Skip to main content
ARS Home » Plains Area » College Station, Texas » Southern Plains Agricultural Research Center » Crop Germplasm Research » Research » Publications at this Location » Publication #86097
Title: EVALUATION OF NEAR INFRARED REFLECTANCE FOR OIL CONTENT OF COTTONSEED

Author
item Kohel, Russell

Submitted to: Journal of Cotton Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/6/1998
Publication Date: N/A
Citation: N/A

Interpretive Summary: Cotton improvement is based on the production and sale of fiber, but cottonseed products are an important secondary product. Cottonseed is a major oilseed in domestic and international markets. At present there are not ready methods that can be used in cotton breeding and testing programs to routinely evaluate oil content of cottonseed. We are reporting the development and evaluation of methods calibrating near infrared reflectanc spectroscopy instrumentation for the measurement of oil content in cottonseed. Six cotton lines were grown at two locations for three years to generate prediction equations for oil content of cottonseed. The equations were generated with near infrared reflectance spectroscopy based on seed-oil contents from continuous wave nuclear magnetic resonance analysis. The prediction equations were evaluated in the test material and the National Cotton Variety Test entries. The prediction equations were capable of identifying differences in seed-oil content within our test populations, but a universal prediction of seed-oil content will require a larger sample base to eliminate spectral variability that interferes with calibration.

Technical Abstract: A data set of 36 samples consisting of a single replication of six entries grown at two locations for three years was used to generate prediction equations for oil content of cottonseed, Gossypium hirsutum L. The equations were generated with near infrared reflectance spectroscopy (NIRS) based on seed-oil contents from continuous wave nuclear magnetic resonance (NMR) analysis. The results in predicting the remaining four replications in the tests, 144 samples, demonstrate the ability to develop prediction equations capable of identifying differences in seed-oil content within specified populations (r2 = 0.89 to 0.92), but they were not successful when tested on the National Cotton Variety Test entries (r2 = 0.00 to 0.84). A universal prediction of seed-oil content will require a larger sample base to eliminate spectral variability that might interfere with calibration.