For two main reasons. First, the NOAEL is determined in animals, not humans. It is therefore prudent to adjust for possible differences by assuming that man is more sensitive than the most sensitive test animal. Secondly, the reliability of toxicity tests is limited by the number of animals tested. Such tests cannot represent the diversity of the human population, subgroups of which may show different sensitivities (e.g. children, the old and the ill). Again, it is prudent to adjust for these differences.