Determination of cell wall carbohydrates in forages by near infrared reflectance spectroscopy

Abstract The use of conventional techniques to analyze neutral sugar constituents of cell wall polysac-charides in forages has been time consuming and not suited for routine analysis. The purpose of this study was to determine the ability of near infrared reflectance spectroscopy (NIRS) to analyze pectin, arabinose, xylose and glucose from forage cell walls. The concentration of pectin, arabinose, xylose and glucose was determined by conventional fractionation and quantification methods on whole-plant tissue from six cool season legumes, two cool season grasses and four warm season grasses commonly grown through the southeastern U.S.A. Forages were harvested at vegetative, first bloom or mature stages of growth. Calibration equations were obtained by multiple linear regression of conventional laboratory analysis values on NIR spectra from 72 randomly chosen spectra using a scanning monochromator NIR spectro-computer system. Calibration equations were verified with 24 additional samples from the original population. From four to eight wavelengths were needed to develop the best equation for the carbohydrate fractions. The coefficients of multiple determination ranged from 0.86 to 0.92. The standard errors of calibration were 0.98, 0.83, 2.48 and 2.21 for pectin, arabinose, xylose and glucose, respectively. Analysis of residuals indicated that error from NIRS prediction was associated with the experimental design. Forage × maturity interactions were significant for arabinose, xylose and glucose residuals. The precision of NIRS analysis was moderately high, but not as precise as previous reports using a single forage species. Better accuracy will probably be achieved by developing NIRS calibration equations for individual species or groups of similar forage species.