With the cumulative gas production technique (Theodorou et al., 1994), the amount of gas released as an end-product of substrate fermentation is measured at regular time intervals, as a measure of the kinetics of fermentation. CO2 is released both directly as a result of the fermentation and indirectly from the bicarbonate buffer in the medium (Beuvink and Spoelstra, 1992). This method was first described for the evaluation of fibrous materials. However, little work has been reported describing the use of cumulative gas production to evaluate concentrates. Soluble fermentable components are rapidly fermented after incubation, while the insoluble ones need to be hydrated and colonized by micro-organisms before they can be degraded (Van Milgen et al., 1993). The varying composition of substrates is one of the reasons for the different rates of gas production seen as these processes are taking place. Concentrates can differ greatly in their composition and therefore in their fermentation characteristics. Here, we describe our first trial with four individual concentrate ingredients, and mixtures of them, to determine whether the parameters from the individual ingredients can be used to predict the fermentation characteristics of the mixed foods.

A major focus for improving the diets in many less developed countries (LDCS) is the provision of rumen fermentable nitrogen (N) using protein supplements to complement N-deficient foods. However, in vitro digestibility methods usually use N-rich environments for the degradation of single foods. This conventional approach may give data which do not reflect the nutritive value of the N-deficient diets often on offer in LDCS, neither is it appropriate for using in vitro gas production to study protein supplementation. Our earlier study indicated that, by using a N-free medium, the gas production technique responded to added ammonium sulphate and urea. The ADAS standardized methodology, which used 10 ml of inoculum instead of the 5 ml used in the earlier study, was found not to be very responsive to N supplementation. The ADAS methodology was therefore investigated in order to develop a modified protocol for fermenting foods in an N-limited environment. The study involved using inocula diluted to different extents in N-free medium for fermenting N-deficient substrates in N-free and N-rich media. The modified protocol was then used for investigating the interactions between N-rich and N-deficient foods from north-west India.

In vitro gas production techniques are becoming increasingly popular as tools to investigate the food quality for ruminant and monogastric herbivores. They are attractive due to their ability to measure the kinetics of digestion associated with the gut microbial fermentation of foodstuffs. It is therefore important to know and understand what factors affect the volume of gas produced, particularly as they relate to repeatability within or between laboratories. In studies involving gas pressure and volume measurement, the chemical composition of the culture medium, the nature of the microbial population and the type and amount of substrate available for fermentation all have a significant influence on gas accumulation. In this study, the manual pressure transducer technique (PTT) of Theodorou et al. (1994) was used to investigate the effect of head-space pressure on the resultant gas production profiles.

Following the development of the Menke technique in 1979, the measurement of gas production in vitro has become increasingly popular for investigating the kinetics of rumen fermentation. The aim of this study was to compare the gas production profiles for three foods using two in vitro gas production techniques; the Menke et al. (1979) technique (MT) and the pressure transducer technique (PTT) (Theodorou et al., 1994). Both techniques involve recording gas production throughout the incubation of a food sample with rumen fluid. The MT incubations are made in gas-tight syringes where the volume of gas produced causes the plunger to move up the syringe barrel. The PTT involves measuring gas production in fermentation bottles using a pressure transducer and syringe assembly to measure the pressure and corresponding gas volume. As the medium to rumen fluid ratios also differ between techniques; 2:1 in the Menke technique and 9:1 in the PTT, both ratios were investigated in this study.

Methodology can play a critical rôle in the measurement of digestion kinetics, especially when the objective is to define kinetic parameters for use in formulating rations or modelling animal responses. Measurement of gas production kinetics provides the opportunity to evaluate the rate of digestion of the soluble, more rapidly fermenting fractions of foods but has the potential for being more sensitive to the in vitro procedure used. Differences among procedures that have little impact on digestion of dry matter after 48 h of incubation, may have dramatic effects on fermentation of soluble matter during the first 20 h. Our objective was to develop a method for measuring the kinetics of gas production that would minimize any detrimental effects associated with the in vitro system and provide estimates of digestion kinetics that can be used to both describe foods for ration formulation systems and provide parameters for models of ruminal digestion.

Fat content is one of the main factors affecting the degradability and fermentability characteristics of oil-rich foods estimated by means of the in situ, in vitro and gas production techniques. Filtered fat, considered degraded/fermented, is of limited value to the rumen microbes and may inhibit fermentation. Pre-extraction of oil may be one way to solve this problem. The rumen organic matter degradability (OMD) and fermentability (FOM) of concentrate foods were evaluated in this study with and without oil extraction.

