and their WL value (r=-0.679; P<0.04). Abdulrazak et al (1996) also observed that the WL of G. sepium and L. leucocephala, as determined by hand-washing, were relatively high in the range of 30.7 - 32.4 and 30.5 - 31.3%, respectively.

The initial and final total SCFA concentrations and pH values of the rumen plus buffer incubated with no added substrate were 1250 and 137) 14 mmol/litre respectively, and 6.920.15 and 7.060.05 (means of nine observations), thus indicating a small increase in the value of both indices. In fact, a small but evident in vitro gas production was noted in these syringes (11.50.8 ml). The increase in total SCFA concentration and gas production was probably due to microbial activity upon the remains of the soluble fraction of the roughages fed to the donor animals.

The effect of washing on overall gas and total SCFA production in leaves incubated during 96 hr is shown in Table 2. There was no treatment effect on the final pH of the solution and, again, a clear shift of pH was observed towards more alkaline values. Khazaal et al (1995) encountered lower pH values (from 6.69 to 6.84) in the buffered rumen fluid after 96 hr of incubation than those found in the present study, and suggested that in a closed system such as that occurring in the syringes, the production of ammonium bicarbonate could increase the buffering capacity of the liquid phase during the fermentation processes. Our experiment could not provide more evidence to support this hypothesis. In fact, the increase of pH could be due either to CO2 release from the incubation medium or by an increase in ammonia concentration due to deamination activity of the rumen micro flora. On the other hand, Huntington and Givens (1997) have observed that increasing the concentration of innoculum can reduce the final pH and increase the asymptote of gas production, whereas Fakhri et al (1997) found that final pH and total SCFA were markedly influenced by the method used for in vitro gas production from starch-rich feeds.

On average, in vitro gas and total SCFA production after 96 hr accounted for 77.9% and 80.2% of the same indices when values determined from washed leaves were compared with those obtained from the same unwashed samples. This overall effect was not significant (P>0.05). However, when individual types of leaves were examined, the depressive effect of washing was significant for gas production in eight of the leaves studied. This same influence was observed for total SCFA production in four types of leaves. This was probably because the variance found within each type of leaf was less than that among the various leaves analysed.

There were no significant differences (P>0.10) in in vitro total gas:SCFA production ratio after 96 hr when data from unwashed and washed samples were compared. The resulting pooled equation predicting in vitro gas production (Y, ml) from SCFA production (X, mmol) was highly significant (R2 = 0.88; P<0.001):

Y = 0.062 + 23.8 X (Syx = 4.59)

It has been argued that the volume of gas produced could be explained either by the amount and proportion of SCFA produced during the incubation or by true organic matter (100 - NDF) disappearance (Blummel and Orskov 1993). In this connection the results obtained here provide more evidence of the existing interdependence between total gas and SCFA production.

Table 3 summarizes the effect of washing the leaves on the cumulative in vitro gas production and sample degradation characteristics defined by the equation p = a + b (1 - e-ct). Potential gas production (a + b) of the nine unwashed leaves was lower (25.7 ml/200 mg DM) than that found by Siaw et al (1993) in leaves from six genera (from 36.5 to 65.9 ml/200 mg DM). In contrast the potential gas production in five browse species appeared to be less than 40 ml/200 mg DM according to Kibon and Orskov (1993). Reasons for explaining this variability are not apparent, but the diverse origin of the leaves tested could partially account for these results.

It is evident that A. auriculiformis and N. fructicans had a very low potential gas production: 12.2 or 5.7, and 8.3 or 4.4 ml/200 mg DM respectively. Furthermore, the duration of the lag phase was from 27.7 to 29.9 hr in A. auriculiformis, and this was a reflection of the absence of gas in the syringes containing this acacia during the first 36 hr of incubation. H. rosasinensis was at the other extreme of the range encountered. Leaves from this malvacea had the highest potential of gas production in this experiment (63.3 ml/200 mg DM).

A trend for washed samples to have a lower potential gas production was observed in this experiment. The exceptions were A. heterophyllus and L. leucocephala. It was observed also that the leaves which had been washed, and which had a low potential gas production, had an increase in the value of the rate constant "c" (A. auriculiformis, A. mangium, A. occidentalis, N. fruticans and T. gigantea). A change in the pattern of potential and rate of gas production could be expected, taking into account that samples from washed leaves had no soluble, readily fermentable substances. However, it must be borne in mind that many anti-nutritional factors present in the leaves of trees and shrubs are water soluble (Kumar 1992). Therefore this factor could be also influencing the pattern of in vitro gas production.

The WL values of the sun-dried leaves were also significantly correlated (P<0.05) with in vitro potential gas production of either the unwashed or the already washed samples. In Table 4 are shown the details of these relationships.

