A rifampicin-resistant poly(G) polymerase has been purified from f2 sus 11-infected cells. The poly (G) polymerase is believed to represent part of the f2 replicase on the basis of several criteria. It is present only in infected cells and shares the characteristic rifampicin resistance of crude f2 replicase activity. Partially purified poly(G) polymerase preparations exhibit replicase activity, synthesizing f2 'plus' strand RNA from denatured, partially doublestranded f2 RNA template. Highly purified poly(G) polymerase preparations, although lacking replicase activity, contain a protein which is electrophoretically identical to the protein product of the viral replicase cistron.The replicative cycle of RNA bacteriophages, particularly those of the f2 serotype, has been analyzed in some detail (1, 2). The structure of the RNA species produced during the infectious cycle is known, as is the temporal sequence of their synthesis (2). Studies of RNA metabolism in cells infected with various phage mutants indicate that the enzyme or enzyme complex responsible for phage RNA synthesis contains both bacterial and viral components (1, 3). In vitro studies of the f2-type phage RNA replicases have, however, been hampered by the instability of these enzymes and the difficulty in obtaining enzyme preparations free of endogenous template (2, 4-6). Thus, almost all of the currently available information on the structure and enzymology of phage replicase has come from studies of the more stable enzyme of the serologically unrelated RNA phage QB. The Q,3 enzyme appears to be quite complex, requiring several bacterial proteins in addition to a phage-specified protein to carry out net synthesis of phage 'plus'* strand RNA from 'plus' strand template (7-9). Several laboratories have shown that the Q, replicase contains a simpler 'core' enzyme defined by its ability to utilize phage 'minus't strands or poly(C) as template and its inability to use phage 'plus' strand template (7-10). Our studies of the f2 replicase were predicated on the assumption that the f2 enzyme is similarly subdivisible and that the 'core' enzyme would prove more amenable to purification and analysis than the complete enzyme.We report here a purification of the f2 replicase based on its rifampicin-resistant poly(C)-dependent poly(G) polymerase activity. The slurry was allowed to stand for 1 hr at room temperature, diluted 50-fold with cold absolute ethanol, filtered under suction through Whatman No. 3 paper, and allowed to dry. The powder was suspended in 20 ml of absolute ethanol/500 mg of cellulose, made 10 mM in magnesium acetate, and UV-irradiated for 30 min as described by Litman (17). The ethanol was removed by filtration and the powder was dried at room temperature. Unbound RNA was removed