A STUDY OF VARIABILITY ASSOCIATED WITH SAMPLING PEANUTS FOR AFLATOXIN

One lot of peanuts known to contain aflatoxin was extensively sampled to study the sources of variability. A nested design was used where sections (50 bag units), subsamples and analytical variation were the variables studied. Sample size was the most critical factor in characterizing this lot. Variability from section to section was not significant indicating random distribution of the contaminant. Three 20 lb samples were taken on a number of lots from the 1968 crop, each of which was subdivided into two equivalent subsamples. The aflatoxin was determined on each of these subsamples. The results indicated that all the significant variation came from the subsamples, further enforcing the thesis that sample size is the critical factor in variability, and not lot inhomogeneity. Analysis of 550 lots from the 1967 crop where triples and analysis were available indicates that the magnitude of the variability is a little greater than was found on the experimental lot. Using the pooled standard deviation of the 1967 crop data, operating characteristic curves were plotted to demonstrate the improvement that can be expected by increasing the sample size. © 1970 AOCS Press.