Cinnamaldehyde, a widely used flavoring agent, has so far been subjected to a limited range of genotoxicity tests, mainly carried out in vitro, which produced contradictory results. Therefore we have examined cinnamaldehyde using additional in vivo genotoxicity end-points. In Sprague-Dawley rats, a single oral dose equal to 1/2 LD50 did not induce DNA fragmentation in liver and gastric mucosa as evaluated by the alkaline elution technique, increased the frequency of micronucleated hepatocytes but not of bone marrow micronucleated polychromatic erythrocytes, and gave rise to a significantly higher incidence of total nuclear anomalies but not of micronucleated cells in forestomach mucosa. In Swiss mice, the equitoxic dose of cinnamaldehyde caused the same clastogenic effect in the liver, whilst a negative response was observed in both bone marrow and forestomach mucosa. Finally, in rats initiated with N-nitrosodiethylamine the administration of 500 mg/kg/day cinnamaldehyde for 14 successive days produced a modest but statistically significant increase of the average diameter and area of gamma-glutamyltranspeptidase-positive foci that, together with changes observed in other parameters, might be considered indicative of a potential promoting activity. Taken as a whole, these findings confirm that high doses of cinnamaldehyde may induce genetic alterations at the chromosomal level, and suggest that the liver is the preferential target of its undesirable effects.