Livestock Research for Rural Development 33 (6) 2021 LRRD Search LRRD Misssion Guide for preparation of papers LRRD Newsletter

Citation of this paper

The rumen in vitro incubation system as a tool for predicting the nutritive value of ruminant diets and the associated emissions of methane

T R Preston, R A Leng1, Sangkhom Inthapanya2 and María E Gomez Z3

Centro para la Investigación en Sistemas Sostenibles de Producción Agropecuaria (CIPAV), Carrera 25 No 6-62 Cali, Colombia
1 University of New England, Armidale, NSW, Australia
2 Soiphanuvong University, Laos PDR
3 Cra 24ª, 3-74, Cali, Colombia


It is proposed that the in vitro rumen incubation system, configured to measure total gas and methane, after a 24-hour fermentation, and combined with a measure of protein insolubility, are effective indicators of the relative carbon footprint, as well as the required balance of nutrients, for efficient milk and meat production from ruminants in tropical countries.

Keywords: bypass protein, cecum fermentation, global warming, glucose precursors


The criteria that should govern the design of ruminant feeding systems are that they support the production of meat and milk and the capacity to reproduce at the maximum rate according to their genetic merit of the targeted animals. When these criteria are met, emissions of methane will be minimized (Figure 1).

Figure 1. Relationship between live weight gain and enteric methane production
per unit live weight gain (Klieve and Ouwerkerk 2007)

The first step is therefore to select the appropriate feeding system that will ensure animals perform at the upper limit of their genetic capacity.

It is frequently argued that in developing countries the available feed sources are of such “poor quality” that they barely support maintenance. This is not true as even the poorest feed resource (eg: cereal straw) can be upgraded by a combination of chemical treatment of the straw and supplementation with bypass protein. The example in Figure 2 shows the application of this strategy to a basal diet of wheat straw in China. The strategy employed by the researchers is an example of the appropriate methodology that should be employed. namely, a “production function” design that relates the output to the level of the critical input, in this case, the level of cottonseed cake. The key feature of this design is that it demonstrates the “diminishing response to the critical input which was the level of the cottonseed cake.

In our first paper in this series (Preston et al 2021) we argued that to respond to our obligation to take all step needed to reduce carbon in the atmosphere, protein requirements for ruminants should be sourced from the leaves of trees and shrubs. As leaves provide both protein and fibre the energy resource should be low in fibre. However, we should have qualified this statement with the clarification that trees that produce fruits are also part of the strategy. Relevant trees in this category are the oil palms, producing palm kernel cake and coconut trees producing coconut cake. Replacing rain forest with oil palm in Indonesia was strongly criticized for the disruptive effects on biodiversity. The future path should be to promote oil palm as a replacement for grasses.

Figure 2. Fattening cattle with ammoniated wheat straw and cottonseed cake (Weixian et al 1994)
Sources of nutrients

The criteria that govern the selection of feed resources as sources of energy and protein were discussed by Preston and Leng (1987) as those that give rise to essential amino acids and precursors of glucose at sites of metabolism. Fortunately, the feed resources that satisfy these requirements are also those that will produce the least methane (Preston et al 2021a, b). We will show later that the in vitro rumen incubation system is an appropriate procedure for ranking feed resources as sources of energy-rich nutrients at sites of metabolism.

In countries where agro-industries have been established (sugar from sugar cane; starch from cassava) there will the byproducts rich in energy that can be used to balance the protein derived from trees and shrubs. These byproducts are cassava pulp from the production of starch, and molasses from sugar cane. When these industries are not established, cassava is the preferred crop to grow, as starch is superior to sugar as an energy source for ruminant, animals and the foliage that is produced by cassava is an excellent source of bypass protein (Promkot and Wanapat 2003).

