In vitro chromosomal aberration (CA) tests have come to play a central role in testing for mutagenic/carcinogenic potential of chemicals in most countries. Guidelines on the conduct of such assays have therefore been published by a variety of sources, and, if anything, recommended protocols have become even more extensive as time has progressed and revisions have been made. Yet there is very little comparative data from within or between laboratories to form a basis for these recommendations. Some claims that certain cell types were more sensitive than others for CA testing has led to comparisons between Chinese hamster ovary (CHO), Chinese hamster lung cells and human lymphocytes, and some examination of critical factors such as exposure periods and sampling times has been undertaken, but much more needs to be done. Clones of CHO and V79 cells from different sources show karyotypic variability are not equally stable, further confounding any comparisons of sensitivity. There would therefore seem to be a need to acknowledge genetic diversity in these cell lines, particularly when making comparisons of sensitivity. As in vitro CA tests employ more and more comprehensive protocols, an understanding or artefacts becomes more critical. Extreme culture conditions (pH shift, high osmolality, high ionic strength) have given rise to significant CA, but are unlikely to occur in vivo. It is possible that chemicals which only produce CA detectable at high levels of cytotoxicity are also not in vivo hazards, although experimental data is urgently needed to confirm or deny this. High toxicity clastogens (HTC) would not have the same biological importance as those inducing CA at low levels of toxicity [(low toxicity clastogens (LTC)]. However, current study design is generally inadequate to discriminate LTC from HTC, and agreement on the level of toxicity that divides them will be a contentious issue. Proposals are made for improved study design, in particular the spacing of doses to permit categorization of chemicals as HTC or LTC. A concept of comparing minimal positive (clastogenic) dose with an arbitrary level of toxicity (e.g. 40-50% inhibitory dose) is introduced to permit this categorization. As an indication of likely in vivo hazard, categorizing a chemical as either HTC or LTC can help with decision making in industry and risk assessment by industrialists and regulators.