Quantification of Mold in Hay by near Infrared Reflectance Spectroscopy

Abstract

Near infrared reflectance spectroscopy was used to quantify mold by spectrally predicting chitin, a fungal cell wall polymer of glucosamine. Chitin data and near infrared reflectance spectra were collected from moldy alfalfa (Medicago sativa L.) hay samples. Chitin concentrations ranged from 75 to 710 μg/g DM and were similar to concentrations in other contaminated products. Second derivative math transformations were made on the spectra, and a calibration equation was developed by regressing transformed spectra against chemical data. The coefficients of determination for calibration and validation were .90 and .82, respectively. Standard errors of calibration and validation were 44 and 65 μg/g DM, respectively. The calibration equation utilized four terms, represented at wavelengths 1630, 2114, 2246,and 2356 nm. The spectra of commercial glucosamine were plotted, and reflectance peaks occurred at 2114 and 2356 nm. The validated equation was used to predict mold (chitin) in bales that had been visually estimated for mold with a relative mold index. Relative mold index was regressed against chitin and was correlated to chitin (R=.86); residual error was small when samples were extremely moldy or clean. We concluded that near infrared reflectance spectroscopy can accurately quantify mold in alfalfa hay. © 1987, American Dairy Science Association. All rights reserved.