Menke et al. (1979), Beuvink et al. (1992) and Theodorou et al. (1994) developed techniques for measuring the time course of gas production of foods fermented in vitro with rumen fluid. These techniques require description of the fermentation profile with an appropriate mathematical model. Although several authors have used these techniques to study the ruminal fermentation of foods, little information is available on the suitability of the model chosen for describing the fermentation profile of the food under study. In this study, the models of Ørskov and McDonald (1979), France et al. (1993) and Beuvink and Kogut (1993) were fitted to the in vitro gas production profiles of 10 whole-crop wheat (WCW) forages (cv. Slepjner) to determine the model most suited to describing the data.

The pressure transducer technique of Theodorou et al. (1984) is becoming of increasing importance in food evaluation. The main advantage over end-point procedures is the collection of kinetic data on a food. Previously such data were only obtained by sequential sacrifice or in situ techniques. Earlier work (Harris, 1996) showed that kinetics in the early stages of the incubation may not accurately simulate the processes occurring in vivo and this led to the use of a priming technique when the microbial innocula is acclimatized for 24 h to a priming food similar to the basal diet of the donor animal. This work investigates the effect of the length of priming on fermentation characteristics of two foods.

Samples of barley grain and straw were ground through a 1-mm screen and a priming food of grass silage and concentrates prepared according to the method of Harris (1996). Gas production was determined from the barley grain and straw using bovine rumen liquor after exposing the microbial population to the priming food for 0 (unprimed), 6, 12, or 24 h. Gas volumes were recorded manually and the blank corrected volumes fitted to the equation of France et al. (1993).

Pell and Schofield (1993) described a gas production technique where cumulative pressure is related to gas production and hence organic matter (OM) fermentation. This technique has been used to describe rates of carbohydrate degradation for use in the Cornell net carbohydrate and protein system (CNCPS, Barry et al., 1994). Increasingly the CNCPS model is being utilized in the United Kingdom (UK) to ration dairy cows and, as a result, a UK foodstuff database has been developed containing the chemical description of the protein and carbohydrate pools. It was necessary to establish whether the gas production technique could be reproduced in a UK laboratory in order to provide rates of fermentation of the carbohydrate pools. Gas production techniques commonly used in the UK do not allow the vessel gas pressure to accumulate so a comparison of methodology was required.

Whilst it is recognized that most foods contain a number of different carbohydrate pools most applications of the gas production technique do not allow them to be differentiated despite their likely different nutritional attributes. Schofield and Pell (1995) reported that improved fermentation kinetics of the carbohydrate pools could be obtained by making gas production measurements on the whole food and on its neutral-detergent fibre (NDF) fraction, the data on the neutral-detergent solubles (NDS) being calculated by difference. More recently Doane et al. (1996) proposed an expansion of this to include incubation of the ethanol insoluble residue, to enable the NDS fraction of the food to be separated into a soluble carbohydrate and organic acids fraction and a starch and pectin fraction. Another approach is that of Groot et al. (1996) who used a multiphasic model of whole food gas production to show that the gas is produced from more than one food fraction. The objective of the present work was to combine these two approaches to examine the value of the multiphasic model applied to gas production from the isolated NDF and NDS fractions.

The gas production (GP) technique (e.g. Theodorou et al., 1994; Cone, 1994) is increasingly used as an indirect measure of substrate fermentability in the rumen. However, there are few data concerning the effect of making changes to the various integral steps of the GP procedure. This experiment was designed to study the effect of the host diet on the gas production profiles of grass hay and high temperature dried grass (HTDG).

Two cows were given either a silage:barley diet (80:20 dry-matter (DM) basis) or barley straw (ad libitum). Four experimental periods with an adaptation period of 3 weeks were used so that each animal experienced each diet type twice. Rumen fluid and solids were collected at the end of each 3-week period, prepared and used to inoculate the cultures. Duplicate samples (milled, 1-mm screen) of either hay or HTDG were accurately weighed (̴ 0-5 g DM) into 250 ml bottles, 85 ml anaerobic buffer (as Theodorou et al., 1994) were then dispensed into each bottle before being connected to the automated GP data logger (as Cone, 1994). After 16 h bottles were inoculated with rumen fluid (15 ml) according to the experimental design and incubated at 39°C for 48 h. The GP data were fitted to the model suggested by France et al. (1993). A representative sample (8 ml) of final culture fluid was preserved with 1 mol/1 orthophosphoric acid (2 ml) and stored frozen (-20°C) to await analysis for volatile fatty acids.

