O número de animais terminados em sistema de confinamento cresce rapidamente no Brasil, concomitante com aumento no teor de amido das dietas utilizadas. A expansão da agricultura e custos mais baixos por unidade de energia digestível proveniente do grão em relação aos volumosos deve acelerar este processo. Os objetivos deste trabalho foram: a) avaliar as perdas de amido em confinamentos comerciais; b) calibrar a metodologia de NIRS para estimativa do teor de amido fecal; c) estudar as relações entre parâmetros fecais (% de amido, MS e pH) com desempenho e eficiência de bovinos Nelore e cruzados recebendo dietas de médio/alto concentrado; d) avaliar o efeito do horário de coleta e da raça dos animais no teor de amido fecal. Foram avaliados os teores de amido fecal de animais de experimentos controlados e de confinamentos comerciais. Resultados de 2.003 amostras para teor de amido fecal foram analisados. Foi observada diferença no teor de amido fecal em função do horário de coleta tanto em experimento quanto em confinamentos comerciais (P < 0,01). Os teores de amido fecal em animais 100% Nelores foram maiores em comparação com o de animais cruzados (P < 0,01). Há necessidade de padronização do horário da coleta e grupo genético para utilizar o teor de amido nas fezes como indicador do uso de amido. Para os dados de experimentos controlados houve efeito de experimento (P < 0,0001), sexo (P = 0,006) e grupo genético (P = 0,0466), sendo maior o teor de amido fecal para machos e animais Nelore, quando comparados com fêmeas e cruzados. A correlação com amido nas fezes foi baixa (r < 0,20) para as variáveis peso médio metabólico (kg0,75), consumo de matéria seca (kg/dia) e eficiência alimentar (kg/kg). Não houve correlação entre teor de amido fecal e consumo alimentar residual (CAR) para novilhas e touros Nelore, demonstrando que talvez o amido nas fezes não seja um boa ferramenta para selecionar os animais mais eficientes em dietas de baixo teor de amido. A correlação entre o teor de amido fecal com pH fecal para animais em confinamentos comerciais apresentou correlação negativa de r = -0,57 e -0,51 quando a fonte de amido foi milho e sorgo, respectivamente. Foi detectada diferença (P =0,0006) no teor de pH fecal em função da fonte de amido da dieta (milho ou sorgo), 6,55 para o milho e 6,04 para o sorgo, sugerindo um menor aproveitamento do amido para os grãos de sorgo, embora o teor de amido fecal para as diferentes fontes avaliadas não tenha sido diferente (P = 0,29). Já para a %MS fecal, houve efeito de fonte de amido (P = 0,06), sendo os teores para milho menores (20,5%) em relação ao sorgo (23,1%). As 1.985 análises de amido nas fezes tanto dos experimentos quanto dos confinamentos comerciais permitiram a obtenção e validação de uma equação de calibração do NIRS para teor de amido fecal. Uma parte dos dados coletados foram utilizados na calibração e outra parte na validação independente da equação. Foram feitas equações para três diferentes conjuntos de amostras: a) todas independente da fonte de amido; b) amostras contendo o milho como principal fonte de amido; c) amostras contendo o sorgo como principal fonte de amido. Todas as equações apresentaram boa habilidade preditiva (R2 > 0,92), alta acurácia (erros sistemático inferior a 2% do teor de amido fecal) sendo que uma única equação pode ser utilizada para estimar o teor de amido fecal independente da fonte de amido da dieta. / The number of animals finished in the feedlot system is rapidly growing in Brazil. At the same time there is an increase in the starch content of the diets. The expansion of grain production and lower costs of digestible energy from grain compared to roughages should accelerate this process. The objectives of this work were: a) evaluate starch losses in commercial feedlots; b) develop the NIRS methodology for fecal starch content; c) study the relationships between fecal parameters (%starch, dry matter and pH) with performance efficiency of Nellore and crossbreed cattle receiving diets of medium/high concentrate; d) evaluate the effect of time at fecal sample collection and animal breed on fecal starch content. The fecal starch content was determined on 2,003 samples. Fecal starch content was affected by time of fecal sample collection both in the controlled experiment and in commercial feedlots (P < 0.01). There was higher fecal starch content in animals 100% zebu compared to B. taurus crossbreds (P < 0.01). There is a need to standardize the time of collection and genetic group to use fecal starch as a management tool. There was an effect of