phi X174 RFI DNA treated with bleomycin (BLM) under conditions permitting nicking does not serve as a template-primer for Escherichia coli DNA polymerase I. Purified exonuclease III from E. coli and extracts from wild-type E. coli strains are able to convert the BLM-treated DNA to suitable template-primer, but extracts from exonuclease III deficient strains are not. Brief digestion by exonuclease III is enough to create the template-primer, suggesting that the exonuclease III is converting the BLM-treated DNA by a modification of 3' termini. The exonucleolytic rather than the phosphatase activity of exonuclease III appears to be involved in the conversion. Comparative studies with micrococcal nuclease indicate that BLM-created nicks do not have a simple 3'-P structure. Bacterial alkaline phosphatase does not convert BLM-treated DNA to template-primer. The endonuclease VI activity associated with exonuclease III does not incise DNA treated with BLM under conditions not allowing nicking, in contrast to DNA with apurinic sites made by acid treatment, arguing that conversion does not require the endonuclease VI action on uncleaved sites.