The results of subacute toxicity tests are the information most commonly used for setting doses for carcinogenicity studies. However, since diet composition is important in regard to tumor incidence, knowledge of the effect of various doses of the compound on the nutritional composition and quality of the diet of the animals on test is also important. It is also important to determine if any of the doses administered in the study exceed the ability of the species under study to excrete it in a manner which is approximately the same as that seen at lower doses. Similarly, it is important to determine if the higher doses administered lead to the formation of metabolites not seen at lower doses which are more nearly equivalent to those to which humans are or will be exposed. The compound under test should, if possible, be administered by the same route by which human exposure occurs or is anticipated to occur. Because of potential difficulties in the interpretation of the results, gastric intubation is, in general, the least desirable method of administration. In selecting the most appropriate animal species to be used in a carcinogenicity study, consideration should be given to the use of the species and strain of experimental animal that is most biologically and metabolically similar to humans, if such data are available.