Use of ten gram samples of corn for determination of mycotoxins

Data were gathered, during a study on the development of an automated system for the extraction, cleanup, and quantitation of mycotoxins in corn, to determine if it was scientifically sound to reduce the analytical sample size. Five, 10, and 25 g test portions were analyzed and statistically compared with 50 g test portions of the same composites for aflatoxin concentration variance. Statistical tests used to determine whether the 10 and 50 g sample sizes differed significantly showed a satisfactory observed variance ratio (Fobs) of 2.03 for computations of pooled standard deviations; paired t-test values of 0.952, 1.43, and 0.224 were computed for each of the 3 study samples. The results meet acceptable limits, since each sample's t-test result is less than the published value of the /t/, which is 1.6909 for the test conditions. The null hypothesis is retained since the sample sizes do not give significantly different values for the mean analyte concentration. The percent coefficients of variation (CVs) for all samples tested were within the expected range. In addition, the variance due to sample mixing was evaluated using radioisotope-labeled materials, yielding an acceptable CV of 22.2%. The variance due to the assay procedure was also evaluated and showed an aflatoxin B, recovery of 78.9% and a CV of 11.4%. Results support the original premise that a sufficiently ground and blended sample would produce an analyte variance for a 10 g sample that was statistically comparable with that for a 50 g sample.