Soja Sekharan, Kevin D. Kephart and E. Brent Turnipseed

Plant Science Department

South Dakota State University, Brookings, SD

Near infrared reflectance spectroscopy (NIRS) can be used as a nondestructive technique to analyze seeds and grains, making it a potentially important tool for several applications. Possible seed-related applications for NIRS may include analysis of seed quality traits to reduce time for routine laboratory testing and help in nondestructive analysis of germplasm and breeding lines. Additionally, storage and archiving laboratories could use NIRS to nondestructively monitor their collections. The objectives of this research are to develop and evaluate NIRS calibrations for seed size, seed vigor, germination, and hard seed content. Results from this research may be advantageous to breeders and to seed laboratories.

Alfalfa and soybeans were chosen for this research project. For soybean, the experiment had 15 cultivars with three levels of accelerated aging at 0h, 48h, 72h and 96h in four replicates. For alfalfa, 15 cultivars were also used with three levels of accelerated aging at 0h, 6h, 12h and 24h in four replicates. Near infrared reflectance spectra were obtained using an NIRSystems model 5000 controlled by software from Infrasoft International, LLC (ISI). Reference procedures included germination percentage at 72 and 168 h, hard and dead seed (AOSA, 1988) and vigor testing with a standardized technique for accelerated aging (AOSA,1983). Calibration equations were developed using modified partial least square procedures of ISI software.

Soybean NIRS calibrations provided encouraging results. Calibration and cross-validation results for seed weight, the two germination periods, and percentage dead seed were excellent. The R² values were in the range of 0.92 to 0.94. Standard errors of calibration were 3, 15, 15, and 35% of the sample set mean for seed weight, 72-h germination, 168-h germination, and percentage dead seed, respectively.

Results for alfalfa were also encouraging (Table 1). The R² values were in the range of 0.49 to 0.93. The standard errors of calibration (SEC) were relatively low for most of the measurements and were near expected variability from seed analysis laboratories. These results suggest that NIRS may be a useful analytical tool for seed analysis.

Table 1. NIRS calibration statistics for seed quality traits of nonground alfalfa seed.

^a= standard error of calibration

^b= standard error of cross validation

^c= 1- variance ratio

References Cited: