Resumen Se realizó un experimento in vitro utilizando un sistema Rumen Simulation Technique (RUSITEC) con el objetivo de evaluar el efecto del fruto completo, del pericarpio y del extracto de saponinas semipurificadas de Sapindus saponaria sobre la fermentación ruminal y la liberación de metano. Como sustrato básico se utilizó una mezcla compuesta por 60% de Brachiaria dictyoneura cv. Llanero (una gramínea de baja calidad) y 40% de hojas de Cratylia argentea (una leguminosa de calidad media), y se observó el efecto de la inclusión de 8% de fruto, 5% de pericarpio ó 1.2% de extracto de saponinas semipurificadas de Sapindus saponaria (pureza aprox. 95%) en base seca. El suministro diario de materia seca se mantuvo constante para la evaluación de los 4 tratamientos durante 2 periodos de 10 días cada uno. Los datos se analizaron como un diseño de bloques completos al azar. No hubo efecto (p > 0.05) de los tratamientos sobre la concentración de amoniaco en el líquido ruminal. El fruto y el pericarpio redujeron el recuento de protozoos ciliados en 50% con respecto al control (p < 0.05), mientras el recuento de bacterias fue similar en todos los tratamientos (p > 0.05). No hubo cambios en la degradación de los nutrientes (p > 0.05), excepto de fibra en detergente neutro que fue 9.3% menor en el tratamiento con fruto (p < 0.05). No hubo efecto de los tratamientos sobre la liberación diaria de metano por fermentador (p > 0.05). Sin embargo, el fruto, el pericarpio y el extracto de saponinas redujeron la liberación de metano por gramo de materia orgánica degradada en 9.5, 5.6, y 4.5% respectivamente (p < 0.05). La reducción que se observó en la liberación de metano en relación con la cantidad degradada de MO podría traducirse en condiciones prácticas, en una reducción de metano emitido por unidad de proteína animal producida que sería útil aun cuando la cantidad total de metano emitido no disminuya. La reducción significativa de la población de protozoos ciliados confirma el potencial defaunante del fruto completo y del pericarpio de Sapindus saponaria, y aun cuando esto no esté asociado necesariamente con una disminución de la metanogénesis, podría ser beneficioso en otros aspectos, como la utilización del nitrógeno, en los cuales se han reportado efectos positivos de la defaunación completa o parcial. Palabras clave: Brachiaria dictyoneura, Cratylia argentea, defaunación, fibra, protozoos. Summary Ruminal fermentation pattern and methane release were determined in a RUSITEC (Rumen Simulation Technique) system using as control substrate 60% Brachiaria dictyoneura cv. Llanero hay (a low quality grass) and 40% Cratylia argentea leaves (a medium quality shrub legume). The effects of the addition of 8% fruit, 5% pericarp and 1.2% semipurified saponins extract of Sapindus saponaria in a dry matter basis to the system were tested. Daily dry matter supply was kept constant to evaluate the four treatments during two 10 days-term periods. Data were analyzed as a randomized completeblock design. There was no effect of treatments on ammonia concentration (p > 0.05). As compared to control, fruit and pericarp reduced ciliate protozoa counts by 50% (p < 0.05) but bacteria counts remained unchanged (p > 0.05). Neutral detergent fiber degradation of the fruit-added treatment was 9.3% lower than control (p < 0.05) but degradation of dry matter, organic matter, protein and acid detergent fiber were not different than control (p > 0.05). Daily methane release (mmol/d) was unaffected by treatments (p > 0.05). However, the addition of fruit and pericarp reduced methane release per gram of degraded dry matter by 8.7 and 4.3% respectively (p < 0.05). Methane release per gram of degraded organic matter decreased by 4.5, 9.5 y 5.6% with the addition of the semipurified saponin extract, the fruit and the pericarp respectively (p < 0.05). The decrease in methane emission per gram of DM and OM degraded, is most likely explained by lower fiber degradation and suggests that there will be less methane emission per unit of edible animal protein produced, eventhough the total amount of methane released is not reduced. The reduction of ciliate protozoa population confirms the defaunation potential of the whole fruit and pericarp of Sapindus saponaria, that may not reduce the production of methane but will definitely result in a more efficient utilization of nitrogen. Key words: Brachiaria dictyoneura, Cratylia argentea, defaunation, fiber, protozoa.

