Observable tumours, risk assessment and experimental design.

The first goal of this article is to outline a design for an efficient experiment for testing each of many chemicals for carcinogenic potential by dermal exposure, including a risk assessment when appropriate. The second is to validate a general mathematical model for the dose-response function in this situation by using the results from a large experiment in skin painting done by the Oak Ridge National Laboratory. These two goals are related because the general model is used as a guide in the choice of an efficient design for testing the many chemicals on which information is required.