A. Fanchone, H. Archimède and M. Boval

Organic matter digestibility (OMD), an essential criterion forthe evaluation of the nutrition of ruminants, cannot be measuredeasily at pasture. Therefore, the objective of this study wasto test and compare 2 methods of OMD prediction based on thefecal CP content (CPf) or near infrared reflectance spectroscopy(NIRS) applied to feces. First, published equations derivedfrom fecal N (Eq. 1_{CP}, n = 40) and from fecal NIRS (Eq. 1_{NIRS},n = 84) were used to predict OMD of an independent validationdata set from which in vivo OMD, ranging from 58 to 74%, wasmeasured for 4 regrowth stages of *Digitaria decumbens*. Second,to establish equations usable in grazing situations and to improvethe efficiency of the predictions, new equations were calculatedfrom a large data set (n = 174) using CPf (Eq. 2_{CP}) or fecalNIRS (Eq. 2_{NIRS}). By applying the CPf method, Eq. 2_{CPf} (OMD,% = 88.4 – 263.9/CPf, % of OM; residual SD = 2.92, r^{2 }= 0.63) showed similar statistical parameters (*P* < 0.01)when compared with Eq. 1_{CP} (OMD, % = 86.6 – 266.2/CPf,% of OM; residual SD = 2.95, r^{2} = 0.79). When using fecal NIRS,Eq. 2_{NIRS} showed decreased SE of calibration (SEC = 1.48) andof cross-validation (SECV = 1.75) and greater coefficient ofdetermination of cross-validation (R^{2}_{CV} = 0.85) than the previouslypublished Eq. 1_{NIRS} (SEC = 1.78, SECV = 2.02, R^{2}_{CV} = 0.77).The validation of the 4 equations on the validation data setwas satisfactory overall with an average difference betweenthe predicted and the observed OMD ranging from 0.98 to 2.79percentage units. The Eq. 2_{NIRS} was nevertheless the most precisewith a decreased residual SD of 2.53 and also the most accurate,because the SD of the average difference between predicted andobserved OMD was the lowest. Therefore, fecal NIRS providedthe most reliable estimates of OMD and is thus a useful toolto predict OMD at pasture. However, an adequate number of referencedata are required to establish good calibration. Indeed, bettercalibration statistics were obtained by increasing the dataset from 84 (Eq. 1_{NIRS}) to 174 (Eq. 2_{NIRS}). In contrast, usingfecal N on a set of 84 or 174 points did not improve the prediction.Both methods are useful for predicting OMD at pasture in certaincircumstances, using fecal NIRS when a large data set (n = 84and n = 174) is available and fecal CP with smaller data sets(n = 40).

digestibility, fecal crude protein, near infrared reflectance spectroscopy, sheep, tropical grass