Bypass protein

The trees-shrubs that have been identified and corroborated as reliable sources of bypass protein include: Leucaena leucocephala, Indigofera sp, Gliricidia sepium, Erythrina spp., Guazuma ulmifolia, Morus alba and cassava (Manihot esculenta). These species have been evaluated under practical conditions and can be recommended as the source of bypass protein. At the same time, they provide the minimum levels of fibre needed to avoid “bloat” and to ensure efficient rumen function. In this respect, tree-shrub foliages should be grazed or offered as branches to facilitate the selection of the leaves (and rejection of the more fibrous elements). Mechanical chopping of foliage is not recommended as it forces the animal to eat elements that it would refuse in a grazing environment. The animal is wiser than the machine!!

The next step is to identify a suitable source of fermentable/digestible carbohydrate-rich in starch or sugar. For farms located close (30km radius?) to agro-industries, the indicated byproducts are cassava root pulp and molasses. These byproducts are low in crude protein (<3% in DM) and should be supplemented with urea as a source of rumen ammonia which needs to be about 20 m/100 ml rumen fluid (Preston and Leng 1987).

Appropriate systems of analysis

The desirable characteristics of feeds that need to be determined are: (i) the potential value of the feeds as a source of energy (digestibility /fermentability); (ii) the solubility of the protein; (which will indicate its potential to escape from [bypass] the rumen fermentation for more efficient enzymic digestion in the small intestine); and (iii) the key elements in the end-products of the rumen fermentation (eg: methane and ammonia). Ammonia levels should be of the order of 200mg/litre (Preston and Leng1987) to ensure efficient growth of rumen bacteria) while methane should be as low as possible as it provides no energy to the animal and more importantly has 20 times more heat- forcing capacity than equivalent amounts of carbon dioxide.

These characteristics of feeds, as discussed above, cannot be estimated by the classical “proximate analysis system”. It is, therefore, appropriate to discuss ways in which the traditional methods of proximal analysis can be supplemented/replaced by more appropriate methods of feed analysis that will provide specific guidance as to how feeding systems can be modified to optimize ruminant animal productivity and reduce greenhouse gas emissions.

Our experience over the last 10 years indicates that for feed resources destined for ruminants the in vitro rumen incubation yields the most useful data, provided, that the feed in question is evaluated as a component of a balanced feeding system and not as an isolated entity.

The in vitro rumen system

In vitro rumen incubation systems have evolved over the 60 years since the method was first reported (Terry and Tilly 1963). The availability of infra-red gas detection meters has facilitated the direct measurement of methane in the gaseous products of fermentation. The recommended method (Photo 1) was developed during research in Laos and Vietnam over the period 2010-2021).

The description of the system and the method of operation is available via the link Sangkhom in vitro.

The incorporation of direct methane measurement in the in vitro system has highlighted two important issues: Over what period should measurements be made? And what is the significance of the loss in DM (DM mineralized) during the period that the substrate stays in the “in vitro” rumen? (as determined by the loss in DM during the process of fermentation)?

The answer to the first question is indicated by the data derived from measuring the methane content of the gas at 6h intervals from 6 to 48h (Figure 3); the methane concentration in the gas increased linearly with the length of the incubation over the period 0 to 48h. The recommendation is to stop the fermentation after 24 hour which corresponds to the average time of retention in a normally functioning rumen. Beyond 24 hours the in vitro rumen begins to acquire the characteristics of a “biodigester”.

Photo 1. The in vitro system uses recycled water bottles for the incubation and gas collection by water displacement (Inthapanya et al 2019)

Figure 3. The methane content of the gas increa7ses linearly with
the length of the incubation (Inthapanya et al 2011)
Figure 4. Effect of increasing substrate fermentability on the methane
concentration in the gas (Preston et al 2019)

There are two ways to measure the progress of events during the proposed 24 hours of incubation: (i) by measuring DM loss from the substrate; and (ii) by the cumulative production of gas during the same 24-hour period. The former measures the mineralization (solubilization) of the substrate and simulates the conventional procedure of the measurement of digestibility. The latter by definition measures that part of the substrate converted to gaseous end-products (Figure 4).

