Still another critical issue is the uncertainty with the empirical data used in the risk analysis conducted here. For example, as mentioned repeatedly in this lecture, all tolerance limits or screening values are toxicity-based because they are derived from the ADI or RfD used, which relies on the toxicity data on hand. These toxicity data, mostly on test animals, are never sufficient and often times do not provide a direct link between exposure and effect in humans. At the same time, data on POC residues in fish are quite vulnerable to methodological or technical errors inherent in sample collection, in sample preparation, as well as in the analytical procedures used.