Application of Near Infrared Reflectance Spectroscopy to Analyses of Poultry Feeds

Abstract

The use of near infrared reflectance spectroscopy (NIR) to determine chemical composition was investigated in poultry feeds. Calibrations for crude protein (CP), calcium (Ca) and phosphorus (P) were developed. A scanning type instrument, the InfraAlyzer 500 was used to select the best set of wavelengths that were used in a filter type instrument, the InfraAlyzer 400+. Feed samples were obtained from 11 laying hen operations in Southern Ontario. The accuracy of the predictions were assessed by the coefficients of determination (r2), standard error of the estimate (SEE), and by the bias or mean difference between chemical and NIR-predicted values. The r2 and SEE values for CP, Ca and P were .90, .94; .80, .58; and .72, .07 for the calibration (C) set samples and .88, .82; .84, .61; and .60, .08 for the prediction (PRE) set samples, respectively. The results of the study suggested that wet chemistry may have influenced the predictions of chemical composition of poultry feeds by NIR.The results of the statistical analyses for CP, Ca, and P indicated that, provided precautions are taken with sample preparation and wet chemistry analyses, NIR may be used as a tool for monitoring quality control in poultry feeds.