Factors Affecting Analysis of the Chemical Composition of Tall Fescue with near Infrared Reflectance Spectroscopy

Open Access
Publisher Website
Google Scholar
Abstract
Tall fescue (Festuca arundinacea Shreb.) samples were collected for 3 consecutive yr and were analyzed chemically and with near infrared reflectance spectroscopy. Concentrations of various components, as analyzed by conventional chemistry and by spectroscopy, differed significantly between oven-dried and freeze-dried replicate samples. The ranges of squared correlation coefficients and standard errors of analysis, respectively, for the various components when near infrared reflectance spectroscopy and chemical procedures were compared for the six sample sets were: CP, .96 to .98 and .34 to .45; ADF, .88 to .92 and .63 to .98; NDF, .87 to .97 and .73 to 1.52; and in vitro digestible DM .91 to .98 and .90 to 2.35. Grinding of tall fescue samples through a cyclone mill following grinding through a Christy-Norris mill did not improve analysis accuracy with near infrared reflectance spectroscopy. When analysis equations prepared from samples collected during a calendar year were used to analyze samples collected during another year, correlation coefficients were generally lower and standard errors of analysis higher than when prediction equations were used within the same year''s samples. This confirms that there is an advantage in recalibrating equations to accommodate new environments of plant growth associated with years.