experiment (P < 0.0001), sex (P = 0.006) and genetic group (P = 0.0466), with higher fecal starch content found for males, Nellore, when compared to females and crossbreds. The correlations between fecal starch and metabolic mid weight (kg0.75), dry matter intake (kg/day) and feed efficiency (kg/kg) were low (r < 0.20). There was no correlation between fecal starch content and residual feed intake for Nellore heifers or bulls, suggesting starch content is not a good tool to determine efficient animals fed medium concentrate diets. The correlation between the level of fecal starch and fecal pH for animals in commercial feedlots showed a negative correlation of r = -0.57 and -0.51 when corn and sorghum were the sources of starch. Fecal pH was different (P = 0.0006) depending on the source of the dietary starch, 6.55 and 6.04 for corn and sorghum respectively, suggesting lower efficiency of starch use with sorghum. However there was no difference in fecal starch percentage for the two starch sources evaluated (P = 0.29). For the fecal DM, there was an effect of source of starch (P = 0.06), lower for corn (20.5%) when compared to sorghum (23.1%). The database obtained with the 1,985 lab analysis of fecal starch from both controlled experiments and commercial feedlots were used to develop a NIRS calibration equation for fecal starch content. Part of the analysis was used to independently validate the equations. Equations were developed from three different sets of samples: a) all samples regardless the starch source; b) samples where corn was the main source of starch c) samples were sorghum was the main source of starch. All equations had good predictive ability, with high regression coefficients (R2 > 0.92) and low bias (< 2%) indicating high accuracy with independent data. Results suggest a single equation can be used to estimate the fecal starch regardless the source of starch in the diet.

A transferência de imunidade passiva colostral é muito importante tanto para a saúde do neonato quanto para o seu desempenho. Com o objetivo de estudar a relação entre a eficiência ou não no estabelecimento da imunidade passiva colostral em neonatos e a suas interrelações com aspectos de desempenho e produção destes animais foram colhidas amostras de sangue de 354 bezerras com 30 horas de vida (24 a 36 h) em 10 propriedades localizadas no centro leste do Estado do Paraná, Brasil. Determinações bioquímicas do soro das amostras foram realizadas para avaliar a qualidade de transferência de imunidade passiva. A proteína sérica total foi determinada pelo método do Biureto e a albumina pelo método do verde de bromocresol para o cálculo das taxas de globulinas obtidas pela diferença entre os teores séricos de proteína e albumina. A fração gamaglobulina foi determinada por eletroforese. Uma análise multivariada, incluindo os teores séricos de proteína total, globulinas e gamaglobulinas, foi utilizada para determinar três grupos conforme à qualidade da transferência de imunidade passiva colostral apresentada (Proc cluster, SAS), a saber: baixa, moderada e alta. Durante o estudo, os produtores anotaram as informações referentes a práticas de manejo adotado na atenção a vaca (parturiente) e ao bezerro, incluindo os dados que poderiam ter influência sobre a transferência de imunidade passiva colostral. Os dados das bezerras, enquanto neonatas e depois quando adultas foram colhidos para a determinação da influência a curto e a longo prazo da falha da transferência de imunidade passiva (FTIP) colostral. Observou-se que os dados de escore de condição corporal da mãe da bezerra, da quantidade de colostro ingerida na primeira mamada e o momento em que ela foi realizada, da morbidade e da mortalidade das bezerras estavam relacionados com a FTIP (P<0,05). Os pesos ao nascer e ao primeiro mês de vida, assim como a freqüência de bezerras analisadas que pariram na propriedade não estavam relacionados diretamente com a qualidade da transferência de imunidade passiva. Fatores como a distocia, idade ao primeiro serviço e a produção média diária de leite não diferiram estatisticamente entre os grupos de baixa, moderada e alta transferência de imunidade passiva (P>0,05). Mas quando se correlacionou por regressão os dados de produção de leite das vacas que quando bezerras apresentaram teores de gamaglobulinas menores que 1,6 g/dL obteve-se valores de r2=0,47 (P=0,0005). Com os resultados desta pesquisa pode-se afirmar que práticas muito simples de manejo como fornecer pelo menos 2 L de colostro até 2 horas após o nascimento e a vaca parir numa boa condição