Recently we and others have reported a high frequency of mutagenesis of shuttle vector plasmids after passage in mammalian cells (Razzaque et al., Proc. Natl. Acad. Sci. U.S.A. 80:3010-3014, 1983; Calos et al., Proc. Natl. Acad. Sci. U.S.A. 80:3015-3019, 1983). The mutation frequency was determined by monitoring the integrity of a bacterial marker gene on the plasmid by standard microbiological procedures. Mutant plasmids contained deletions, insertions of cell DNA, and point mutations. The observed mutation frequency of 1% is much higher than that of cellular markers and could be due to the induction of a mutagenic environment by infection with a replicating plasmid. Alternatively, the hypermutagenesis may be due to some critical transient or persistent difference between the DNA in the plasmid and the cellular chromosome. We performed a number of experiments designed to distinguish between these alternatives, with particular reference to deletion mutagenesis. We conclude that mutagenesis was specific to the plasmid and propose that the majority of the deletion and insertion mutants were generated very early in the infection, before replication of the vector. However, some deletion mutagenesis also occurred after plasmid replication had begun.