The aim of this research was to evaluate the efficiency of two sampling plans for fumonisins (B1+B2) analysis in 11 lots of maize. For sample collection the two systems used were a manual system in whole grains, using sampling spear, and an automatic system in milled grains, using the continuous flow sampling system (CFSS). For comparison purpose, two dispersion measurements were used: variance and coefficient of variation. For each sampling plan the variance associated to the three steps of fumonisins quantification (sampling, sample preparation, and analysis), and the total coefficient of variation were determined. The correlation between the average fumonisins concentration and the variances for each phase of the detection procedure was evaluated for each sampling plan using analysis of regression. At the automatic sampling plan in milled grains for fumonisins analysis in maize both sampling variance (0.0226mg kg-1)² and total coefficient of variation (6.37%) were lower (P<0.01) than in the manual sampling plan in whole grains whit (0.0685mg kg-1)² of variance and 8.94% of coefficient of variation, showing so better efficiency for fumonisin detection on maize.
mycotoxin; contamination; homogeneity; variability; variance; granulometry
