Within and between field variability of residue data and sampling implications.

Information on the variability of residues in individual fruits and vegetables is required to estimate the acute dietary exposure to pesticides. The distribution of residues in apples, kiwi fruits, potatoes and butter beans was studied in field experiments representing commercial farming practice. No correlation was found between the residue concentration or surface residue and the mass of apples. The relative frequency distributions of residues in crop units were continuous and skew positive. The log-normal transformation did not result in a normal distribution in any of the trials. Consequently, 299, 120 and 59 random samples should be analysed to estimate 99th, 97.5th and 95th percentile of the residues at 95% confidence level. The distribution of residues is not significantly influenced by the mean residue, pre-harvest interval, chemical and physical properties of the active ingredient, formulation of pesticide or, on top fruit, the foliar application method. However, the residue distribution is likely to be influenced by the size, shape and density of the plants, and mode of application of pesticides. The variability factor should be defined as the ratio of the highest value at a specified percentile of residues occurring in unit crops and the population mean. Generic variability factors may be determined for various groups of commodities. Variability factors of 5 and 9 are recommended for medium size fruits, and potato following granular application of pesticides, respectively.