Data generated in the UK have indicated that pesticide residue levels can be highly variable between the individual fruit or vegetables from the same crop or lot in trade. Statistical experiments with these data showed that residues in composite samples, taken according to Codex recommendations, are unlikely to differ by more than a factor of 3–4 from the mean level in the lot. This was corroborated by results obtained from real composite samples. Many fruit or vegetables in trade are mixed after harvest to form combined lots. Analysis of composite samples provides a good indication of average residues but, where the lot has been mixed, such average values are potentially misleading. Residues monitoring is the best means available for general control of pesticide use and consumer exposure, but new strategies for sampling and analysis are required to address the combined effects of residues variability and mixing of lots.