corporal (ECC = 3 ou 3,5) podem prevenir a falha na transferência de imunidade passiva. A FTIP tem como conseqüências: maiores taxas de morbidade e mortalidade, primeiro parto mais tardio e diminuição do número de novilhas de reposição, podendo ainda estar correlacionada a menores produções leiteiras. / Adequate passive transfer of maternal immunoglobulin is important for optimal health and performance in newborn calves. Blood samples were collected from 354 dairy calves, ranging from 24 to 36 hours of age, between July 2005 and May 2006 on 10 farms in the middle-eastern region of the state of Parana, Brazil. The objective was to study the relationship and effectiveness of the transfer of colostral passive immunity and its contributing factors as related to the development and production of animals. For each sample collected, total serum protein was determined by the biuret method and albumin by bromocresol green method, and the difference was used to evaluate the globulins. Electrophoresis was used to determine the -globulin fraction of the sample. A multivariable analysis, including total serum protein, globulins and gamma globulin, was used to create 3 groups to classify the quality of the transfer of colostral passive immunity (cluster procedure, SAS): failure or inadequate group, marginal group and adequate group. During the study, breeders were asked to provide information on calf and pre-partum cow management practices, including details on colostrum feeding. Data from the calves while newborn and as heifers was gathered to determine the long and short term effect of the failure of passive immunity transfer (FPIT). Body condition score of the mother at calving, quantity of colostrum ingested, timing of ingestion, morbidity and mortality of calves and age at calving time were related to FPIT (P<0,05). The weight of the calves after birth and at one month of age and the frequency of calves that became cows in the farm were not directly related to failure. Dystocia, age at first service in days, and milk production did not differ statistically (P>0,05). However, when a regression was performed based on data of milk production from calves that had serum gamma globulins levels below 1,6g/dL, a correlation was identified (r2=0.47; P=0.0005). Basic management practices can prevent failure of passive immunity transfer by feeding calves 2 L of colostrum within 2 hours of life and ensuring that the cow calves with a good body condition score (BCS = 3 or 3.5). FPIT is responsible for higher morbidity and mortality rates, a delay in first parturition, a decrease in the number of replacement heifers and it can also be responsible for less milk production.

Objetivou-se avaliar o efeito da inclusão de óleo de soja e óleo de linhaça na dieta de vacas em lactação sobre a produção, composição de leite e parâmetros sanguíneos. Foram utilizadas 18 vacas de raça Holandês, com produção média de 25 ± 4 kg/dia em 153 ± 62 dias em lactação. As vacas foram distribuídas em delineamento quadrado latino, com 6 quadrados contemporâneos, 3 períodos e 3 tratamentos e arraçoadas de acordo com os seguintes tratamentos: 1) Controle (CT): sem adição de óleo; 2) Óleo de Soja (OS): 5% de adição de óleo de soja no concentrado como fonte de ômega-6 (% MS); 3) Óleo de Linhaça (OL): 5% de adição de óleo de linhaça no concentrado como fonte de ômega-3 (% MS). A produção de leite não foi alterada mas a suplementação com óleo diminuiu o consumo de matéria em 1.93 kg/dia/vaca. A produção de gordura foi significativamente menor quando as vacas receberam suplementação com óleo vegetal (3,37, 2,75 e 2,89 % para CT, OS e OL, respectivamente). No entanto, ambos os óleos de soja e linhaça diminuíram a concentração de ácidos graxos saturados (65,72, 54,09 e 54,26 para CT, OS e OL respectivamente) e aumentaram a quantidade de ácidos graxos insaturados no leite (34,24, 45,91 e 45,74 para CT, OS e OL respectivamente) e diminuíram a relação entre os ácidos graxos saturados/insaturados (18,28, 10,41 e 4,52 para CT, OS e OL respectivamente). Além disso, o óleo de soja e de linhaça aumentaram significativamente a teor de ácidos graxos mono (24.95, 30.89 e 31.61%, para CT, OS e OL respectivamente) e poli-insaturados (2,4, 3,0 e 2,97 para CT, OS e OL respectivamente) na gordura do leite. As variáveis sanguíneas aspartato aminotransferase, gama glutamil transferase, ureia, albumina, creatinina e proteínas totais não foram alteradas. Por outro lado, o colesterol total, HDL e LDL foi maior no grupo suplementado com os óleos vegetais. A suplementação com óleos vegetais ocasionou diminuição do consumo de matéria seca, melhorou relação &omega;6/&omega;3 do leite, sem alterar profundamente as variáveis sanguíneas de vacas lactantes. / The use of polyunsaturated oils in the ruminant diet has been shown to be an efficient alternative to improve the lipid composition of cows\' milk, aiming to improve human health. The objective of this study was to evaluate the effect of soybean oil and linseed oil in the diet of lactating cows on the production, milk composition and blood parameters. 