In a previous collaborative work, we have recently shown that the cytochalasin-B (Cyt-B) concentration used in the human lymphocytes cytokinesis-block micronucleus (CBMN) assay is an important variable in the baseline micronuclei (MN) frequency as well as in the percentage of binucleated cells obtained. Now we have investigated how Cyt-B concentration modulates the MN frequency induced in whole blood human lymphocyte cultures by two clastogens (ethyl methanesulphonate and mitomycin-C) and two aneugens (colchicine and vincristine sulphate). The experimental design includes six donors, two concentrations of Cyt-B (3 and 6 g/ml), two concentrations of the four chemicals tested and the exchange of slides between laboratories. The statistical analysls of the results shows: (i) non-significant differences in the MN frequencies and in the toxicity results between scorers from each laboratory, except for 0.06 M colchicine at 3 g/ml Cyt-B; (ii) an induction of MN by all genotoxic agents tested, the frequencies being lower with 6 than with 3 g/ml Cyt-B, in control and aneugen-treated cultures; and (iii) significant differences between Cyt-B concentrations in several treatments, obtaining lower MN frequencies and higher values for nuclear division index and % binucleated cells when 6 g/ml Cyt-B was used. Bearing in mind these results as well as the toxicity data showing that 6 g/ml Cyt-B is much more effective in blocking cytokinesis, we can conclude that the use of 3g/ml Cyt-B may overestimate the induced frequency of MN. © 1994 Oxford University Press.