Actual gas production from unwashed leaves at 3, 6, 12, 24, 48 and 96 hr could be predicted from the WL of the leaves (Table 5). The highest correlation between these parameters was obtained at 6 hr (R2 = 0.77; P<0.001) as can be seen in Figure 1. This could be a reflection of the nature of the preference of the rumen microflora for the substances contained in the soluble fraction. This same interdependence could not be established with washed samples at any time (R2 < 0.10; P>0.10).

The work described in this paper supports the work of others (Kibon and Orskov 1993; Siaw et al 1993) indicating the potential of the in vitro gas production technique of Menke et al (1979) for ranking leaves from tropical trees and shrubs. Moreover, it has been shown that the WL test has a good potential as a predictor of in vitro gas production in the rumen ecosystem. In agreement with these findings, Chermiti et al (1996) have found that the neutral soluble fraction (NDS = 100 - NDF) is highly identified with the WL value, and an interdependence has been claimed to exist between the degradation of the so-called true organic matter and the in vitro gas production (Blummel and Orskov 1993).

Nevertheless, more work is required to identify any interaction between method and type of sample, such as the influence of the variations in the pH media on the characteristics of in vitro gas production.


This research will be submitted by Nguyen Van Lai to the University of Tropical Agriculture Foundation in partial fullfillment of the requirements for the degree of Master of Science in Sustainable Use of Natural Renewable Resources, a programme financed by the Danish Embassy in Vietnam, with support from the British Council, Hanoi, Vietnam. Dr J Ly is currently on sabbatical leave from the Instituto de Investigaciones Porcinas, Havana, Cuba with financial support from the Swedish Agency for Research Cooperation with Developing Countries (SAREC).


Abdulrazak S A, Muinga R W, Thorpe W and Orskov E R 1996 The effects of supplementation with Gliricidia sepium or Leucaena leucocephala forage on intake, digestion and live weight gains of Bos taurus x Bos indicus steers offered napier grass. Animal Science 63:381-388

Blummel M and Orskov E R 1993 Comparison of in vitro gas production and nylon bag degradability of roughage in predicting feed intake in cattle. Animal Feed Science and Technology 40:109-119

Chen X B 1995 XBC Laboratory. An utility for processing data of feed degradability and in vitro gas production. Version 4.0. Rowett Research Institute. Aberdeen

Chermiti A, Nefzaoui A, Teller E, Vanbelle M, Ferchichi H and Rokbani N 1996 Prediction of the voluntary intake of low quality roughages by sheep from chemical composition and ruminal degradation characteristics. Animal Science 62:57-62

Fakhri S, Moss A R, Givens D I and Owen E 1997 Comparison of four in vitro gas production methods to study rumen fermentation kinetics of starch rich feeds. Proceedings of the British Society of Animal Science p 196

Huntington J A and Givens D I 1997 The effects of inoculum concentration and method of inoculation on the gas production profile of high temperature dried grass. Proceedings of the British Society of Animal Science p 192

Khazaal K, Dentinho M T, Ribeiro J M and Orskov E R 1995 Prediction of apparent digestibility and voluntary intake of hays fed to sheep: comparison between using fibre components, in vitro digestibility or characteristics of gas production or nylon bag degradation. Animal Science 61:527-538

Kibon A and Orskov E R 1993 The use of degradation characteristics of browse plants to predict intake and digestibility by goats. Animal Production 57:247-251

Kumar R 1992 Anti-nutritional factors. The potential risks of toxicity and methods to alleviate them. In: Legume trees and other fodder trees as protein sources for livestock (A Speedy and J L Pugliese ed.) FAO Animal Production and Health Paper 102. Rome p 145-160

Ly J and Preston T R 1997 An approach to the estimation of washing losses in leaves of tropical trees. Livestock Research for Rural Development 9(3):5-9

McDonald I 1981 A revised model for the estimation of protein degradability in the rumen. Journal of Agricultural Science, Cambridge 96:251-252

Menke K H, Raab L, Salewski A, Steingass H, Fritz D and 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. Journal of Agricultural Science, Cambridge 93:217-222

Menke K H and Steingass H 1988 Estimation of the energetic feed value obtained from chemical analysis and in vitro gas production using rumen fluid. Animal Science and Development 28:7-55

Orskov E R and Ryle M 1990 Energy nutrition in ruminants. Elsevier Applied Science. London pp 149

Siaw D E K A, Osuji P O and Nsahlai I V 1993 Evaluation of multipurpose tree germplasm: the use of gas production and rumen degradation characteristics. Journal of Agricultural Science, Cambridge 120:319-330

SAS 1982 SAS user's guide. Statistical Analysis Systems Institute. Cary NC

Steel R G D and Torrie J H 1980 Principles and procedures of statistics. McGraw-Hill Book Co. New York

Undersander D, Mertens D R and Thiex N 1993 Forage analysis procedures. National Forage Testing Association. Omaha pp 154

Received 11 June 1997

Return to top