Soja Sekharan

Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School



Seed samples of soybean [Glycine max L. (Merr.)] and alfalfa [Medicago sativa (L.)] were scanned with a scanning monochromator near infrared reflectance spectrometer (NIRS). The objective was to determine if NIRS couldbe used for rapid, non-destructive seed analysis and for the evaluation of NIRS calibrations for seed weight, seed vigor, and germination in soybean and alfalfa. Three independent sets of samples, Sets 1, 2, and 3 in soybean and alfalfa were involved in germination testing (AOSA, 1988) and accelerated aging (AOSA, 1983). Common measurements for soybean and alfalfa included seed, percent dead seed and germination count at 168 h. The percentages of abnormal seed and germination at 72 h was measured for soybean, whereas percentages of hard seed, germination at 96 h and viable seed were measured for alfalfa. Measurements from Sets 1 and 2 in soybean and alfalfa were used for creating calibration equations for seed weight, germination and seed vigor. Set 3 was used for external validation. In soybean, equations from Set 1 recorded low R2 as compared to Set 2 which recorded R2 in the range o( 0.92-0,-94 for seed weight and all germination measurements. The validation statistics in samples of Set 3 indicated that equations from Set 1 predicted satisfactorily. In alfalfa, equations from Sets 1 and 2 had R2 ranging from 0.79-0.93 and 0.85-0.95, respectively, for seed weight and all germination measurements. Validation statistics in samples of Set 3 for hard seed content, germination at 168 hand viable seed percentage were excellent with calibration equations from Set 1. The results indicate the potential for NIRS to be used for routine seed analysis.

Library of Congress Subject Headings

Soybean -- Seeds -- Analysis
Alfalfa -- Seeds -- Analysis
Near infrared reflectance spectroscopy



Number of Pages



South Dakota State University