18 Holstein cows were used, with a mean production of 25 ± 4 kg / day in 153 ± 62 days in lactation. Animals were distributed in a Latin square design, with 6 contemporary squares, 3 periods and 3 treatments, and ratified according to the following treatments: 1) Control (CT): without addition of oil; 2) Soybean oil (SO): 5% addition of soybean oil in the concentrate as a source of omega-6 (% DM); 3) Flaxseed oil (FO): 5% addition of linseed oil in the concentrate as a source of omega-3 (% DM). Milk yield was not altered but the supplementation with oil decreased the DM intake by 1.93 kg/day. Fat production was significantly lower when cows were supplemented with vegetable oil (3.37, 2.75 and 2.89% for CT, SO and LO, respectively). However, both soybean and linseed oils decreased the concentration of saturated fatty acids (65.72, 54.09 and 54.26 for CT, SO and LO respectively) and increased the amount of unsaturated fatty acids in milk (34.24, 45.91 and 45.74) and decreased the relationship between saturated / unsaturated fatty acids (18.28, 10.41 and 4.52). In addition, soybean and flaxseed oil increased significantly in monounsaturated fatty acids (24.95, 30.89 and 31.61% for CT, SO and LO respectively) and polyunsaturated fatty acids (2.4, 3.0 and 2.97) in milk fat. Blood variables aspartate aminotransferase, gamma glutamyl transferase, urea, albumin, creatinine and total proteins were not altered. On the other hand, total cholesterol, HDL and LDL were higher in the group supplemented with vegetable oils. Supplementation with vegetable oils caused a reduction in dry matter intake, improved n-6/n-3 ratio of milk, without profoundly altering the blood variables of lactating cows.

I evaluated some of the factors affecting livestock distribution by conducting experimental (Chapters II and III) and observational (Chapter IV) studies. In Chapter II, I described the effect of locations of familiar foods and social interactions on choice of feeding location by lambs. Lambs were exposed to a pasture as subgroups of strangers and companions with different dietary habits (i.e., three lambs that preferred milo with three lambs that preferred wheat). Milo was placed on one end and wheat on the other, about 100 m apart. Strangers typically fed in different locations, reflecting dietary preferences. Conversely, companions fed in both single subgroups and in separate subgroups because both social interactions and dietary preferences affected choice of feeding location. The objective of Chapter III was to evaluate the effect of experience with a pasture on choice of feeding location. Lambs with different levels of familiarity with the pasture were exposed as subgroups with different dietary habits. Lambs familiar with the pasture typically fed in separate locations, reflecting dietary preferences. Lambs naive to the pasture always fed in one subgroup and consumed both foods because social interactions overrode dietary preferences. Results in Chapter IV describe cattle observations on a 1,030-ha grazing allotment. Cattle home ranges were similar in location (i.e., site fidelity) between 1990 and 1991, even though home ranges increased in size and (or) shifted in location in response to water availability. Moreover, forage availability did not account for changes in home range size or location. Site fidelity probably developed because of experiences early in life while foraging with mother and (or) peers. Managers may be able to improve distribution by manipulating foraging experiences. Placing familiar foods/supplements in underutilized areas, controlling the amount of experience livestock have with different habitat types, and culling animals that spend a disproportionate amount of time in riparian zones may improve · distribution. Herding could also improve distribution. Herding integrates social interactions and experience with foods by controlling the exposure of social groups to particular foods and habitats. Nevertheless, livestock may still spend considerable time in riparian zones unless other watering points are available.

