Sampling, sample preparation, and analytical variability associated with testing wheat for deoxynivalenol

The variability associated with testing wheat for deoxynivalenol (DON) was measured using a 0.454 kg sample, Romer mill, 25 g comminuted subsample, an d the Romer Fluoroquant analytical method. The total variability was partit ioned into sampling, sample preparation, and analytical variability compone nts. Each variance component was a function of the DON concentration and eq uations were developed to predict each variance component using regression techniques. The effect of sample size, subsample size, and number of aliquo ts on reducing the variability of the DON test procedure was also determine d. For the test procedure, the coefficient of variation (CV) associated wit h testing wheat at 5 ppm was 13.4%, The CVs associated with sampling, sampl e preparation, and analysis were 6.3, 10.0, and 5.3%, respectively. For the sample variation, a 0.454 kg sample was used; for the sample preparation v ariation, a Romer mill and a 25 g subsample were used; for the analytical v ariation, the Romer Fluoroquant method was used. The CVs associated with te sting wheat are relatively small compared to the CV associated with testing other commodities for other mycotoxins, such as aflatoxin in peanuts. Even when the small sample size of 0.454 kg was used, the sampling variation wa s not the largest source of error as found in other mycotoxin test procedur es.