RNA polymerase (nucleosidetriphosphate: RNA nucleotidyltransferase, EC 2.7.7.6) was purified from rifampicin-resistant Bacillus subtifis, from both uninfected cells and cells infected with bacteriophage SPOl. The enzyme from infected cells lacked all traces of the sigma subunit, contained several polypeptides absent from the enzyme made in uninfected cells, and had an altered template specificity in a transcription assay. A cell-free protein synthesizing system from Escherichia coli, when poisoned with rifampicin, was completely dependent on addition of either of these RNA polymerase preparations for DNA-dependent protein synthesis. Under these conditions, the SPOI-modified RNA polymerase preferentially stimulated the synthesis of functional mRNA for the phage enzyme dCMP deaminase (deoxycytidylate aminohydrolase, EC 3.5.4.12), whereas unmodified B. subtilis RNA polymerase could stimulate synthesis of this mRNA in small quantity and only after prolonged incubation. This mRNA belongs to a class of phage transcripts (m) which cannot be transcribed in vivo in the absence of phage-specific protein synthesis.Transcription of the genome of bacteriophage SP01 has been extensively studied by competitive RNA-DNA hybridization (1). Six classes of transcripts have been described, which are independently controlled with respect to the turning on or turning off of their synthesis. Two of these (e and em) are synthesized after infection without detectable lag, and are subsequently repressed if phage protein synthesis is allowed to occur. Three additional classes of RNA (m, mll and m21) are made independently of phage DNA replication, but require prior phage protein synthesis. In particular, the protein product of cistron 28 is required for m and m1l expression, and the protein products of cistrons 33 and 34 are required for M21 RNA synthesis (2,3). After phage DNA synthesis begins, another population of transcripts (1) is made.The existence of three pre-replicative classes of phagespecific transcripts (m, m1l, and M21), whose synthesis is under phage genetic control, suggested the possibility that mature phage DNA extracted from purified virus might be an appropriate template for a new RNA polymerase (nucleosidetriphosphate: RNA nucleotidyltransferase, EC 2.7.7.6) of the correct specificity for these classes. Since it had been shown that the rifampicin sensitivity of RNA synthesis throughout SP01 infection was determined by the sensitivity of the host's RNA polymerase (4) it was logical to expect that the new specificity for middle RNA synthesis should reside in the host's RNA polymerase, which had been modified in some way. Duffy and Geiduschek (5) demonstrated that the expected in vitro selectivity for phage middle transcripts does, indeed, reside in an enzyme which contains the host core and other peptides which are made after phage infection.Competitive RNA-DNA hybridization is a very powerful tool which can be used to determine the relative amount of specific nucleotide sequences in a population of in vitro synth...