"January 2003" / Includes bibliographical references (leaves 172-204) / xx, 204 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, Dept. of Animal Science, 2003

Includes bibliographical references (leaves 202-21). / 2 v. (xxiv, 409 leaves) : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Experiments were designed for the early detection of vitamin B12 deficiency in sheep and cattle. Metabolic indicators were examined in animals where deficiency was induced with either nitrous oxide (N2O) or by feeding diets of low cobalt content. Results showed that vitamin B12 dependent enzymes in the liver of ruminants appeared to be affected only at a late stage of deficiency. However changes in concentrations of metabolites of these pathways occured earlier. / Thesis (Ph.D.)--University of Adelaide, Dept. of Animal Sciences, 1994

Bibliography: leaves 107-130. This study evaluates quality of fermentation of the silage and also its nutritive value to ruminants on the basis of "in vitro" and "in sacco" digestability. Animal response to the silage is also evaluated in an "in vivo" digestability and nitrogen balance trial with Australian Merino sheep. The objective of this thesis is to investigate the opportunities for utilising citrus pulp and poultry litter to improve the nutritive value of wheat straw.

Five trials were conducted to assess the nutritive value of high fiber canola meal (HFCM) as a protein supplement for ruminants. Trial one determined the chemical composition of HFCM relative to low fiber canola meal (LFCM) and canola meal in a completely randomized design using samples obtained from five different crushers. In the second trial, the in situ nylon bag technique was used to determine disappearance of dry matter (DMD), crude protein (CPD), neutral (NDFD) and acid (ADFD) detergent fiber from canola meal, LFCM and HFCM samples derived from five different crushers following 24 h of nunen incubation. The effect of crushing plant of origin on in situ nutrient disappearance in the three meals was also determined. In the third trial, rumen nutrient kinetic parameters and effective degradabilities were determined for the blended canola meal, LFCM and HFCM. In situ disappearance of different amino acids canola meal, LFCM and HFCM following 12 h of rumen incubation was determined in the fourth study. The fifth trial was designed to determine nutrient digestibility coefficients by growing lambs of seven diets containing dehydrated alfalfa and 0, 25, 50 and 75% HFCM or canola meal in a randomized complete block design. Milk yield and composition responses of early to mid lactation dairy cows to diets supplemented with HFCM, canola meal or soybean meal were also determined. Results of the first trial indicated that tail-end dehulling of canola meal resulted in more fractionation of fiber than protein between HFCM and LFCM. Results of the second trial showed differences in CPD and DMD within LFCM and HFCM but not within canola meal samples. However, no meal by crusher interaction was observed indicating that in situ nutrient disappearances in the three meals were consistent across crushers. At 5% h^-1 rumen flow rate, effective DM degradability (trial three) was higher (P $<$ 0.05) in LFCM than in canola meal and was higher in canola meal than in HFCM while effective CP degradability was higher in LFCM and canola meal than in HFCM. Effective NDF degradability was higher (P $<$ 0.05) in LFCM than in canola meal and MFCM. However, effective ADF degradability was higher in LFCM than in canola meal and was higher in canola meal than in HFCM. Tail-end dehulling had little effect on in situ disappearance of amino acids from canola meal, LFCM and HFCM following 12 h of rumen incubation. Results from the fifth trial indicated that at 75% inclusion rate, the diet containing HFCM had lower (P $<$ 0.05) DM, CP and gross energy digestibility coefficients relative to canala meal diet. Results of the dairy trial showed no treatment effect on feed intake. (Abstract shortened by UMI.)

Nematode parasites have become the biggest problem for small ruminant production in South Africa due to their resistance to the commercial anthelmintics. Notable, wattle tannin has been used as an alternative strategy for control. However, the concentration and the frequencies can likely influence its effect on the parasites control. The objective of this study was to determine the degree of pasture infestation and nematode infection in sheep and goats, as well as investigate nematode resistance to the anthelmintics, and the potential of wattle tannin in nematode control. The first study dealt with the epidemiology of internal parasites. Eight Merino ewes and eight Nguni does averaged 7-18 months of age were observed for 1 year during the months of February 2008 to January 2009 at the University of KwaZulu-Natal Research Farm (Ukulinga). Egg count per gram (EPG) and coccidian oocysts per gram (OPG) were counted according to Mc Master Technique (Hansen & Perry, 1994) by magnifying parasitic eggs from monthly rectal faecal samples dissolved