Abstract
Near-infrared reflectance spectroscopy (NIRS) was used to predict chemical compositions of forage paddy rice silage samples (n=116). Eighty-six samples were selected for calibration development, and remaining samples were used for equation validation. Accuracy of prediction was evaluated by correlation coefficients (r) and standard deviation of prediction (SDP). NIRS accurately predicted crude protein (CP), dry matter (DM), crude ash (CA), organic cellular contents (OCC), organic cell wall (OCW) and low digestible fiber (Ob) contents of samples, but prediction of high digestible fiber (Oa) was markedly less accurate. The r and SDP values for CP, DM, CA, OCW, OCC and Ob were respectively 0.97 and 0.43, 0.97 and 0.29, 0.98 and 0.73, 0.98 and 2.33, 0.98 and 2.82, and 0.99 and 1.86. The results indicate that CP, DM, CA, OCW, OCC and Ob contents in forage paddy rice silage can be predicted by using NIRS techniques with an acceptable accuracy.
