Skip to main content
ARS Home » Research » Publications at this Location » Publication #75364
Title: PROGRESS IN PREDICTING RICE MILLING YIELD USING NEAR-INFRARED TECHNOLOGY

Author
item McClung, Anna
item Delwiche, Stephen - Steve
item Webb, Bill

Submitted to: Rice Technical Working Group Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 2/1/1996
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: In rice breeding programs, it is necessary to mill thousands of samples annually to select for milling quality. This is usually done by using a labor intensive process which involves several time consuming steps. A 'quick and dirty' method which would classify breeding lines into low, intermediate and high milling yield categories would be of considerable value for use in varietal improvement programs. Previous reports have show that near-infrared technology offers promise in estimating whole grain milling yield in Australian medium-grain rice varieties. We initiated a study to see if this technique could be used as a tool for predicting milling quality of U.S. rice germplasm. Rough rice and brown (dehulled) rice samples were scanned using near-infrared transmission (NIT) and near-i ed reflectance (NIR) equipment. Prediction of whole milling yield by scanning rough rice with NIT was similar to that reported by Australian researchers indicating that the method was applicable to U.S. germplasm. Higher levels of precision for predicting total milling yield were observed using NIR scans of rough rice and were within about 1% accuracy of that determined by the milling process. The regression equation for predicting whole milling yield using rough rice was less accurate than for total milling yield. Scanning brown rice instead of rough rice somewhat improved the prediction of whole milling yield but also added the hulling step to the data collection process. As the database increases to reflect more environmental and genetic variation, it is expected that the statistics for prediction will improve. However, it appears that breeders could use the NIR method to eliminate lines with low whole milling yields and determine actual milling yield on just a portion of their breeding selections.