Predicting Corn Wet-Milling Characteristics with Near Infrared Reflectance Spectroscopy

Proceedings of the Corn Utilization Conference IV.
National Corn Growers’ Association. St. Louis, MO.
Brumm, T.J., L.A. Johnson, and C.R. Hurburgh.


The yields and qualities of products from the wet milling of corn vary considerably depending on grain properties, such as kernel hardness, density, starch content, and protein content (Fox et al 1991). A rapid test for predicting wet-milling characteristics would be a valuable aid in identifying lots and developing corn hybrids with greater value to the wet-milling industry. The objective of this research was to predict wet-milling characteristics using near infrared reflectance (NIR) spectroscopy.



Twenty-seven hybrids, encompassing a wide range of physical properties and compositions, were wet-milled using a laboratory procedure. They were analyzed using an NIR (1100nm to 2500 nm) scanning monochrometer, which was calibrated to relate the absorbance of light at selected wavelengths to starch yield and the protein content of recovered starch. In a previous study (Fox et al 1991) using the same samples, physical properties and proximate analyses were used to predict these two characteristics. The NIR spectroscopy calibrations (table below) gave correlations approximately equal to or better than those of the previous study (Fox et al 1991). Although these results indicate that it may not be possible to use NIR spectroscopy for sorting commercial lots or hybrid seed by their relative wet-milling potential. We can determine whether a particular hybrid is high, medium, or low in wet-milling potential.



Independent Variable

Wavelengths (nm)


Fox et al R-squared


Starch yield

1584, 1888, 1976




Protein content in starch

1448, 1696, 1728





1Standard error of calibration.


Fox, S.R., L.A. Johnson, C.R. Hurburgh, and C. Dorsey-Redding. 1992. Cereal Chem. 69(2): 191-197.