The results are reported of the second stage in a programme for a systematic in vitro study on the carcinogenic potential of metal compounds with Balb/3T3 clone A31-1-1 mouse fibroblasts. Nineteen metal compounds that exhibited a strong cytotoxic effect during a previous screening run with a 100 microM fixed dose were tested with a 72-hour exposure over a wide range of concentrations (from 0.1 microM to 1000 microM), to produce dose-effect curves to permit extrapolation of the 50% inhibition concentration (IC50) values for each metal compound. This allows the establishment of a suitable range of doses for individual metal species, for use in the subsequent Balb/3T3 assay based on a two-stage concurrent cytotoxicity and morphological transformation protocol. Another 13 metal compounds were also tested, to determine whether the Balb/3T3 cell transformation assay is really a valuable in vitro model in relation to the problem of metal speciation. Of the metal compounds assayed, 26 showed a dose related cytotoxic response with calculated IC50 values ranging from 0.25 microM (CH3HgCl) to 140 microM [(C5H5)2TiCl2], whereas six metal compounds, namely (NH4)6Mo7O24*4H2O, CH3AsO(OH)2, C2H6AsNaO2(3H2O, KBr, CrCl3*6H2O and (NH4)2[TiO(C2O4)2]*H2O, displayed no observable cytotoxicity or low cytotoxicity at all the doses tested. The determination of IC50 values permits a ranking of the cytotoxicity responses of metal compounds with the highest cytotoxicities. Dose-effect curves and IC50 values of different chemical forms of individual metal compounds of As, Br, Cr, Hg, Ir, Pt, Te, Ti and V (cationic/anionic inorganic or organometallic species) showed clearly how the chemical nature of the metal strongly influences the toxic response. This confirms that the Balb/3T3 cell line is a valuable in vitro model with respect to the problem of metal speciation. This is a fundamental aspect to be considered when incorporating the results from in vitro cell transformation assays of the carcinogenic potential of metal compounds into regulatory testing schemes. In this context, the choice of test metal species for the development and validation of such assays cannot disregard the possibility that humans will be exposed to specific chemical forms of individual metal compounds (different oxidation states, and inorganic or organometallic natures) that can profoundly affect their toxicity.