The C3H/10T1/2 CL8(10T1/2) cell transformation assay has been used as an important in vitro tool for screening potential carcinogens. In this paper we describe an experimental design that increases the sensitivity and specificity of the assay. One half of the total dishes was allocated to the solvent control and the other half was equally subdivided into four treatment dose groups of low to high cytotoxic effects. The transformation frequency was calculated on the basis of the number of dishes with Type III foci. Each treatment group, as well as the pooled treatment groups, was compared to the solvent control using Fisher's exact test. The sensitivity of our design, as evaluated by power analyses, greatly exceeded that of a standard test design in which about 20 dishes are allocated to each of the control and treatment groups. Furthermore, our use of an expanded number of control and treatment dishes reduces the chance for both false positive and false negative responses. Our experimental design is illustrated with data from experiments in which the transforming potential of two drugs, dimenhydrinate and SC-32006, was examined.