The induction of aneuploidy in cultured mammalian cells has been studied using two chemical agents, diethylstilboestrol dipropionate and Colcemid, which act on the mitotic spindle, in order to investigate different methods of scoring and statistical analysis. Two cell lines, one human and the other Chinese hamster, were each exposed to one test agent, and metaphase preparations scored (a) counting all the chromosomes and (b) counting only the chromosomes of a selected group within the karyotype, to assess which would be the more rapid means of scoring metaphases for aneuploidy. Both the numbers of aneuploid cells and, where appropriate, the total numbers of chromosomes lost and gained were analysed using regression analysis, G-statistics and chi 2 or Fisher's exact test. The results indicate that counting selective chromosomes may in certain circumstances be advantageous, but will result in many more metaphases being required to be scored. Regression analysis to demonstrate dose response is suitable for the analysis of strong positives, but it is suggested that with weak inducers of aneuploidy, the additional tests cited above, using either the number of aneuploid cells or the actual number of chromosomes lost and gained are more sensitive measures of the effects.