The results of the in vitro incubation described in Figures 5 and 6 are an example of how the in vitro system identifies important characteristics of different sources of carbohydrate and protein Sugar is more fermentable than cassava root but the gas it produces is richer in methane when supplemented with the same level and source of protein and NPN.

Figure 5. Effect of source of carbohydrate and increasing proportions of bitter
replacing sweet cassava leaf on the methane content of the gas
in an in vitro rumen incubation (Phuong et al 2020)
Figure 6. Effect of the substrate (cassava root or sugar)
and source of cassava leaf (sweet or bitter) on DM
solubilized in 24h (Phuong et al 2020)

Because of the negative relationship between methane and propionic acid (Figure 7), the end-products of the fermentation of cassava would be of higher nutritive value. In this case, the classical “proximate” analysis system would have predicted the opposite result.

Figure 7. The negative relationship between methane production and
propionic acid in the rumen of cattle (Whitelaw et al 1984)

This example shows how the in vitro rumen incubation system provides relevant information about the extent and nature of the rumen fermentation; it is apparent that the substrate (diet) with the energy in the form of starch was only slightly less digestible than when sugar was the energy source. In contrast, the composition of the gas showed major differences with the gas from the sugar diet having almost 100% more methane than when the energy source was starch derived from the cassava root.

Protein solubility

The other major criterion of feed nutritive value for ruminants is the protein content and specifically its relative solubility as this will determine the extent to which this protein will escape the rumen fermentation and be available for more efficient enzymatic digestion in the intestine. This protein in many cases will be linked with fibrous entities especially when it is sourced from the leaves of trees and shrubs. This residual fibre can be fermented in the cecum and large intestine where the end product of fermentation will be acetic acid and not methane (Dmeyer 1991; Immig 1996; Popova et al 2013).

Two factors determine the nutritive value of dietary protein for ruminant: s the first is the amino acid balance and especially the concentration of lysine and methionine. The second factor is the solubility of the protein. The interaction between these two factors is demonstrated in the research presented in Figure 8. Groundnut meal (providing protein of low biological value) was heated to reduce the solubility of the protein and compared with fish meal (low solubility; high biological value) and the same fish meal after it had been hydrolyzed to make the protein soluble. The four sources of protein with contrasting degrees of solubility and amino-acid balance were compared in maize-based diets fed to early-weaned calves. Retention of nitrogen was highest when the protein source was fish meal of low solubility and lowest when the fish meal was of high solubility.

Figure 8. Protein sources that have higher biological value and lower
solubility (Insol FM) ) support the highest N retention in young,
early-weaned calves (Whitelaw and Preston 1963)
Figure 9. Methane percentage in the gas increase linearly with
increasing solubility of the protein (Preston et al 2019)

There is no conflict between high levels of ruminant productivity and the carbon footprint of the feeding system. High levels of growth or milk production require feeds rich in starch and “bypass” protein (eg: protein of low solubility). This strategy will decrease enteric methane because starch-rich diets are fermented to VFA rich in propionic acid and hence produce less methane (Figure 7); while diets rich in bypass protein (ie: protein of low solubility) also produce less methane (Figure 9).

Rate of gas production and methane content of the gas

It has been observed in several in vitro incubations in which dietary manipulations have affected the rate of gas production, that there was a positive relationship between the rate of gas production and the methane content of the gas (Figure 9; Inthapanya et al 2019, 2020; Sina and Preston 2021). A similar relationship existed between the solubility of the protein supplement (and therefore of the fermentability of the substrate) and the percentage of methane in the gas (Figure 9). To generalize: the more fermentable is the substrate in the rumen, the higher is the content of methane in the gas. Substrates that have more of the protein “insoluble” form will therefore produce less methane.



The research discussed in this paper was supported by the MEKARN III project financed by the Swedish International Development Authority (SIDA).