The polyphenol concentration in various tree fodders has been assayed by the gas production technique using rumen fluid inoculum either by fermentation of whole samples +/- a polyphenol adsorbent (Khazaal et al., 1994) or simply by the reduction of gas production caused by acetone extracts of the fodders (Wood and Plumb, 1995). Successful replacement of rumen fluid as a source of anaerobic bacteria by faeces from ruminant animals has been reported both for two-stage Tilley and Terry dry-matter digestibility (DMD) (El Shaer et al., 1987) and for the gas production technique (Davies, 1991). The use of faeces based inoculum +/- polyphenol adsorbents was therefore tested and optimized for fallen paulownia and poplar leaves using the gas production technique (the trees were commercial hybrids of Paulownia and Populus species respectively). A series of trials was carried out to: develop the use of preliminary incubation of the faeces inoculum; optimize the adsorbent type, molecular weight (MW) and quantity; optimize the faeces concentration in the inoculum and the use of glucose as a pre-incubation stimulant; and evaluate the use of differences in Tl /2 for gas production and end DMD +/- adsorbent to assess the effect of anti-nutritive compounds (ANCs) on the relative rate and extent of fermentation.

All trials were carried out using 100 ml plastic vaccine flexibottles as containers maintained at 39°C in water baths. Gas was manually extracted with disposable syringes, generally at 3, 6, 9,12,15,18, 21, 24, 30, 36, 42, 48, 60, 72 h and bottles were well shaken at these times.

A significant portion of ruminant production in the tropics relies on the grazing of native grasses growing in acid, low-fertility soils. Under these conditions, animal production is generally limited by the low quality and availability of forage supply. The introduction of forage legumes into grazing systems is commonly accepted as a way to alleviate this problem. However, many of the tropical forage legumes adapted to acid soils and selected on the basis of good agronomic performance have limited feeding value. This could be associated with increased production of secondary metabolites such as condensed tannins (CT), which have the ability to bind and precipitate proteins, carbohydrates and other molecules (Mueller-Harvey and McAllan, 1992). High intake of CT by ruminants has been associated with depressed intake and reduced digestibilities of protein and cell wall of temperate (Barry and Duncan, 1984; Pritchard et al, 1988) and tropical (Carulla, 1994) forage legumes. However, more recent work showed that intake and digestibility of tropical legumes was not only influenced by CT concentration but also by cell wall content (Barahona et al, 1997). In the ongoing process of identifying legume germplasm for acid soils, the use of in vitro techniques has been of great benefit. Gas production methods have considerable appeal in this respect due to their ability to evaluate digestion kinetics and their potential to simulate digestion processes in the rumen. Pendong et al. (1996) using temperate forages, showed that the pressure transducer gas technique of Theodorou et al. (1994) could be used to assess digestible organic matter disappearance as well as fermentation kinetics. A similar assessment was undertaken in the current study in which the tannin contents and cell-wall characteristics of tanniniferous legumes from Colombia were determined and related using measurements made in laboratories in Colombia and the United Kingdom.

There is increasing demand to obtain fast and accurate dynamic nutritional information from forages. Near-infrared reflectance spectroscopy (NIRS) offers the possibility for obtaining such information for a range of nutritional constituents of foods. Herrero et al. (1996 and 1997) calibrated in vitro gas production measurements of a single grass species by NIRS. There would be greater practical benefit if the gas production predictions could be obtained using calibrations derived from a wide range of plant species, since a single equation could be used for all forages. The objective of this study was to investigate if in vitro gas production measurements of a broad based sample population could be calibrated by NIRS.

Browse species are important food resources in semi-arid areas, especially during the dry season when the nutritive value of grass is at its lowest. However, browse plants often contain secondary plant compounds which limit their nutritive value. Proanthocyanidins (PAs) (also called condensed tannins) and related flavonoids are a common constituent of woody plants in tropical regions (Mangan, 1988). PAs cause a bitter and astringent taste which lowers food palatability and they also lower the digestibility of proteins and carbohydrates (Jackson et al., 1996). PAs also interfere with current chemical methods that are used for estimating nutritive value of foods (Reed, 1995). Biological assays, especially , in vitro techniques, have a potential to reflect better the nutritive value of foods that contain PAs. The in vitro gas production technique has been found to reliably predict the nutritive value of temperate forages (Makkar et al., 1996). There has been limited research on its efficacy with tropical forages. The experiment reported here was conducted to test the hypothesis that gas production of tropical browse species reflects their content of fibre, protein and/or PAs.

Legumes have a significant role in many farming systems of the tropics and subtropics through their contribution to enhanced nutritive value of the animal diet, biological nitrogen fixation and landscape stability (Humphreys, 1995). The great potential of legumes to increase productivity of livestock is being related with their high content of nutrients, especially protein and other nutrients often deficient in grass or low quality hay diets for ruminants. In Indonesia, legumes are often used for replacement of costly concentrate in ruminants diets. Norton and Poppi (1995) reported that quality of tropical legumes varies between and within species but is generally higher than that of tropical grasses. It is well known that the major factors limiting intake and digestibility are those associated with rate and extent of forage degradation by microbial and physical factors in the rumen, primarily the amount of cell wall constituents and the extent of lignification. Association of polysaccharides of cell wall with lignin hinders attack by microbial enzymes and prevents the physical attachment of bacteria to the cell wall. Menke et al. (1979) reported that the amount of gas released when a food is incubated in vitro in the rumen fluid, is also closely related to digestibility of the food and could be used to predict food intake. There is very little information available on the rumen degradation characteristics and in vitro gas production of the tropical legumes. Objectives of the present study were to assess differences in in vitro degradation and gas production between five commonly used legumes in south Sumatra, Indonesia.

