Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School

Plant Science


Plant breeding programs evaluate large numbers of breeding lines for quality characteristics. Traditional methods of determining quality have two major disadvantages. First, most traditional methods require the destructive testing of seeds, therefore limiting the testing to later generations in a breeding program. Secondly, traditional methods require a great deal of time. This experiment was designed to use Near-Infrared Reflectance Spectroscopy (NIRS) for rapid, non-destructive analysis of whole wheat seed. Comparisons were made between the results of NIRS and traditional quality analysis characteristics such as test weight, protein, hardness, ash, mix absorption and loaf volume. Material analyzed included advanced lines from the Uniform Regional Hard Red Spring Wheat Nursery (URN) and from the South Dakota State University Spring Wheat Breeding Program (SD). The URN was classified as a wide and diverse genetic and environmental base while SD was classified as a relatively narrow genetic and environmental base. The URN was used as a validation set to predict its own material and also the SD material. The SD material was also used as a validation set to predict its own material and the URN material. After analysis by NIRS, all sample data were used to determine a set of calibration equations. These equations were then used to identify a validation set of data points for comparison. The data from this validation set was paired with corresponding traditional quality data and analyzed using the least squares regression method. Final data has shown that wheat protein can be predicted accurately by the NIRS system. Other quality factors that showed promise for accurate predictions and could be useful in a breeding program are test weight, flour protein and possibly wheat ash. In this study, we found that using a very narrow environment/genetic base to predict a very narrow environment/genetic base had the better R2 values and proved to be the best prediction set. We also found out that a wide environment/genetic base predicting a narrow or a wide environment/genetic base produced relatively lower values when predicting the quality factors.

Library of Congress Subject Headings

Hard red spring wheat -- Quality Hard red spring wheat -- Seeds -- Analysis Near infrared reflectance spectroscopy



Number of Pages



South Dakota State University