APPARENT AMYLOSE CONTENT OF MILLED RICE BY NEAR-INFRARED REFLECTANCE SPECTROPHOTOMETRY

Amylose content is considered to be the most important factor influenc ing cooking and processing characteristics of rice. Traditional method s for measurement entail titrimetric, amperometric, or colorimetric me asurement of an iodine-binding complex of solubilized starch, but thes e methods are time-consuming and prone to random error. A method for d etermining apparent amylose content (AAC), based on the near-infrared (NIR) reflectance spectrum (1,100-2,498 nm) of ground milled rice, was developed on a diverse population of 50 commercial rice varieties or experimental lines grown in the five major rice-producing states of th e United States. Three laboratories possessing NIR spectrophotometers participated in a cooperative study. Each laboratory ground, scanned, and developed AAC models by partial least squares analysis, using comm on sets of samples for calibration (n = 150) and validation (n = 97). Validation set statistics typically indicated r(2) = 0.95 and standard error = 1.0% AAC for models consisting of 16-18 factors. Within-labor atory repeatability error of the NIR method was comparable to the refe rence colorimetric method. Between-laboratory reproducibility error (0 .5% AAC) indicated that the NIR models were consistent across laborato ries. Such accuracies are considered sufficient for selection in rice breeding programs.