The C3H 10T 1/2 clone 8 cell transformation assay was evaluated for possible use in a battery short term assay for determining the carcinogenic potential of chemicals. The assay was originally developed for use with polycyclic aromatic hydrocarbons (PAHs). The assay was tested against 3-methylcholanthrene (56495) (MCA), 7,12- dimethylbenz(a)anthracene (57976) (DMBA), benzo(a)pyrene (50328) (BaP), dibenz(a,h)anthracene (53703) (DBA), the noncarcinogenic PAHs phenanthrene (85018) and anthracene (120127), and N-methyl-N'-nitro- N-nitrosoguanidine (70257) (MNNG). Positive responses were obtained with MCA, DMBA, and BaP. The response to DBA was marginal. MNNG, phenanthrene, and anthracene gave negative responses. The assay was used to test 20 coded carcinogens by two laboratories. Less than 50 percent of the compounds showed a clear transformation response in either laboratory. Forty five percent of the known carcinogens were identified as inducing either positive or suspect activity in one laboratory, whereas 85 percent were identified as either positive or suspect agents in the other laboratory. To broaden the responsiveness of the assay and to limit false negative results, the assay was modified by varying the target cell density, increasing the duration of treatment, delaying the treatment of plated target cells, amplifying the phenotypic transformation by delaying replating of treated target cells, or using exogenous metabolic supplements such as liver S9 mix. Eighteen compounds were evaluated by the C3H 10T 1/2 system after modification. The results were compared with the known in-vivo carcinogenicity of the compounds. Two results were inconclusive. Fourteen compounds, 88 percent, yielded results that agreed with the available in-vivo data. One false negative, pararosaniline (569619) was found and one agent, dimethylnitrosamine (62759) yielded suspect activity. The authors suggest that the modified C3H 10T 1/2 assay can be used as a reliable screening test to complement in-vitro cell assays.