The question of how many and which short-term tests (STT) are necessary for a satisfactory characterization of the genotoxic properties of chemicals is still open. The answer is important for both basic mutagenicity research and risk assessment. This paper, aimed at giving a rational answer to the problem, analyzes with multivariate statistical methods the data generated by the International Program for the Evaluation of Short-Term Tests for Carcinogens (IPESTTC). Although it has been found that this data base has a limited reliability for assessing the ability of STTs to predict carcinogenicity, the IPESTTC results are an important source of information on the relationships among different assays, and their ability to identify genotoxic chemicals. A scale of genotoxicity of the chemicals was established by studying with factor analysis their results in 20 IPESTTC tests. The next step of the analysis consisted in the identification of the STT batteries which are the most able to reproduce the genotoxicity scale based on the entire set of STTs. Different batteries were ranked according to their ability to quantify genotoxicity. As a general conclusion, this study indicated that an articulated range of STTs is necessary, and it is not possible to use only one assay (e.g., Salmonella) as an exhaustive indicator of genotoxicity.