Cells of the mouse T-lymphoma line GRSL13 were treated with 8-methoxy-psoralen plus longwave ultraviolet light (PUVA) under conditions where the biological effects are mainly due to non-persistent DNA cross-links (PUVA-CL treatment). Fluctuation analysis showed that PUVA-CL treatment resulted in an enhancement of the mutation rate in the progeny of treated cells, which persisted until the eleventh generation after treatment. Since only 5 cross-links are available to account for 52 mutational events observed in the coding region, about 90% of the induced mutational events must have been untargeted. This was confirmed by molecular analysis of these mutations, which showed that 53% of the point mutations arose at sites which are not a target for psoralens. This supports the hypothesis that stress responses may give rise to untargeted mutagenesis. Further support for this hypothesis is provided by the observation that 8-methoxy-psoralen (8-MOP) or UVA alone (both of which are known to induce many pleiotropic effects) each acted as indirect mutagen by enhancing the mutation rate 2-4 fold in the progeny of treated cells.