in saturated sodium chloride. Faecal samples also were cultured for 15 days to identify infective nematode larvae (L3) using Baermann technique. Herbage samples were collected monthly from four paddocks as well to count L3 on the pasture. Sheep live weight was also recorded monthly. Seasonal effects was significant (P<0.05) on the EPG, OPG, faecal culture L3 and pasture L3. A higher level of infection was observed in summer (wet) than in winter (dry season). Trichostrongylus spp larvae were the most prevalent larvae (26.5%) while Strongyloides, Haemonchus contortus, Nematodirus and Cooperia spp occurred in the faecal culture by percentage of 20.9%, 16%, 16% and 14.5%, respectively. For parasite resistance, Ivermectin 1% (IVM), Closantel 5% (CST) and a combination of Abamectin 0.08% and Praziquantel 1.5% (CPA) were evaluated. Twenty four sheep (12 females and 12 males) aged between 7-18 months were used for 21 days. Animals were naturally infested by gastro-intestinal parasites. EPG and faecal culture L3 were counted on day 0, 7, 14 and 21. Closantel was the most effective. Haemonchus spp. were least affected whilst Trichostrongylus spp. were the most affected by all drugs. In the third study, wattle tannins were evaluated as an alternative nematode control drug. Three experiments (Exp.) were conducted to determine the effect of tannin concentration (Exp.1 and 2) and frequency (Exp.3) on nematode parasites. In Exp.1, 0, 0.8, 1.6 and 2.4 g tannin/kg BW were drenched for three consecutive days per sheep (16 females and 8 males, aged 8-9 months) for 21 day. In Exp.2, 30 sheep (14 males and 16 females, aged 9-18 months) were randomly allocated into three tannin treatments (0, 0.8 and 1.6 g tannin/kg BW) and drenched for a day. In Exp.3, 26 sheep (11 males and 15 females aged 9-18 months) were divided into three groups of 9, 9, and 8 sheep each. These groups were drenched with 1.6 g tannins/kg BW/day; once, twice or thrice for the 3 groups respectively. For the three experiments, EPG and L3 larvae were counted in individual feacal samples. For all tannin treatments, EPG decreased (P<0.05) over time. Though the differences among tannin levels and frequencies varied (P<0.05) over time, EPG consistently decreased with increasing tannin level and frequency. Thus 1.6 and 2.4 g tannin /kg BW for 3 consecutive days had nearly similar effects on the EPG. The results of this study are rather inconclusive that weather conditions such as rainfall had a direct effect on internal parasites development. Nematode parasites at Ukulinga Research Farm were resistant to the commercial anthelmintics used. Drenching with 1.6g wattle tannin/kg BW over three successive days is enough to reduce EPG and reduce the degree of pasture contamination. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2010.

Le méthane (CH4) est un des principaux gaz à effet de serre. L'élevage est à l'origine d'un tiers du CH4 produit par l'activité humaine en Europe. En plus, la production de CH4 représente une perte de 2% à 12 % de l'énergie consommée par l'animal. La méthanogenèse est le résultat de l'activité d'un groupe de microorganismes particuliers - les Archaea méthanogènes. La production de CH4 permet de d'éliminer du milieu ruminal l'hydrogène produit au cours de la fermentation des aliments par les autres microorganismes (bactéries, protozoaires, champignons). En effet, l'accumulation d'hydrogène affecte le fonctionnement optimal du rumen. La réduction des émissions de CH4 par les ruminants présente donc un intérêt économique et environnemental non négligeable et passe inévitablement par une modification de l'écosystème microbien du rumen. L'objectif de ce travail de thèse était de relier la production de CH4 avec la structure et l'activité de la communauté méthanogène du rumen. Différents modèles de manipulation de l'écosystème microbien ruminal comme la défaunation (élimination des protozoaires) et l'utilisation d'aliments connus pour modifier la méthanogenèse ont été utilisés. Le rumen étant un écosystème complexe, les interactions fonctionnelles entre les Archaea méthanogènes et les autres microorganismes présents (bactéries et protozoaires) ont également été étudiées. Dans cette optique, des outils de biologie moléculaire, permettant de cibler les principales communautés microbiennes, ont été optimisés. Nos travaux permettent de conclure sur l'absence de relation claire entre le nombre (et/ou la concentration) des Archaea méthanogènes et la méthanogenèse dans le rumen. Cependant les réductions des émissions de CH4 ont été attribuées aux changements dans la diversité de la communauté méthanogène et la disponibilité en hydrogène. Ce travail de thèse a mis en évidence que les modifications de la composition et/ou de l'activité métabolique de la communauté des Archaea méthanogènes seraient à l'origine des réductions des émissions de CH4 par les ruminants. Une meilleure connaissance des mécanismes microbiens impliqués dans la production de méthane permettra d'envisager de nouvelles pistes pour diminuer les émissions chez les ruminants.