The following summary represents a consensus of the working group except where noted. The items discussed are listed in the order in which they appear in the OECD guideline (473) for easy reference. Metabolic activation. S9 from animals induced either with Aroclor 1254 or with the combination of phenobarbital with beta-naphthoflavone is acceptable, and other systems could be used with suitable justification. Exposure concentrations. The upper limit of testing should be 10 mM (or 5 mg/ml where molecular weight is not known or mixtures are being tested), whichever is lower. Where this limit is inappropriate the investigator should give detailed justification of the choice of top concentration. Cytotoxicity should be measured not only in range-finding tests but also concurrently with the assay for chromosomal aberrations. Cytotoxicity should be assessed by measurements of cell growth such as cell counts or confluence estimation. Mitotic index data alone are not a sufficient measure of cytotoxicity, except in the case of blood cultures for which other methods are impractical. Cytotoxicity at the top dose should be greater than 50% of concurrent negative/solvent controls, if this can be achieved without exceeding a concentration limit of 10 mM or 5 mg/ml. There should be at least three concentrations scored for aberrations (each with and without S9), covering a toxicity range down to a concentration giving little or no cytotoxicity. This will usually mean that the concentrations scored will be quite closely spaced. It was not possible to reach a consensus on the issue of solubility limits. The group did not agree on whether (a) solubility rather than cytotoxicity should be the limiting factor, such that only one top dose with evident precipitate should be scored even if toxicity is not observed, or (b) several concentrations with evident precipitate should be scored for aberrations if this were necessary to obtain cytotoxicity. It was agreed that evidence of precipitation should be determined in the final culture medium. Controls. Concurrent positive controls are required but the working group thought it inappropriate to specify the control chemicals or the degree of response that should be obtained, leaving it up to the test laboratory to demonstrate that the system was working adequately based on historical data within the laboratory. It is not necessary to include both negative and solvent controls concurrently with the aberration test; solvent controls alone are acceptable provided that the laboratory has data to demonstrate that there is no effect of the solvent on baseline values. Preparation of cultures.(ABSTRACT TRUNCATED AT 400 WORDS)