Our inter- and intralaboratory results showed that the SHE transformation assay is not yet sufficiently developed for a large scale screening program. The major source of variability is largely due to the low number of transformed colonies induced by test chemicals in any one experiment. Although this limitation presumably can be overcome by scoring a larger number of colonies, we feel such an approach would defeat the practical purpose of a screening assay. Of the experimental variables examined, we believe that medium supplements other than fetal calf serum and alternative detection/selection methods for transformation are two areas that need further development in order to improve the reproducibility of the SHE assay system for testing diverse chemicals. Until then, this system under defined conditions should remain a useful tool for research purposes.