The objective of this study was to evaluate the capacity of the ensilage of Jatropha curcas L. expeller cake to reduce the phorbol esters and its effect on fermentative losses, by adding soluble carbohydrates or microbial inoculants. The design was completely randomized with four replications in a 3 × 2 factorial arrangement, with three sources of soluble carbohydrates (SC, control, 50 g sucrose/kg or 50 g crude glycerin/kg as fed) and two doses of microbial inoculants (MI, 0 or 5 × 10(5) ufc Lactobacillus plantarum + 3.33 × 10(5) ufc Propionibacterium per g as fed). Twenty-four mini-silos (982 cm³) of polyvinyl chloride were created and opened after 60 days of fermentation at room temperature. The pre-hydrated Jatropha curcas L. cake (282 g of water/kg) contained 0.424 mg of phorbol esters/g of dry matter. Ensiling reduced the phorbol esters in 47.4%, on average, regardless of the SC or MI. There was no interaction effect between SC and MI on effluent, gases or total dry matter losses. However, both losses were increased when SC were added, and it was higher with glycerin that than sucrose. The addition of MI reduced all fermentation losses. The process of ensiling, although partially to reduce the phorbol esters of pre-hydrated Jatropha curcas L. cake, is not indicated as a biodestoxification procedure.

Avaliaram-se diferentes modelos lineares (quadrático, linear-platô, quadrático-platô e raiz quadrada) e não lineares (exponencial-Mitscherlich e Michaelis-Menten) para descrever o crescimento microbiano ruminal em função do suprimento de nutrientes. Utilizou-se banco de dados compostos por 76 observações de meios de cultivo ruminal in vitro (mix de bactéria) contendo 0,0; 0,375; 0,75; 1,50; 3,0; 6,0; 12 e 24g de sacarose/L. O crescimento microbiano foi avaliado por meio de mensuração da densidade óptica (OD - 600nm) após 0; 6; 12; 18 e 24 horas de fermentação. A avaliação conjunta dos parâmetros da regressão linear dos valores preditos e observados indicou adequação de todos os modelos na predição do crescimento microbiano. A partição do quadrado médio do erro de predição não permitiu identificar diferenças na qualidade da predição do crescimento microbiano entre os modelos. Porém, ao avaliar o crescimento marginal em função do suprimento de glicose verificou-se inadequação do modelo linear-platô e do modelo quadrático (em baixos níveis substratos). Os modelos quadrático-platô, raiz quadrada, Mitscherlich ou Michaelis-Menten podem ser utilizados para descrever o comportamento do crescimento microbiano ruminal ao suprimento de nutrientes. Entretanto, diferenças são observadas quanto às estimativas das eficiências de uso nos diferentes níveis de substratos, o que implica em diferenças nas recomendações dos níveis ótimos de nutrientes. Neste sentido, a escolha desses modelos poderá ser modulada pelo nível de suprimento de nutrientes desejado.

Linear (linear-plateau, quadratic, quadraticplateau, square root) and non-linear (exponential and Michaelis-Menten) models to describe the ruminal microbial growth as a function of sucrose levels were evaluated. Is was used a database of 76 observations of microbial growth in roll tubes (mix of bacterial) containing final concentrations of 0.0; 0.375; 0.75; 1.50; 3.0; 6.0; 12 and 24g/L of sucrose. The microbial growth was evaluated by measuring optical density (OD - 600nm) at 0; 6; 12; 18 and 24 hours of fermentation. The evaluation of linear regression parameters of predicted and observed values indicated adequacy of all models in the prediction of microbial growth. The partition of the mean square prediction error did not allow identifying differences in relation to prediction quality of the models. However, the marginal growth as a function of sucrose supply was inadequate to models linear-plateau and quadratic (in low sucrose levels). The square root, Mitscherlich and Michaelis-Menten models can be used to describe biological responses to nutrients. However, differences are observed in estimates of substrates efficiencies, which implies differences in the recommendations of optimum levels of nutrients. In this sense, the choice of these models can be modulated by the level of nutrient supply desired.