Aniline (in the form of its hydrochloride) has been shown to induce a rather rare spectrum of tumors in the spleen of Fischer 344 rats. The dose levels necessary for this carcinogenic activity were in a range where also massive effects on the blood and non-neoplastic splenotoxicity as a consequence of methemoglobinemia were to be observed. This review aimed at clarifying if aniline itself or one of its metabolites has a genotoxic potential which would explain the occurrence of the spleen tumors in rats as a result of a primary genetic activity. The database for aniline and its metabolites is extremely heterogeneous. With validated assays it ranges from a few limited Ames tests (o- and m-hydroxyacetanilide, phenylhydroxylamine, nitrosobenzene) to a broad range of studies covering all genetic endpoints partly with several studies of the same or different test systems (aniline, p-aminophenol, p-hydroxyacetanilide). This makes a direct comparison rather difficult. In addition, a varying number of results with as yet not validated systems are available for aniline and its metabolites. Most results, especially those with validated and well performed/documented studies, did not indicate a potential of aniline to induce gene mutations. In five different mouse lymphoma tests, where colony sizing was performed only in one test, aniline was positive. If this indicates a peculiar feature of a point mutagenic potential or does represent a part of the clastogenic activity for which there is evidence in vitro as well as in vivo remains to be investigated. There is little evidence for a DNA damaging potential of aniline. The clastogenic activity in vivo is confined to dose levels, which are close to lethality essentially due to hematotoxic effects. The quantitatively most important metabolites for experimental animals as well as for humans (p-aminophenol, p-hydroxyacetanilide) seem to have a potential for inducing chromosomal damage in vitro and, at relatively high dose levels, also in vivo. This could be the explanation for the clastogenic effects that have been observed after high doses/concentrations with aniline. They do not induce gene mutations and there is little evidence for a DNA damaging potential. None of these metabolites revealed a splenotoxic potential comparable to that of aniline in studies with repeated or long-term administration to rats. The genotoxicity database on those metabolites with a demonstrated and marked splenotoxic potential, i.e. phenylhydroxylamine, nitrosobenzene, is unfortunately very limited and does not allow to exclude with certainty primary genotoxic events in the development of spleen tumors. But quite a number of considerations by analogy from other investigations support the conclusion that the effects in the spleen do not develop on a primary genotoxic basis. The weight of evidences suggests that the carcinogenic effects in the spleen of rats are the endstage of a chronic high-dose damage of the blood leading to a massive overload of the spleen with iron, which causes chronic oxidative stress. This conclusion, based essentially on pathomorphological observations, and analogy considerations thereof by previous authors, is herewith reconfirmed under consideration of the more recently reported studies on the genotoxicity of aniline and its metabolites, on biochemical measurements indicating oxidative stress, and on the metabolism of aniline. It is concluded that there is no relationship between the damage to the chromosomes at high, toxic doses of aniline and its major metabolites p-aminophenol/p-hydroxyacetanilide and the aniline-induced spleen tumors in the rat.