Variability Among Peanut Subsamples Prepared for Aflatoxin Analysis with Four Mills
Abstract The variability in aflatoxin concentration among peanut subsamples ground with 4 different mills was evaluated. Twenty 2 kg samples of naturally contaminated peanuts were ground in a Dickens subsampling mill (DM), a Stephan model UM-12 vertical cutter mixer (SM), and a Robot Coupe model RSI6Y-1 vertical cutter mixer (RC1). Twenty 4 kg samples were ground in the DM, SM, and a Robot Coupe model R10P vertical cutter mixer (RC2). From each 2 kg sample, ten 100 g subsamples were withdrawn, and from each 4 kg sample, ten 200 g subsamples were withdrawn. Subsamples were analyzed for aflatoxin by liquid chromatography. The coefficient of variation (CV) among each set of 10 subsamples was determined for each sample, and the CVs for each sample size were ranked and analyzed by the Kruskal-Wallis test of ranks. For 2 kg samples, the CVs for the samples ground in RC1 ranked significantly lower than those for samples ground in DM and SM. For 4 kg samples, the CVs for samples ground in RC2 and SM were significantly lower than that for samples ground in DM. The averages of the CVs for 2 kg samples were 17.2% (RC1), 32.8% (SM), and 40.6% (DM). The averages of the CVs for 4 kg samples were 21.2% (RC2), 26.0% (SM), and 47.0% (DM).