NEAR-INFRARED SPECTROSCOPIC ANALYSIS OF UNDRIED SILAGES AS INFLUENCED BY SAMPLE GRIND, PRESENTATION METHOD, AND SPECTRAL REGION

The objectives of this work were to evaluate how different grinding procedures, spectral regions (680 to 1234 nm, 1100 to 2498 nm, and 680 to 2498 nm), and sample presentation methods influence near infrared reflectance analysis of undried silages. Grinds included 1) undried with dry ice in a Vita-Mix 3600 Plus blender, 2) undried in a Hobart chopper, 3) undried in a Wiley grinder, 4) in a Wiley grinder (20 mesh) after oven drying 60-degrees-C, and 5) intact. Presentation methods included 1) rectangular cell, 15-cm path, with and without polyethylene bags, and 2) a nonrotating circular cell. Undried (including dry ice regrinds of Hobart-grind and Wiley-grind) and dried silages were analyzed for DM, CP, and ADF. Each calibration was performed three times (average used) with each third of the samples used once as the validation set. Results, for undried samples, for DM and CP were about the same for all methods of grinding, presentation, and source of chemistry. Overall, the undried Wiley ground samples in the rectangular cell gave better results than any other grind-cell combination among the Wiley, Hobart, and intact undried materials. Intact samples generally did not result in satisfactory estimates of composition, even when multiple samples were scanned and averaged, nor did wavelengths below 1100 nm.