Demeyer D 1991 Differences in stoichiometry between the rumen and hindgut fermentation. Adv. 1023 Animal Physiology and Animal Nutrition 22:50-66

Immig I 1996 The rumen and hindgut as a source of ruminant methanogenesis. Environmental Monitoring and Assessment Volume 42, Issue 1-2, pp 57-72

Inthapanya S, Preston T R and Leng R A 2011 Mitigating methane production from ruminants; effect of calcium nitrate as a modifier of the fermentation in an in vitro incubation using cassava root as the energy source and leaves of cassava or Mimosa pigra as a source of protein. Livestock Research for Rural Development. Volume 23, Article #21.

Inthapanya S, Preston T R, Ngoan L D and Phung L D 2019 Effect of a simulated rice distillers’ byproduct on methane production in an in vitro rumen incubation of ensiled cassava root supplemented with urea and leaf meal from sweet or bitter cassava. Livestock Research for Rural Development. Volume 31, Article #164.

Klieve A V and Ouwerkerk, D 2007 Comparative greenhouse gas emissions from herbivores. In: Q.X. Meng, L.P. Ren and Z.J. Cao, Proceedings of the 7th International Symposium on the Nutrition of Herbivores. 7th International Symposium on the Nutrition of Herbivores, 17-22 September, Beijing, China, (487-500). 2007.

Leng R A 2016 Immuno-modulation of the ecology of gut biofilm microbiota may determine the extent of either methanogenesis or acetogenesis in the processes of fermentative digestion in ruminants, macropod marsupials and equids. Submitted for publication

Phuong L T B, Preston T R and Leng R A 2020 Effect of contrasting sources of carbohydrate (sugar or starch) and bypass protein (leaf meal from sweet or bitter cassava) on methane production in a rumen in vitro incubation. Livestock Research for Rural Development. Volume 32, Article #70.

Popova M, Morgavi D P and Martin C 2013 Methanogens and methanogenesis in the rumen and cecum of lambs fed two different high concentrate diets Applied Environmental. Microbiology doi:10.1128/AE

Preston T R and Leng R A 1987 Matching ruminant production systems with available resources in the tropics and sub-tropics Penambul Books Armidale (2009 New online edition;

Preston T R, Silivong P and Leng R A 2019 Methane production in rumen in vitro incubations of ensiled cassava (Manihot esculenta Cranz) root supplemented with urea and protein-rich leaves from grasses, legumes and shrubs. Livestock Research for Rural Development. Volume 31, Article #112.

Preston T R, Leng R A and Gomez M E 2021 Adapting systems of livestock production to be compatible with global commitments to restore the health of planet Earth; ecosystems that remove atmospheric carbon and provide, food, feed and renewable energy. Livestock Research for Rural Development. Volume 33, Editorial. Retrieved April 4, 2021, from

Promkot C and Wanapat M 2003 Ruminal degradation and intestinal digestion of crude protein of tropical protein resources using nylon bag technique and three-step in vitro procedure in dairy cattle. Livestock Research for Rural Development. Volume 15, Article #81.

Sina V and Preston T R 2021 Methane production in vitro rumen incubation of ensiled cassava root and urea is reduced by supplementation with low levels of Brewers’ spent grains and by supplementation with cassava foliage rather than water spinach. Livestock Research for Rural Development. Volume 33, Article #45.

Tilley J M A and Terry R A 1963 A two-stage technique for the in vitro digestion of forage crops. Journal of the British Grassland Society. 18: 104

Weixian Z, Xue G C, Dolberg F and Finlayson P M 1994 Supplementation of ammoniated wheat straw with hulled cottonseed cake. Livestock Research for Rural Development. Volume 6, Article #9.

Whitelaw F G and Preston T R 1963 The nutrition of the early-weaned calf III. Protein solubility and amino acid composition as factors affecting protein utilisation. Animal Production, 5, pp 131-145 doi:10.1017/S0003356100021620

Whitelaw F G, Eadie J M, Bruce L L A and Shand W J 1984 Methane formation in faunated and ciliate-free cattle its relationship p with rumen volatile fatty acid proportions. British Journal of Nutrition 1984 1984 Sep;52(2):261-75. PMID: 6433970. 1984 Sep;52(2):261-75. doi: 10.1079/bjn19840094.