EVALUATION OF NEAR INFRARED REFLECTANCE SPECTROSCOPY FOR PREDICTING FORAGE QUALITY IN TIMOTHY

Abstract

Near infrared reflectance spectroscopy (NIRS) is a newly attractive technology introduced for the analysis of agricultural products and for which new instruments have been developed. The objective of this experiment was to evaluate the capabilities of the Technicon InfraAlyzer 500 (I/A-500), a scanner monochromator type instrument, with regard to its potential utilization in the breeding of timothy (Phleum pratense L.) for higher nutritional qualities. Special features of this instrument were also used to further assess its capacities once the wavelengths were deliberately reduced and set to make the I/A-500 comparable to a 19 wavelength filter instrument such as typically found in a Technicon I/A-400R model. The investigation was performed using a total of 120 timothy samples collected from spaced plant nurseries in 1982 and 1984 and analyzed for acid detergent fiber (ADF), neutral detergent fiber (NDF), crude protein (CP), digestibility (DMM), water solubility (WS), water retention (WR) and packed volume (PV) in the laboratory. The equations developed with the I/A-500 had R2 and r2 larger than 0.85 for ADF, NDF, CP, DDM and WS in both 1982, and 1984, and combined 1982–1984 with the exception of DDM and WS in 1982 that had r2 of 0.76. The lower R2 and r2 obtained for WR and PV were attributed to poor laboratory procedures. The standard errors of calibration of ADF, NDF, CP and DDM were all as good or better than those previously reported, with other cool season grasses. The equations developed by using only the 19 wavelengths typically found in an I/A-400R provided R2 standard errors of calibration, r2 and standard error of analysis which were quite similar to those obtained with the I/A-500. Thus, we concluded that both the unrestricted wavelength selection of the I/A-500 and the restricted wavelength selection available in an I/A-400R would be adequate to evaluate forage quality in timothy. In this experiment, the main limitations appeared to have been related to the precision of the laboratory procedures and to the lack of variation in the populations under evaluation. As a general guideline to evaluate these two factors, it is proposed to use the ratio of the standard deviation of the population over the standard deviation of the laboratory procedure. In our experimentation a ratio larger than 5.0 appeared suitable to obtain adequate calibrations.Key words: Fiber, crude protein, digestibility, forage physical properties, Phleum pratense L.