Analysis of Forage Research Samples Utilizing a Combination of Wet Chemistry and near Infrared Reflectance Spectroscopy

Publisher Website
Google Scholar
Abstract
The possibility of utilizing near infrared reflectance spectroscopy (NIRS) combined with wet chemistry analysis to determine chemical composition and in vitro digestion of forage samples from research experiments was investigated. For every experiment where NIRS was used in combination with wet chemistry, wet chemistry analyses were conducted, and NIRS equations were developed and validated utilizing a selected set of samples that represented all treatments evaluated in the experiment. The optimum NIRS equations developed were used to predict chemical composition or in vitro digestion of samples in the experiment that had not been selected for wet chemistry analysis. Chemical composition or in vitro digestion was determined in 15 forage experiments, consisting of tropical grasses and legumes, utilizing a combination of wet chemistry and NIRS analysis. Standard errors of calibration (SEC) and analysis (SEA) ranged from .14 to .79% and .32 to .83%, respectively, for NIRS analysis of crude protein (CP), and from 1.25 to 2.82% and 1.47 to 2.88%, respectively, for NIRS analysis of in vitro organic matter disappearance (IVOMD). Analysis of neutral detergent fiber (NDF) yielded SEC and SEA values ranging from .75 to 1.21% and 1.24 to 1.61%, respectively. Analysis of acid detergent fiber and acid detergent lignin gave SEC values of 1.10 and .61%, respectively, and SEA values of 1.42 and .94%, respectively. A subset of samples that had undergone wet chemistry analysis was selected from every experiment where a given quality trait was analyzed, and general NIRS equations were developed for CP, IVOMD and NDF. General NIRS equations were used to predict chemical composition or in vitro digestion of samples from the validation groups of the individual experiments. General CP and NDF equations yielded similar validation results compared with equations developed from each experiment, indicating that a general CP or NDF equation could be used to predict CP or NDF content of a specific group of samples. Validation results from the general IVOMD equation were not as acceptable as those from equations developed from each experiment, indicating that separate IVOMD NIRS equations should be developed for each specific group of samples.