Six adult African-type hair sheep (BW = 40.3 +/- 6.3 kg) fitted with ruminal and duodenal cannulas were subjected to four treatments. Sheep were offered basal diets at a rate of 80 g of DM/kg of metabolic BW (equivalent to ad libitum access) consisting either of a low-quality grass hay (Brachiaria dictyoneura, 3.7% CP, DM basis) alone or in combination with a forage legume (Cratylia argentea, 18.6% CP, DM basis) in a 3:1 ratio (DM basis). In addition, 0 or 8 g of DM of Sapindus saponaria fruits (12.0% crude saponins, DM basis) per kilogram of metabolic BW was administered intraruminally. Supplementation of C. argentea increased intakes of OM (+21%; P < 0.01) and CP (+130%; P < 0.001), as well as ruminal fluid ammonia N concentrations (from 2.40 to 8.43 mg/dL; P < 0.001). Apparent OM and N digestibilities were not affected by legume addition, but ADF digestibility decreased by 10% (P < 0.01). Total ruminal VFA concentration was unchanged, but acetate:propionate was lower (P < 0.01) and isobutyrate proportion was greater (P < 0.001) with the legume addition. Legume supplementation increased duodenal flows of total N (+56%; P < 0.001), nonammonia N (+52%; P < 0.001), ruminal escape N (+80%; P < 0.001), and microbial N (+28%; P < 0.05). Microbial efficiency was not affected by legume addition. Supplementation of S. saponaria increased (P < 0.05) dietary OM intake by 14%, but had no effect on CP intake and ruminal fluid ammonia concentration or on OM and N digestion. Digestibility of ADF was decreased (P < 0.01) by 10% with S. saponaria as was acetate:propionate (P < 0.001) and the isobutyrate proportion (P < 0.001). Ruminal protozoa counts increased (P < 0.01) by 67% with S. saponaria. Duodenal N flows were not significantly affected by S. saponaria supplementation, except for microbial N flow (+34%; P < 0.01). Microbial efficiency was greater (P < 0.05) by 63% with the addition of S. saponaria. Few interactions between legume and S. saponaria supplementation were observed. The NDF digestibility was decreased with S. saponaria in the grass-alone diet, but not in the legume-supplemented diet (interaction; P < 0.05). Interactions were absent in ruminal fermentation measures and duodenal N flow, indicating that effects were additive. Results suggest that, even when not decreasing ruminal protozoa count, supplementation of S. saponaria fruits is a beneficial way to improve ruminal VFA profile, microbial efficiency, and duodenal flow of microbial protein in sheep fed tropical grass-alone or grass-legume diets.

A comparison of the effects on rumen fermentation of three saponin rich tropical fruits supplemented to forage-based diets was completed using a rumen simulation technique (Rusitec). The diets contained either no tropical fruit or 100mg/g of Sapindus saponaria (crude saponins, 120mg/g), 200mg/g of Enterolobium cyclocarpum (crude saponins, 19mg/g) or 200mg/g of Pithecellobium saman (crude saponins, 17mg/g). The four diets were evaluated with faunated and defaunated rumen fluid obtained from a single donor cow. Compared to the control diet, P. saman decreased (P<0.05) ammonia concentration of rumen fluid, and E. cyclocarpum and P. saman increased (P<0.05) n-butyrate proportion of total volatile fatty acids. Defaunation enhanced (P<0.05) propionate proportion with corresponding reductions of acetate and n-butyrate. Organic matter degradation of the S. saponaria diet did not differ from that of the control diet but was higher (P<0.05) with P. saman and E. cyclocarpum. Only one of the saponin rich fruits evaluated, S. saponaria, decreased (P<0.05) protozoal count (by 54%) and daily methane release (by 20%) relative to control, but without affecting the methanogen count. Defaunation suppressed methanogenesis by 43% over all diets (P<0.05), and the effect of S. saponaria on methane was more pronounced in defaunated (29%) versus faunated rumen fluid (14%). When related to organic matter apparently fermented, differences relative to the control diet persisted (P<0.05), but methane release per unit of fibre degraded did not differ between the S. saponaria diet and the control diet. This study demonstrated that supplementation with S. saponaria is effective against ruminal methanogenesis, but that this was not exclusively an effect of the associated depression in protozoal count.

In two in vitro experiments with the RUSITEC-apparatus, Brachiaria dictyoneura was tested alone and with legumes at dietary proportions of 1/3, 2/3, or 3/3 of Arachis pintoi (Expt 1) and 1/3 of Arachis pintoi, Cratylia argentea, or Calliandra calothyrsus (Expt 2). In Expt 2, all diets were evaluated with and without 80 mg/g diet of Sapindus saponaria fruits. In Expt 1, the stepwise replacement of the grass by A. pintoi curvi-linearly increased rumen fluid concentrations of ammonia, volatile fatty acids, bacteria, and protozoa. Methane release rates were 1.7, 7.3, 8.8, and 9.0 mmol/day. With increasing legume proportion, more organic matter and protein were degraded, the latter being only partially recovered as ammonia. In Expt 2, 1/3 of A. pintoi basically had the same effects as in Expt 1. Cratylia argentea was less effective in modifying the fermentation pattern. In association with a higher nutrient degradation and rumen ammonia concentration, C. argentea and A. pintoi increased methane release to about 3- and 4-fold levels. Calliandra calothyrsus reduced nutrient degradation and methane release per gram of organic matter degraded. Tannins, predominant in C. calothyrsus, might have affected methanogenesis. Sapindus saponaria reduced methanogenesis by 11% on average in grass-alone and legume-supplemented diets.