Three different poly(dC)s with modifications that block the N-3 of deoxycytidine were used as templates for polymer synthesis by Escherichia coli DNA polymerase I (EC 2.7.7.7). In contrast to previously reported results with transcriptases, the hydrated form of 3,N(4)-ethenodeoxycytidine (epsilondC.H(2)O) did not mispair. Both 3,N(4)-ethenodeoxycytidine (epsilondC) and 3-methyldeoxycytidine (m(3)dC) led to dTMP misincorporation: 1/20 epsilondC and 1/80 m(3)dC. No other misincorporations appeared to be significant in amount. Thus, both qualitatively and quantitatively, replication errors resulting from carcinogen-modified bases are less frequent than errors in transcription of the same deoxypolynucleotides. Replication of comparable ribopolynucleotide templates by cucumber RNA-dependent RNA polymerase (EC 2.7.7.48) was strongly inhibited by epsilonrC.H(2)O and epsilonrC, so that the fidelity of this enzyme could not be assessed. However, both poly(dC) and poly(rC) containing dU or rU led to incorporation of rA. The presence of even small amounts of purines in poly(rC) greatly depressed synthesis, but the complementary base was incorporated. The finding that an RNA replicase can utilize a deoxypolynucleotide template is a further indication that, at least in vitro, the specificity of the relationship of enzymes and their natural templates is not absolute.