Abstract
Near infrared reflectance (NIR) technique was used to determine moisture, protein and ash contents in 60% extraction flour samples from home-grown wheat harvested over 3 years (1979-1981). In order to ascertain a feasibility of an universal calibration set, a performance of the calibration developed using the 37 flour samples from the 1980 wheat was evaluated. It is concluded that standard error of difference between standard-analytical and NIR predicted values are in the range of acceptable limits except ash measurement, which means that the NIR technique is applicable to a routine analysis, however, a further study should be conducted to solve the problems with ash measurement and an annual bias adjustment in the instrument.
