ABSTRACT It is fundamental that before new chemicals may enter the environment of man, they should seem to be (i) desirable and (ii) reasonably safe. Although new chemicals should be treated as separate entities, there is little evidence in general terms to suggest that there is an alternative to screening tests in intact animals, despite the deficiencies of techniques about which much is written. The object of toxicological studies is two-fold, to demonstrate (i) possible reactions irrespective of dosage of new compounds and (ii) the probability in biological terms of these reactions occurring near therapeutic doses (an attempt to prove a negative, the absence of toxicity). Criticism of techniques is easy, especially when considering individual compounds. This introductory paper has attempted to draw attention to the logic which forms the basis of animal screening tests, whilst at the same time commenting on (i) deficiencies due to lack of information and (ii) changes that are likely in the future which will improve extrapolation of results to man. Evidence that traditional tests will be superseded in the foreseeable future appears to the author to be minimal when the whole field of testing new chemicals is considered. The general principles relating to animal studies are "guidelines" which should allow an element of flexibility to treat individual chemicals on their own merits.