Short-term tests investigating genetic end-points in prokaryotes have been extensively used worldwide not only for risk assessment purposes but also for evaluating the modulation of the mutagenic response. In spite of some intrinsic limitations, such as the lack of cell compartmentalization or the need for an exogenous metabolic system working extracellularly, experimental systems in bacteria can provide useful preliminary indications and some information on the mechanisms involved. In the large majority of studies the putative modulator is mixed with a known mutagen and then assayed in target bacteria, with suitable controls. However, under natural conditions exposure of target cells to modulators may either precede, co-exist with, or follow exposure to mutagens. Therefore, a variety of methodological variations, involving pre-treatment, co-treatment, or post-treatment of bacteria with the putative modulator, have been designed. Application of these procedures showed that the effects of modulators can be completely upset, from inhibition to enhancement, or vice versa, by changing the experimental conditions. Use of methodological variations may provide more complete information on the spectrum of possible effects in bacteria as well as a better insight into modulation mechanisms. Several examples illustrating the flexibility of the Salmonella test in this field of research are available. On the other hand, the widespread use of these relatively simple techniques, yet requiring skillfulness and experience, may lead to some misuse or oversimplifications. A rather common inadequacy is to use excessive amounts of test mutagens, or to express the results in terms of revertants/survivors, rather than revertants/plate. In fact, in the Salmonella test the number of revertants is rather unrelated to the initial number of plated bacteria, provided a normal background lawn of bacterial growth is formed. Thus, a 50% killing of bacteria will not appreciably influence the number of revertants/plate, but expressed as revertants/survivors the effect will look twice as large.