Abstract: 【Objective】This study was conducted to investigate the feasibility of measuring in vitro dry matter digestion (IVDMD), neutral detergent fiber (NDF), acid detergent fiber (ADF), crude protein (CP) and ether extract (EE) concentrations in ensiled maize by near infrared diffuse reflectance spectroscopy (NIDRS). 【Method】The ensiled stover and whole plant samples from normal oil maize, high oil maize and super oil maize were used to establish NIDRS calibration models of ensiled maize with the techniques based on the principal component space of spectra and partial least square regression (PLS). 【Result】 The results showed that the determination coefficients of cross validation (R2cv) and validation (R2val , in parentheses) of these yielded models were 0.9133 (0.8879), 0.9764 (0.9455), 0.9789 (09635), 0.7294 (0.7333) and 0.9254 (0.9387) for IVDMD, NDF, ADF, EE and CP respectively. The root mean square error of estimation, cross validation and prediction (RMSEE, RMSECV and RMSEP) ranged from 0.26 for CP to 2.23 for NDF. 【Conclusion】 It has been demonstrated that it is feasible to use NIDRS as a rapid and an accurate technique to predict ensiled maize quality traits. These calibration models could meet the needs of silage quality evaluation in livestock industry and screening of various samples in silage maize breeding.

Key words: Maize, Ensilage, Quality traits, Near infrared reflectance spectroscopy (NIRS), Calibration models