Recently in vitro and in situ techniques have been used for assessing forage nutritive value for ruminants instead of laborious in vivo trials. Although, Blummel and Ørskov (1993) have shown that an in vitro gas production technique gives reliable estimates of forage nutritive value, whether the estimates could be correlated with in situ degradation are not well established. There is little known about the differences in degradation characteristics of neutral-detergent fibre (NDF) and cellular contents (CC: soluble carbohydrates, lipids and crude protein) in the rumen, and also about the relationships between component degradation rate in the rumen and gas production rate for different forages. This is because most in situ kinetic studies have followed the disappearance of insoluble cell wall constituents and comparative studies of in vitro and in situ estimates was not performed sufficiently.

The objective of this experiment was to investigate the relationship between in situ degradation of forage components and in vitro gas production rate for the three forages.

Quantitative expressions of the kinetics of digestion are needed to estimate more precisely the quantity and composition of nutrients digested from feeds and their subsequent efficiency of utilization by the animal (Mertens, 1993). Degradation kinetic parameters are estimated by fitting appropriate models to data resulting from the measurement of either the undigested residue or the fermentation end-products after exposure of the feed to digestion. Kinetic data can be collected using either in vitro or in situ procedures. The in situ porous synthetic fibre bag technique has been extensively used to describe the kinetics of feeds degradation in the rumen. Several in vitro gas production methods have been developed to assess fermentation kinetics. The aim of this work was to study the relationship between rumen degradation parameters of forages determined either by the in situ procedure or by the gas production technique and to evaluate the potential of gas production measurements to assess the extent of dry matter degradation in the rumen.

In addition to assessing rate and extent of gas production from fermenting forages in vitro with rumen micro-organisms, gas production methods (e.g. Theodorou et at, 1994) may be used also to study the degradation kinetics of forage dry matter and its fractions. As the substrate dry matter can be lost only through fermentation or solution, this removes the error inherent in the polyester bag method caused by fine particle losses from bags being deemed part of the ‘soluble’ fraction.

The pressure transducer technique (PTT) of Theodorou et al. (1994) was used to measure gas production from nine tropical forage samples (Table 1). Nine bottles were prepared from each of the forage samples. Two bottles of each forage were harvested after 8, 24 and 48 h of incubation and a further three bottles at 72 h, to determine dry matter (DM) and organic matter (OM) losses. The same nine forages were assessed using the polyester bag method (Mehrez and Ørskov, 1977) to obtain DM and OM disappearance after rumen incubations of 4, 8, 24, 48 and 72 h. The soluble fraction was determined by hand washing. The reproducibility of measured losses, at given times, was examined using concordance (rc) correlation (Lin, 1989) and mean square prediction error (MSPE, Bibby and Toutenberg, 1977). Also the simple exponential model was used to estimate the fractional rate of DM degradation (kd) and asymptote A (%) for each forage and the values obtained using PTT (kd,ptt) and in situ (kd,bag) compared using rc and MSPE.

In horses, forages account for the greatest proportion of the diet (Micol and Martin-Rosset, 1995). As a result, the energy value of forages has to be determined very accurately to meet the horse requirements. In the UFC system (Horse Feed Unit) proposed by INRA (Martin-Rosset et al., 1994), the energy value of forages is closely related to organic matter digestibility (OMD).

In ruminants, OMD has been related to the in vitro gas production after 24 h of fermentation of foodstuffs with rumen fluid (Menke et al., 1979). The prediction of OMD with the gas test method is less accurate than with the nylon bag method but better than prediction performed with in vitro method or from chemical composition of foods (Khazaal et al., 1993 and 1995).

In the UFC system, three routine methods have been performed by our laboratory and proposed by INRA to predict routinely OMD of foodstuffs. OMD can be predicted either from chemical composition (Martin-Rosset et al., 1996a), by a pepsin cellulase method (Aufrère, 1982; Martin-Rosset et al., 1996b) or by near infra-red spectrometry (NIRS) (Andrieu et al., 1996).

The gas test method was subjected to be adapted in horses for predicting OMD. The accuracy of the prediction with the gas test method was compared with the other routine methods.