Extension Forage Testing Using near Infrared Reflectance Spectroscopy

Near infrared reflectance spectroscopy (NIRS) was evaluated for its accuracy in analyzing the nutritive value of tropical grass hays made by Florida livestock producers. From 1982 to 1987, a unique set of producer hays was obtained representing four tropical grass species, and an unknown range of maturities, fertilization schemes, weather conditions, and hay making and storage procedures. General NIRS equations representing the above sources of variation were developed for crude protein (CP), invitro OM digestion (IVOMD) and neutral detergent fiber (NDF). Results from general equations were compared to results from species-specific NIRS equations to determine if broadbased calibrations could analyze quality traits of specific sample groups with similar accuracy as calibrations developed from specific groups. Standard errors of calibration and analysis ranged from .73 to .96% and .74 to .92%, respectively, for CP, 2.30 to 3.14% and 1.87 to 4.17%, respectively, for IVOMD, and 1.45 to 1.71% and 1.45 to 2.18%, respectively, for NDF. Numbers of unacceptable analyses obtained from general NIRS calibrations were no greater than those obtained from species-specific calibrations. Other validation statistics from general NIRS calibrations were intermediate to those from species-specific calibrations, indicating that broad-based calibrations analyzed the nutritive value of individual species with a degree of accuracy similar to that of speciesspecific calibrations. Broad-based NIRS calibrations have application in extension forage testing programs where identification and purity of producer hays may not be known.